12/6/2020 --------- [bash-5.1 released] 12/7 ---- Makefile.in - bashline.o: add dependency on ${DEFDIR}/builtext.h. Report from Fazal Majid 12/11 ----- builtins/wait.def - wait_builtin: don't assign the variable given with -p if there are no jobs to wait for. Report and fix from Ouz arrayfunc.c - kvpair_assignment_p: return non-zero if argument L appears to be a key-value pair associative array compound assignment - expand_and_quote_kvpair_word: run a single word in a key-value pair associative array compound assignment through the appropriate expansions and single-quote the result arrayfunc.h - kvpair_assignment_p, expand_and_quote_kvpair_word: extern declarations subst.c - expand_oneword: detect whether VALUE appears to be a key-value pair compound assignment and call the appropriate function to expand each word in the resulting list. Fixes inconsistency reported by oguzismailuysal@gmail.com 12/12 ----- subst.c - command_substitute: don't reset pipeline_pgrp to shell_pgrp if we are already forked to run a command (SUBSHELL_FORK). Fixes SIGINT in command substitution in here-document in forked child issue reported by oguzismailuysal@gmail.com 12/18 ----- execute_cmd.c - execute_pipeline: execute the lastpipe code even if prev == 0. It can only be 0 here if stdin was closed when this function was executed - execute_pipeline: if prev == 0, set lstdin to a sentinel (-1) that means to close fd 0 after executing the command, and call close(prev) before restoring stdin. restore_stdin now understands -1, and closes fd 0. Fixes issue reported by Tomas Janousek 12/21 ----- doc/bashref.texi - PROMPT_COMMANDS: clean up a couple of remaining instances of this name. Report from Eli Schwartz 12/26 ----- subst.c - command_substitute: make sure that the child process always has pipeline_pgrp == shell_pgrp: if pipeline_pgrp is non-zero when we get to the child, meaning that we're part of an already-forked child that is, for instance, running redirections, we need to reset shell_pgrp to it to preserve the invariant. Fixes bug with terminal being set to the incorrect pgrp reported by oguzismailuysal@gmail.com 12/29 ----- configure.ac,builtins/shobj-conf,m4/threadlib.m4 - midnightbsd: update auto-configuration to treat MidnightBSD like FreeBSD. From https://savannah.gnu.org/patch/?10006 12/30 ----- examples/loadables/stat.c - stattime: use strftime with a default format or the format supplied with the -F option to format the file time information - stat_builtin: now takes a -F fmt option for a strftime format string; change to function calling sequences to pass the format down to stattime examples/loadables/Makefile.in - stat: now built and installed by default 12/31 ----- command.h - SUBSHELL_IGNTRAP: new flag value jobs.c - make_child: set SUBSHELL_IGNTRAP in subshell_environment in the child process, meaning that we should not execute trap handlers for trapped signals execute_cmd.c - execute_in_subshell, execute_simple_command, execute_disk_command: make sure to unset SUBSHELL_IGNTRAP after a child process restores the original signal dispositions - execute_simple_command: make sure to set SUBSHELL_IGNTRAP after make_child returns and we're setting subshell_environment directly subst.c - command_substitute,process_substitute: unset SUBSHELL_IGNTRAP after the child process has reset the trapped signal dispositions trap.c - trap_handler: if we get a signal for which we have set a trap, but SUBSHELL_IGNTRAP is set in subshell_environmnent, make sure we restore the original signal disposition and resend the signal to ourselves. Fixes issue reported by Nikolay Borisov (or at least makes the race window much smaller) sig.c - initialize_terminating_signal: set the original signal handler from the return value from sigaction; a minor optimization that saves a system call or two 1/5/2021 -------- builtins/declare.def - declare_internal: make some option combinations that don't make sense errors (e.g., -f and -a/-A/-i/-n) - declare_internal: if we build a new variable name by expanding the value of a nameref variable, make sure to chop the `+' in a `+=' off before going on 1/7 --- doc/{bash.1,bashref.texi} - bind: add an example to the synopsis making it clear that you can use any readline command line as a non-option argument, as it says in the text. From a report from Dan Jacobson 1/12 ---- locale.c - local_shiftstates -> locale_shiftsates in the non-multibyte code branch. Reported by Henry Bent subst.c - expand_compound_assignment_word: make sure to call dispose_words on the WORD_LIST * returned from expand_oneword after turning it back into a string. Fixes memory leak reported by Alexander Mescheryakov 1/13 ---- variables.c - bind_variable_internal: when performing an assignment to a subscripted array variable that was the value of a nameref (used in the original assignment), don't call make_variable_value on the value, since that messes up +=. Just call assign_array_element and let that take care of calling make_variable_value appropriately. Fixes bug reported by Oguz 1/14 ---- findcmd.c - search_for_command: if `checkhash' is set, don't add non-executable files to the command hash table, since we will just remove them later lib/sh/winsize.c - get_new_window_size: set *rp and *cp even if READLINE is not defined 1/15 ---- lib/sh/winsize.c - get_new_window_size: call rl_set_window_size only if we can determine we're using readline: an interactive shell without no-line-editing, or if we've already initialized readline, presumably in a non- interactive shell support/Makefile.in - man2html: add LDFLAGS_FOR_BUILD to the recipe. Report from Jeffrey Walton 1/17 ---- lib/readline/misc.c - rl_operate_and_get_next: fix old K&R function declaration. Report from Tom Tromey lib/readline/readline.c - _rl_internal_char_cleanup: move code that cleans up the active region and deactivates the mark inside this function so callback mode applications get the intended functionality. Report and fix from sparrowhawk996@gmail.com lib/readline/bind.c - rl_parse_and_bind: when using the arithmetic comparison operators on the version, make sure to invert the tests so that we stop parsing if the test fails. Report and fix from Tom Tromey 1/19 ---- Makefile.in - pipesize.h: add dependency on ${BUILTINS_LIBRARY} to avoid parallel makes trying to create it twice. Report and fix from Richard Purdie 1/21 ---- subst.c - param_expand: if a nameref expands to array[@] or array[*], make sure to call chk_atstar so the right variables are set to split the result. Report from Oguz 1/22 ---- builtins/declare.def - Rewrote to reduce complexity. Still needs some work. lib/readline/*.h, lib/tilde/tilde.h - PARAMS: remove, rely on compilers understanding function prototypes lib/readline/{undo.c,histlib.h} - _hs_replace_history_data: move extern declaration to histlib.h lib/readline/readline.c - _rl_parse_colors: instead of an extern declaration for this, include "parse-colors.h" for it lib/readline/{histfile.c,histlib.h} - _hs_append_history_line: move extern declaration to histlib.h lib/readline/history.h - HS_HISTORY_VERSION: define to 0x0801 (current library version) if it's not already defined. We conditionally define it in case an application has (unwisely) chosen to use it, since the history library doesn't do anything with it yet lib/readine/{rlprivate.h,{search,misc,readline}.c} - _rl_free_history_entry: add extern declaration to rlprivate.h, remove extern declaration from c source files. Use HS_HISTORY_VERSION as a sentinel that it's ok to use HIST_ENTRY in rlprivate.h lib/readline/{rlprivate.h,{isearch,search,undo}.c} - _rl_saved_line_for_history: add extern declaration to rlprivate.h, remove extern declaration from c source files, using HS_HISTORY_VERSION in the same way 1/24 ---- lib/readline/signals.c - _rl_handle_signal: make sure that all sigprocmask calls are protected by HAVE_POSIX_SIGNALS. Report and fix from Tom Tromey 1/26 ---- lib/readline/callback.c - rl_callback_read_char: make sure rl_linefunc is non-NULL before calling through the pointer. The line function could have been removed by the application before readline processes any typeahead input. Bug reported by Matthias Klose , pointer to root cause from Koichi Murase lib/glob/glob.c - skipname,wskipname: put in some additional checks for `.' to ensure that we don't get false positives (or incomplete tests) that can affect the results of extglob patterns. Reported by gregrwm 1/27 ---- lib/glob/glob.c - extglob_skipname,wextglob_skipname: fixed an off-by-one error (SE was being set to one character before the end of the pattern string, not the null character at the end of the pattern string like other callers of glob_patscan) that caused the last character of the last subpattern to be cut off when calling skipname - extglob_skipname,wextglob_skipname: some cleanups so the code is closer to identical for the single-byte and wide character versions 2/1 --- execute_cmd.c - execute_simple_command: in posix mode, if we have a variable assignment error while assigning into the temporary environment (e.g., assigning to a readonly variable), a non-interactive shell running a special builtin exits; a non-interactive shell running anything else jumps back to the top level. A shell compiled with -DSTRICT_POSIX exits unconditionally. - execute_simple_command: make sure posix mode sets $? to non-zero if a variable assignment error occurs preceding a non-special builtin subst.c - do_assignment_statements: take the code from expand_word_list_internal that performs assignment statements, either standalone or preceding simple command names, and factor it out into this function - expand_word_list_internal: call do_assignment_statements where appropriate 2/2 --- lib/glob/glob.c - dequote_pathname: fix function definition for non-multibyte systems. Report and fix from Marc Aurèle La France Makefile.in,doc/Makefile.in - for certain targets, remove files before creating them to deal with symlinked build trees. Report and fix from Marc Aurèle La France examples/loadables/accept.c - include limits.h before typemax.h Report and fix from Marc Aurèle La France builtins/gen-helpfiles.c - if USING_BASH_MALLOC is defined, make sure to undefine malloc as well as free. Fixes bug reported by George R Goffe builtins/Makefile.in - install-help: now depends on $(HELPFILES_TARGET) so we make sure the separate helpfiles are created before we try to install them if we don't go through the `all' makefile target configure.ac - HELPDIR: now ${datadir}/bash/helpfiles 2/3 --- parse.y - parse_string_to_word_list: before expanding a compound assignment statement body, make sure to save any alias that's currently being expanded. Restore the alias after the compound assignment is parsed. Reported back in 11/2020 by Alex fxmbsw7 Ratchev 2/4 --- lib/readline/histexpand.c - history_expand_internal: when calling the history_inhibit_expansion function, make sure to call it using the string as we've expanded it to that point (RESULT), adding the expansion and next characters temporarily, since we make expansion decisions based on what we've accumulated, not what we started with. This makes things like echo abc!$!$ work, where before the second `!' inhibited expansion because bash_history_inhibit_expansion mistakenly took it as the second character in a `$!' word expansion. Fixes bug reported back in 10/2020 by Paul Fox array.h - array_pop: instead of calling array_dispose_element from this macro, just call array_shift with the AS_DISPOSE flag 2/5 --- bashhist.c - shell_comment: move condition to return 0 if the delimiter stack is not empty or the shell is parsing a here document into the function itself, don't have the callers check so the check is in one place. Fixes bug reported by Oguz array.h,variables.c - ARRAY_ELEMENT_REPLACE: convenience define for modifying an array element's value variables.c - pop_args: a couple of code simplifications 2/7 --- lib/malloc/malloc.c - pagesz: at least MALLOC_PAGESIZE_MIN (4096) bytes - union mhead: now 16-byte aligned on all systems, 32-bit and 64-bit pointers - binsizes: since the smallest allocation overhead is now 16 bytes, redo the buckets so binsizes[0] == 32; adjust the thresholds for split/coalesce/prepopulate/mmap (NBUCKETS = 28; STARTBUCK = 0). Sizes stay pretty much the same; indices change - consistently use MALLOC_SIZE_T instead of long/unsigned int/int - use MAGIC8_NUMBYTES as the length of the mh_magic8 buffer, in case it changes later for alignment - internal_remap: new function, calls mremap to reallocate a chunk of memory allocated using mmap(); called from internal_realloc if the old size and new size are both bigger than the mmap threshold - internal_realloc: call internal_remap if the old size and new size are both above the threshold where we use mmap for allocation 2/10 ---- include/timer.h - new file, declaration for a timer struct to be used by a set of functions to implement timers using SIGALRM or select/pselect lib/sh/timers.c - new file, set of functions to manipulate timer objects and timeouts using SIGALRM or select/pselect. Inspired by a patch from Koichi Murase . Not used yet 2/11 ---- builtins/read.def - read_builtin: if there is a timeout set, block SIGCHLD around calls to zread and its siblings, or calls to readline for `read -e', so SIGCHLD (and the consequent waitpid) doesn't interrupt the read. Fixes bug reported by Koichi Murase , but there may be a different fix coming Makefile.in,builtins/Makefile.in - fix up dependencies, especially on builtins.h and builtext.h 2/12 ---- lib/readline/input.c - rl_read_key: if we set rl_done == 1, set RL_STATE_DONE as well Reported by Koichi Murase lib/readline/isearch.c - _rl_search_getchar: only call _rl_read_mbstring if rl_read_key returns >= 0, avoid some work lib/readline/vi_mode.c - _rl_vi_callback_change_char,_rl_vi_change_char: don't overwrite the last replacement string if _rl_vi_callback_getchar returns -1. It will likely make no difference, since the next read will return an error or EOF, but being careful - rl_vi_overstrike: if _rl_overwrite_char doesn't return 0, break out of the loop lib/readline/text.c - _rl_overwrite_char: return 1 if _rl_read_mbstring returns < 0 so we don't try to insert garbage bashline.c - posix_edit_macros: handle rl_read_key() returning <= 0 2/15 ---- parse.y - read_comsub: make sure to turn on the LEX_RESWDOK flag if we are in a case statement and read a `)', since we can get a valid `esac'. Fixes bug reported by Oguz - read_comsub: if we're in a case statement, recognize `}' as a reserved word and set the LEX_RESWDOK flag for the next word, since we can get an esac (or another reserved word) after it 2/16 ---- parse.y - reserved_word_acceptable: add ARITH_CMD and COND_END to the list of tokens that can precede a reserved word, so you can use reserved words after ((...)) and [[...]]. Reported by Koichi Murase 2/17 ---- parse.y - parse_comsub: use new LEX_CASEWD flag to track when we are reading the WORD in `case WORD in' and turn on the LEX_RESWDOK flag when that word ends. This allows $(case x in esac), which no one uses. - parse_comsub: use LEX_PATLIST flag to track when we are reading a case pattern list so `|' doesn't turn on the LES_RESWDOK flag - parse_comsub: case_level: simple counter to count the number of esacs we need to see before we're no longer in a case statement; analog of esacs_needed_count from the lexer 2/19 ---- parse.y - CHECK_FOR_RESERVED_WORD: don't return ESAC if we read `esac' after a left paren in a case pattern list. From an austingroup-bugs discussion about https://www.austingroupbugs.net/view.php?id=1454 - parse_comsub: if we read a `(' while looking for a case pattern list and LEX_CKESAC is set, we have a leading left paren in the pattern list and should turn off LEX_CKESAC so (esac) doesn't prematurely terminate the case command. From an austingroup-bugs discussion about https://www.austingroupbugs.net/view.php?id=1454 2/26 ---- builtins/history.def - history_builtin: when checking negative offsets to -d, which are supposed to count back from the end of the history list, check the range against 0 instead of history_base, because the calculation is done against history_length, which is independent of history_base. Report and fix from Christopher Gurnee 2/28 ---- doc/bashref.texi - replaced a number of uses of @var with a mixture of @env and @dfn to better match up with the texinfo standards doc/{bash.1,bashref.texi} - clarify some aspects of the coproc description, especially the use of NAME and when it's optional 3/1 --- subst.c - read_comsub: fix off-by-one error in mbrtowc that causes a read one character past the end of buf. Report and fix from Platon Pronko in https://savannah.gnu.org/patch/?10035 3/3 --- builtins/ulimit.def - ulimit_builtin: Posix compatibility: if the last command specified by an option does not have an option argument, but there is an operand remaining after all the options are parsed, treat the operand as an argument to that last command. From an austin-group discussion and a Geoff Clare suggestion back in November, 2020. Austin Group interpretation 1418 examples/shellmath - a package of shell functions to perform floating-point math entirely in bash. Contributed by Michael Wood . Available at https://github.com/clarity20/shellmath 3/4 --- support/shobj-conf - darwin: take out the -arch-only option in SHOBJ_XLDFLAGS and SHOBJ_ARCHFLAGS; no longer needed doc/{bash.1,bashref.texi} - coprocesses: suggested changes from rms@gnu.org; recommend the `coproc NAME { commands; }' form as the simplest and most flexible 3/5 --- builtins/exec.def - exec_builtin: set last_command_exit_value before calling exit_shell so any exit trap gets the right value for $?. From Matthew Bauer via https://savannah.gnu.org/patch/?10039 3/8 --- include/timer.h - SHTIMER_ALRMSET: new flag, indicates that there is an active alarm associated with this timer (falarm() was called) lib/sh/timers.c - shtimer_set: set the SHTIMER_ALRMSET flag after calling falarm - shtimer_unset: don't call falarm(0,0) unless the SHTIMER_ALRMSET flag is set 3/9 --- include/posixtime.h - added some BSD convenience defines if they are not present parse.y,shell.h - {save,restore}_parser_state: save and restore shell_eof_token and pushed_string_list; change callers (e.g., xparse_dolparen) so they don't have to manage them 3/10 ---- builtins/common.h - extern declarations for moving to timers (sh_timer) for read builtin timeouts quit.h - CHECK_ALRM: remove, no longer used trap.c - check_signals: call check_read_timeout instead of CHECK_ALRM bashline.c - bash_event_hook: use read_timeout instead of checking `sigalrm_seen'; that no longer exists - bash_event_hook: accommodate readline timing out (not used yet) lib/sh/zread.c - zread: call read_builtin_timeout() to check for a timeout before calling a blocking read() builtins/read.def - sigalrm,reset_timeout,check_read_timeout,read_builtin_timeout: new and modified functions to use sh_timers for timeouts instead of SIGALRM. Based on work contributed by Koichi Murase - read_builtin: use sh_timers for read timeouts (-t N) instead of using SIGALRM - edit_line: simulate receiving SIGALRM if readline times out (not used yet) 3/11 ---- lib/readline/readline.c - rl_initialize: call _rl_timeout_init to set things up for any timeout that was set with rl_set_timeout - readline_internal_charloop: if we longjmped because of a timeout, make sure to set rl_done/RL_STATE_DONE and return; we are abandoning this call to readline(). The readline timeout changes were based on work contributed by Koichi Murase lib/readline/readline.h - extern declarations for new timeout functions and hook - rl_clear_timeout: new define lib/readline/callback.c - rl_callback_read_char: if we longjmped because of a timeout, make sure to set rl_done/RL_STATE_DONE and return; we are abandoning this call to readline() lib/readline/util.c - _rl_abort_internal: if we time out, don't ring the bell; let the caller handle it lib/readline/input.c - extern declarations for public and readline-library-private functions and hooks to implement timeouts - rl_set_timeout,rl_timeout_remaining: new public functions - _rl_timeout_select: new function, uses select/pselect to implement read timeouts that take timeouts set with rl_set_timeout into account; calling hook function if a timeout occurs - rl_gather_tyi, _rl_input_available: use _rl_timeout_select, taking any existing timeout into consideration if it expires before the timeout passed as an argument - rl_getc: use _rl_timeout_select and handle any timeouts by calling _rl_timehout_handle - set_alarm,reset_alarm: new functions to implement timeouts using SIGALRM for systems that lack a working select/pselect - _rl_timeout_init: new function, sets things up for reading input with a specified timeout - _rl_timeout_handle: a timeout handler; calls any event hook and sets up to abort the current readline() call - _rl_timeout_handle_sigalrm: a timeout handler for systems using SIGALRM to implement timeouts lib/readline/parens.c - rl_insert_close: use _rl_timeout_select to take timeouts into account lib/readline/rlprivate.h - extern declarations for readline-library-private timeout functions lib/readline/rltty.c - rl_deprep_terminal: don't print a newline after the bracketed paste disable sequence if we timed out lib/readline/signals.c - _rl_handle_signal: if sig is SIGALRM, call _rl_timeout_handle_sigalrm() lib/readline/doc/rltech.texi - rl_set_timeout,rl_timeout_remaining: document new public functions - RL_STATE_TIMEOUT: document new possible state value for rl_readline_state - rl_timeout_event_hook: document new hook function, called when readline times out builtins/read.def - read_builtin: changes to use the readline timeout functions to implement timeouts with `read -e'; these use rl_set_timeout and sh_timer structs together 3/12 ---- subst.c - expand_string_dollar_quote: new function, expands $'...' and $"..." in a string for those code paths that don't expand it themselves subst.h - expand_string_dollar_quote: extern declaration parse.y - read_secondary_line: if $'...' or $"..." appears in the line, call expand_string_dollar_quote to expand them. This now returns new memory, need to change callers make_cmd.c - make_here_document: account for read_secondary_line returning newly allocated memory, free `full_line' appropriately bashline.c - shell_expand_line,history_and_alias_expand_line: expand $'...' and $"..." in the line by calling expand_string_dollar_quote, since that happens after history expansion and before alias expansion in normal processing 3/15 ---- subst.c - expand_string_dollar_quote: fix out-of-order initialization Makefile.in - {TAGS,tags}: add ETAGS/ETAGSFLAGS/CTAGS/CTAGS flags; make sure to cd to the source directory before running them to get source files that don't have absolute paths. Fix from Mike Jonkmans parse.y - xparse_dolparen: don't longjmp if FLAGS includes SX_NOLONGJMP. From a report by Xu Lu 3/16 ---- subst.c - process_substitute: set startup_state and parse_and_execute_level to see if we can avoid a fork() bashline.c - bash_spell_correct_word: bindable command (spell-correct-word) to perform spelling correction on the current `shellword', using the same code as the `cdspell' option and directory spelling correction during completion. Feature suggested by in 10/2020 by Karl Kleinpaste - bash_spell_correct_word: bound to "C-x s" by default in emacs mode lib/readline/display.c - rl_redisplay: fix redisplay problem that occurs when switching from the rl-digit-argument prompt "(arg: N)" back to the regular prompt, and the regular prompt contains invisible characters doc/bash.1,lib/readline/doc/rluser.texi - spell-correct-word: document new function and its default binding 3/17 ---- doc/{bash.1,bashref.texi} - cd: slight changes to specify that it sets PWD and OLDPWD - {pushd,popd}: make it clear that these builtins use cd to change the current working directory; change wording to simplify the description and clarify the exit status 3/18 ---- execute_cmd.c - execute_disk_command: after performing redirections, call unlink_all_fifos() to remove the FIFOs created as part of expanding redirections. They should have been opened by then, and we're going to call shell_execve right away anyway, so we won't be around to remove the FIFOs. From a report from Michael Felt 3/22 ---- parse.y - alias_expand_token: slight tweak to check for alias expansion: perform expansion unconditionally if PST_ALEXPNEXT is set, and disable it in case statement pattern lists if the previous token indicates a command name is acceptable. From a report by Oguz config-bot.h - HAVE_MKDTEMP: fix typo 3/25 ---- lib/readline/terminal.c - look in terminfo for key sequences for page up (kP) and page down (kN) and bind them to history-search-{backward,forward}, respectively. From a patch from Xose Vazquez Perez 3/30 ---- doc/bashref.texi - expand the node describing $"..." string translation with additional details and examples 3/31 ---- misc.c - rl_fetch_history: moved here from vi_mode.c - rl_fetch_history: negative arguments count back from the end of the history, instead of taking you to the beginning of the history list - rl_fetch_history: in vi mode, an out-of-range argument rings the bell and doesn't change the line vi_mode.c - rl_vi_fetch_history: call rl_fetch_history readline.h - rl_fetch_history: new extern declaration doc/bash.1,lib/readline/doc/{readline.3,rluser.texi} - rl_fetch_history: add description builtins/setattr.def - show_var_attributes: if a variable's value indicates that it should be ANSI-C quoted, use ansic_quote instead of sh_double_quote to format the value string. From proposal by Greg Wooledge 4/5 --- arrayfunc.c - unbind_array_element: if FLAGS includes VA_ONEWORD, don't use skipsubscript to parse the subscript, just assume the entire SUB is the subcript and that it contains the closing `]', so we just want everything in SUB except the last character. parse.y: - select_command: use compound_list instead of list, like for_command. Report by Greywolf - list: move this into compound_list (replacing the instance of `list' in the compound_list production), remove from the grammar 4/6 --- arrayfunc.c - unbind_array_element: use VA_NOEXPAND instead of literal 1 4/7 --- lib/readline/funmap.c - default_funmap: add missing `vi-undo' to the list of vi-mode bindable functions. Reported by Xirui Zhao 4/8 --- config-top.h - DEFAULT_LOADABLE_BUILTINS_PATH: default value for BASH_LOADABLES_PATH doc/{bash.1,bashref.texi} - enable: note that it uses $BASH_LOADABLES_PATH, and that the default is system-dependent variables.c - initialize_shell_variables: initialize BASH_LOADABLES_PATH to the default given in DEFAULT_LOADABLE_BUILTINS_PATH 4/12 ---- doc/{bash.1,bashref.texi} - add link to git master tar file as a place to get the current version 4/14 ---- bashline.c - attempt_shell_completion: use -1 as a sentinel value for in_command_position indicating that we cannot be in a command position (e.g., because we're the target of a redirection) and should not check for a programmable command completion or tell the programmable completion code to use command completion. Report and fix from Marc Aurèle La France 4/16 ---- builtins/bind.def - bind_builtin: reverse sense of strvec_search return value when deciding whether or not to remove a unix-command binding from the cmd keymap. Bug report by Dale Sedivec lib/readline/doc/rltech.texi - RL_PROMPT_{START,END}_IGNORE: document current values of \001 and \002. Report from Mingye Wang 4/19 ---- arrayfunc.c - assign_assoc_from_kvlist: fix memory leak reported by konsolebox 4/20 ---- command.h,subst.c - W_ITILDE: remove, replace with a variable since it's only used inside a single call to expand_word_internal 4/21 ---- {subst.c,make_cmd.c,parse.y} - W_DQUOTE: no longer used, use W_NOPROCSUB and W_NOTILDE directly (for arithmetic commands and words in arithmetic for commands) 4/24 ---- bashline.c - executable_completion: since this function gets an unquoted filename from rl_filename_completion_function, we need to quote special characters before passing it to bash_directory_completion_hook. Report from Alex fxmbsw7 Ratchev 4/26 ---- lib/readline/search.c - _rl_nsearch_abort: move function calls around so _rl_restore_prompt happens before rl_clear_message, like when aborting an incremental search. Suggested by sparrowhawk996@gmail.com subst.h - ASS_ALLOWALLSUB: new assignment flag value, means to allow @ and * as valid array subscripts when assigning to existing associative arrays arrayfunc.c - assign_array_element: allow assignment of key `@' to an existing associative array if the caller passes ASS_ALLOWALLSUB - assign_compound_array_list: allow ( [@]=value ) to an existing associative array builtins/declare.def - declare_internal: allow assignment of key `@' to an existing associative array by passing ASS_ALLOWALLSUB to assign_array_element as part of local_aflags. This affects declare, local, and typeset subst.c - do_assignment_internal: allow a[@]=value to an existing associative array by passing ASS_ALLOWALLSUB to assign_array_element 4/27 ---- builtins/common.[ch] - builtin_bind_var_to_int: wrapper for bind_var_to_int to be used by builtin commands; placeholder for future work builtins/wait.def - builtin_bind_var_to_int: use instead of bind_var_to_int builtins/common.c - builtin_bind_variable: allow assignment of key `@' to an existing associative array by passing ASS_ALLOWALLSUB to assign_array_element. This affects printf and read builtins/variables.[ch] - bind_var_to_int: add third `flags' argument to pass to bind_variable instead of always passing 0 redir.c,builtins/common.c,builtins/printf.def - bind_var_to_int: change callers, add third flags argument builtins/common.c - builtin_bind_var_to_int: pass ASS_ALLOWALLSUB to bind_var_to_int so builtins like wait can assign to assoc[@] and assoc[*] 4/28 ---- bashline.c - command_word_completion_function: make sure to initialize old_glob_ignore_case before trying to restore from it - command_word_completion_function: if we are completing a glob pattern, make sure to set rl_filename_completion_desired, so we get quoting and appending -- we are completing a filename, after all. From a report from Manuel Boni lib/readline/bind.c - enable-active-region: separate control of the active region and bracketed paste. Still set to the same default value as bracketed paste, and enabling bracketed paste enables the active region. Now you can enable bracketed paste and then turn off the active region. doc/bash.1,lib/readline/doc/{readline.3,rltech.texi} - enable-active-region: document new bindable readline variable and its effects 4/30 ---- command.h - W_ARRAYREF: new flag, meaning the word is a valid array reference with subscript, replaces W_DOLLARSTAR, which was unused subst.c - expand_subscript_string,expand_array_subscript: new functions to parse and expand-and-quote array subscripts. For future use 5/3 --- builtins/mapfile.def - mapfile: if the delimiter is a newline, set unbuffered_read = 1 for any file descriptor that isn't seekable and lseek sets errno to ESPIPE (pipes, FIFOs, maybe terminal devices). If it's not a newline, only allow buffered reads if the file descriptor is a regular file. Report from Koichi Murase builtins/read.def - read_builtin: only set unbuffered_read = 1 if the input is coming from a pipe (which we can't seek on) or the input is a terminal and we want to read a specified number of characters or we're using a non-standard delimiter 5/4 --- builtins/mapfile.def - mapfile: call zsyncfd before calling the callback. Suggested by Koichi Murase ; we'll see how it goes 5/5 --- subst.h - expand_subscript_string: extern declaration {arrayfunc,subst}.c - expand_subscript_string: replace expand_assignment_string_to_string with calls to this when expanding array subscripts subst.c - cond_expand_word: call expand_word_internal with Q_ARITH if `special' says we should quote for an arithmetic expression context - expand_word_internal: call expand_array_subscript when we see `[' in arithmetic or array subscript contexts, make conditional on the compatibility level later - expand_word_internal: make sure W_ARRAYREF makes it through this function and into the returned word 5/6 --- arrayfunc.c - array_expand_index: call evalexp with a flag of 0 since we call expand_arith_string with Q_ARITH and we want evalexp to remove the quotes execute_cmd.c - eval_arith_for_expr,execute_arith_command: now that Q_ARITH has an effect on array subscripts (it quotes the special expansion characters), call evalexp with 0 as the flags arg so quote removal is performed on this quoted argument. Make this conditional on the shell compatibility level later - execute_cond_command: make sure to expand the argument to -v by calling cond_expand_node with Q_ARITH, and correspondingly turn off assoc_expand_once when calling unary_test with that argument, since we want it to be expanded again to remove the quotes - execute_cond_command: expand the arguments to the arithmetic operators with Q_ARITH and pass TEST_ARITHEXP to binary_test to ensure that it lets evalexp expand the arguments to remove the quoting test.c - arithcomp: if TEST_ARITHEXP is in FLAGS, call evalexp with a flag if 0 to force evalexp to expand the arguments to remove the quoting subst.c - param_expand: since we call expand_arith_string with Q_ARITH, we need to call evalexp with 0 instead of EXP_EXPANDED for $((...)) expansion - expand_word_internal: if we recursively call expand_word_internal to expand the contents of a double-quoted string, make sure we pass through Q_ARITH if it appears in QUOTED - verify_substring_values: call expand_arith_string with Q_ARITH and correspondingly call evalexp with 0 instead of EXP_EXPANDED execute_cmd.c - execute_cond_node: if -v is the operator, and the operand is a valid array reference, pass TEST_ARRAYEXP flag to unary_test 5/7 --- builtins/common.[ch] - set_expand_once: set array_expand_once to the value passed as the first argument, returning the original value builtins.h - ARRAYREF_BUILTIN: new flag for shell builtins, means that the builtin can take array references, with subscripts, as arguments builtins/mkbuiltins.c - set ARRAYREF_BUILTIN flag on builtins given in the arrayvar_builtins array execute_cmd.c - execute_cond_node: use set_expand_once to set array_expand_once to 0 before calling unary_test with -v (see change from 5/6) arrayfunc.c - unbind_array_element: allow the caller to choose whether or not a subscript of `*' or `@' unsets the entire array by passing VA_ALLOWALL in FLAGS. Right now, since the unset builtin doesn't pass VA_ALLOWALL, those subscripts unset individual elements for associative arrays. We preserve the old behavior of unsetting indexed arrays for the time being with new indexed-array-specific code 5/9 --- builtins/shopt.def - expand_once_flag: "assoc_expand_once" option now sets this flag, calls set_assoc_expand on change - set_assoc_expand: sets assoc_expand_once to mirror expand_once_flag; placeholder for future changes builtins/common.h - expand_once_flag: extern declaration 5/10 ---- doc/{bash.1,bashref.texi} - note that case patterns undergo quote removal. Reported by AlvinSeville7cf 5/11 ---- builtins/bashgetopt.c - list_optflags: flags associated with the word corresponding to list_optarg, assuming list_optarg is a separate argument builtins/bashgetopt.h - list_optflags: extern declaration builtins/set.def - unset_builtin: set VFLAGS each time through the loop, since we take whether or not the word has W_ARRAYREF set into account - unset_builtin: don't pass VA_ALLOWALL to unbind_array_element for now test.c - unary_test: in the -v case, use assoc_expand_once in the first call to valid_array_reference () builtins/printf.def - printf_builtin: only set VA_ONEWORD if the option argument to -v has W_ARRAYREF set (look at list_optflags) 5/12 ---- subst.c - expand_array_subscript: don't quote @ or * in the expanded subscript at all, even when they are the only character in the subscript. See how this works out -- it might uncover places where we need to allow `*' and `@' as subscripts where they are not allowed now expr.c - expr_bind_variable: pass ASS_ALLOWALLSUB to bind_int_variable so we can allow (( A[@]=value )) when A is an existing associative array arrayfunc.h - AV_ATSTARKEYS: new flag value, means to accept a[@] and a[*] but treat them as keys/expressions and not special values arrayfunc.c - array_value_internal: check AV_ATSTARKEYS and don't treat them as ALL_ELEMENT_SUB values; they fall through to use as keys/indices test.c - unary_test: if -v is passed an array reference, add AV_ATSTARKEYS to the flags passed to array_value so we treat @ and * as keys for an existing associative array 5/13 ---- subst.c - expand_cond_word: if SPECIAL == 3 (arithmetic expression), dequote the resulting WORD_LIST * as if special == 0, because we don't want to quote the list for pattern matching. Report from Adjudicator Darren 5/14 ---- subst.c - expand_array_subscript: add double quote (") to the list of characters that are backslash-quoted in subscripts after word expansion. skipsubscript treats them specially, so you have to quote them to do things like `key='"' ; array[$key]=1 ; [[ -v array[$key] ]]' 5/16 ---- builtins/wait.def - wait_builtin: if we longjmp to wait_intr_buf, call unset_waitlist if we have called set_waitlist (wflags & JWAIT_WAITING). Fixes bug with wait -n interrupted by a trapped signal (not SIGINT) reported by Jonas Alfredsson jobs.c - wait_sigint_cleanup: restore the old sigint handler before we longjmp out by calling restore_sigint_handler() 5/17 ---- builtins/read.def - bind_read_variable: now takes an additional argument, flags to pass to builtin_bind_variable; change callers - SET_VFLAGS: set vflags and bindflags - read_builtin: call SET_VFLAGS to set vflags and bindflags from each word before calling valid_array_reference and bind_read_variable builtins/common.c - builtin_bind_variable: set vflags (for valid_array_reference) and bindflags (for bind_variable/assign_array_element) separately for clarity arrayfunc.c - assign_array_element: sanity check: make sure that the subscript returned by array_variable_name consumes the entire NAME, otherwise flag it as a subscript error. This keeps things like `KEY=' ]'; read assoc[$KEY] <<< hello' from assigning to incomplete subscripts builtins/printf.def - printf_builtin: if LIST_OPTFLAGS includes W_ARRAYREF, set VA_NOEXPAND in VFLAGS 5/17 ---- lib/readline/complete.c - compute_lcd_of_matches: move a couple of strlen calls out of a loop in calls to mbrtowc; performance improvement only. Report and fix from sparrowhawk996@gmail.com lib/readline/bind.c - rl_trim_arg_from_keyseq: take a key sequence and its length and return the index into the key sequence following any initial numeric argument. Return -1 if there is no numeric argument (the caller is expected to make sure) or if the key sequence consists *only* of the numeric argument. The caller should use the remainder of the key sequence to look up the desired key binding. lib/readline/readline.h - rl_trim_arg_from_keyseq: extern declaration bashline.c - bash_execute_unix_command: if the argument count is > 1 or we have an explicit argument, call rl_trim_arg_from_keyseq to get past the numeric argument and deal with the rest of the key sequence. We still need a way to pass it to the invoked program or function. From a report from Jesper Nygards 5/18 ---- bashline.c - bash_execute_unix_command: if the user supplied a numeric argument when invoking bash_execute_unix_command, pass it to the command in the READLINE_ARGUMENT variable lib/readline/readline.[ch] - _rl_del_executing_keyseq: convenience function to `delete' the last character added to the executing key sequence. Intended to be used before calling rl_execute_next or similar functions that push input back to be re-read doc/{bash.1,bashref.texi} - READLINE_ARGUMENT: documented new variable available for commands defined using `bind -x' keybindings lib/readline/doc/rltech.texi - rl_trim_arg_from_keyseq: documented new function 5/19 ---- builtins/evalstring.c - should_suppress_fork: suppress the fork if we're in a process substitution subshell, in addition to being a simple command without redirections. From a report back in 10/2020 from Hyunho Cho bashline.c - command_word_completion_function: if we're trying to complete an absolute program (one containing a slash), don't run strcmp or strcasecmp on the return value from rl_filename_completion_function, since that duplicates work the filename completion function already does. From a report back in 1/2021 by awa54@cdak.net 5/22 ---- parse.y - CHECK_FOR_RESERVED_WORD: if we are returning an ESAC and unsetting PST_CASESTMT, decrement esacs_needed_count parse.y,shell.h - sh_parser_state_t: save and restore esacs_needed_count and expecting_in_token in the shell parser state struct and save_parser_state/restore_parser_state print_cmd.c - print_simple_command: don't bother to call command_print_word_list with an empty list - print_simple_command: don't print a space before a redirection list if there weren't any command words to print 5/24 ---- lib/sh/input_avail.c - nchars_avail: make sure SET and OSET are declared on systems with select(2). Reported by Larkin Nickle parse.y - cond_term: if we read a `!' toggle CMD_INVERT_RETURN instead of setting it unconditionally. Report and patch from Vincent Menegaux via https://savannah.gnu.org/patch/?10070 5/25 ---- doc/{bash.1,bashref.texi} - test: add the ( $2 $3 ) case to the description of the four-argument behavior. Inspired by a discussion with Christoph Anton Mitterer 5/27 ---- doc/bashref.texi - replace most of the GNU parallel section with a reference to the tutorial on gnu.org lib/glob/glob.h - GX_NEGATE: new flag; indicates whether the pattern is being negated as part of an extglob pattern. Not used yet lib/glob/glob.c - extglob_skipname,wextglob_skipname: pass GX_NEGATE to the skipname functions if the pattern is being negated. Not checked yet 5/28 ---- doc/{bash.1,bashref.texi} - dotglob: add text to clarify that `.' and `..' have to be matched by a pattern beginning with `.' or -- and this is the sketchy part -- that a pattern beginning with `.' has to be one of the patterns in an extended glob expression lib/glob/glob.c - skipname,wskipname: perform the special checks for `.' only if the pattern is not being negated 6/3 --- parse.y,shell.h - eol_ungetc_lookahead: save and restore as part of the parser state included in {save,restore}_parser_state 6/7 --- lib/readline/display.c - puts_face: make sure to cast each member of STR to unsigned char, so it's not misinterpreted as EOF, which putc_face does not display. Report and fix from Volodymyr Prodan in https://savannah.gnu.org/patch/?10076 examples/shobj-conf - aix*gcc: change SHOBJ_LD to ${CC}, prefix the SHOBJ_LDFLAGS flags with -Wl, so gcc will pass them to the linker. Report from lehel@maxer.hu in https://savannah.gnu.org/support/?110505 6/11 ---- doc/{bash.1,bashref.texi} - cd: additional arguments are not ignored; they raise an error. Report from Douglas McIlroy lib/glob/strmatch.h - FNM_DOTDOT: enable special handling for `.' and `..': if FNM_PERIOD is not set, `.' and `..' at the start of a string or as a pathname component need to be matched by a `.' in the pattern and cannot be matched by `?', `*', or a bracket expression lib/glob/glob.c - glob_vector: pass FNM_DOTDOT to strmatch() if noglob_dot_filenames is not set to enable special handling of `.' and `..'. Prompted by a discussion with Nora Platiel - skipname,wskipname: remove special checks for `.' and (disabled) checks for `..' lib/glob/sm_loop.c - GMATCH: implement special handling for FNM_DOTDOT and `.' and `..': make sure they can't be matched by any special glob characters at the start of the string or as a pathname component (if FNM_PATHNAME is set). This also means that !(pattern) won't return `.' or `..' if dotglob is set - GMATCH,EXTMATCH: don't pass FNM_DOTDOT down to recursive calls, like FNM_PERIOD, once we process the first character in the string or pathname component lib/glob/smatch.c - ISDIRSEP,PATHSEP,SDOT_OR_DOTDOT,PDOT_OR_DOTDOT: provide definitions for single-byte or wide character strings for sm_loop.c to use 6/16 ---- doc/{bash.1,bashref.texi},lib/readline/doc/{rluser,rltech}.texi - slight wording changes to rewrite gender-specific language. From a report by Vipul Kumar , suggested language from G. Branden Robinson , Lawrence Velázquez , and Andrew Church builtins/fc.def - fc_builtin: make sure an entry in the history list is non-null before trying to write it to stdout or the temporary file. From a report from Sibo Dong builtins/common.[ch] - builtin_arrayref_flags: given a WORD_DESC * (which includes flags) and a base set of flags, return a set of flags to pass to valid_array_reference and similar functions builtins/set.def - unset_builtin: call builtin_arrayref_flags to set vflags 6/19 ---- parse.y - parse_comsub: rewrite to recursively call the parser (yyparse()) and return a string constructed from the resulting parse tree. Probably will only work with bison. Error reporting is more accurate about line numbers and invalid tokens, and command substitution errors are caught earlier, before expansion - DOLPAREN: new token, never created by yylex; only ever set by parse_comsub and xparse_dolparen to indicate we're recursively calling the parser for a command substitution - comsub: new grammar production that's triggered by DOLPAREN and parses a command substitution, returning a . It's one of the possible end states for the top-level parser - grammar: only call rewind_input_string if the shell's input is coming from a string - shell_ungets: push a string back onto the shell input; only used by make_here_document for backwards compatibility -- allowing a here- document to be terminated by a token at the end of a command substitution - yylex: don't need any more special handling when returning shell_eof_token, but we keep the clause for future work - read_token_word: don't handle backslashes (leave them in the input stream) if we are reading a command substitution (PST_NOEXPAND) - reserved_word_acceptable: allow reserved words after DOLPAREN - report_syntax_error: better error handling if we hit EOF while looking for the ending right paren in a command substitution - parse_string_to_word_list,parse_compound_assignment: make sure to turn off parse_comsub sentinel temporarily make_cmd.c - make_here_document: backwards compatibility: if we end a here document on the same line as the end of a command substitution, allow the token to terminate the here document (without requiring a newline) and push the remainder of the line back for the parser to consume and terminate the command substitution builtins/evalstring.c - parse_string: if we read shell_eof_token and use it to terminate a command, rewind the input string here before returning, instead of guessing where to rewind it in the caller 6/22 ---- redir.c - do_redirection_internal: if VARASSIGN_REDIR_AUTOCLOSE is non-zero, add a redirect to automatically close {var} parser.h - PST_NOERROR: don't print error messages in yyerror(), just reset the parser and return parse.y - yyerror: if parser_state & PST_NOERROR, don't print an error message - xparse_dolparen: if the flags includes SX_COMPLETE, add PST_NOERROR to parser_state, to inhibit error messages. Fixes bug with adding lines with incomplete command substitutions to the history list reported by Koichi Murase subst.c - skip_matched_pair: make sure to pass the SX_COMPLETE flag to extract_delimited_string 6/28 ---- lib/readline/bind.c - rl_trim_arg_from_keyseq: handle rl_vi_arg_digit if VI_MODE is defined. Report and fix from Koichi Murase - rl_trim_arg_from_keyseq: rework to handle case where the digit argument is given as a discrete sequence of multiple rl_digit_argument commands (e.g, "M-1 M-2" instead of "M-1 2"). Report and fix from Koichi Murase subst.c - pat_subst: avoid calling RESIZE_MALLOCED_BUFFER with STRLEN(s) as an argument, since it ends up going into a loop. Report and fix from Koichi Murase lib/sh/casemod.c - cval: take the string length as an argument, to avoid having to call strlen every time (in HANDLE_MULTIBYTE mode). The caller already has this info. Report and fix from Koichi Murase builtins/complete.def - print_cmd_name: single-quote the command name if it contains any shell metacharacters - print_one_completion: call print_arg with a second arg of 1 if the function name contains shell metacharacters. Report and fix from Koichi Murase 6/30 ---- configure.ac - substitute BASE_CFLAGS_FOR_BUILD and STYLE_CFLAGS into Makefile targets - take STYLE_CFLAGS out of the CFLAGS_FOR_BUILD variable that gets substituted - remove STYLE_CFLAGS from CFLAGS - use `:+' expansion to set variables based on whether $GCC is set or null, since configure can do GCC= support/Makefile.in - man2html: remove CCFLAGS_FOR_BUILD from the recipe, leaving only LDFLAGS_FOR_BUILD. Report from Jay K - gen-helpfiles: remove CCFLAGS_FOR_BUILD from the recipe, leaving only LDFLAGS_FOR_BUILD - man2html: use $(STYLE_CFLAGS) in BASE_CCFLAGS variable for C file compilation options {,builtins}/Makefile.in, lib/{sh,readline,malloc,glob}/Makefile.in - use STYLE_CFLAGS so specifying CFLAGS=-g to make doesn't clutter the output with warnings about parens and format strings 7/9 --- lib/readline/search.c - make_history_line_current: call _rl_free_saved_history_line to clean up _rl_saved_line_from_history and get all the code that frees it into one place lib/readline/misc.c - _rl_free_saved_history_line: if rl_undo_list points to the data member of _rl_saved_line_from_history, set it to NULL to avoid having it point to freed memory, since the next thing we do now is to free the undo list the data member points to - _rl_start_using_history: call _rl_free_saved_history_line instead of calling _rl_free_history_entry directly. Fixes memory leak reported by Trung Dam 7/12 ---- lib/readline/search.c - make_history_line_current: free rl_undo_list before replacing the current line with the line from history, since it is a private undo list from reading the search string lib/readline/rlmbutil.h - Since wchar_t/mbrtowc/wcrtomb are limited to 16 bits on Windows with MSVC, start abstracting the differences using WCHAR_T/ MBRTOWC/WCRTOMB 7/13 ---- lib/readline/{complete,display,input,text,util,vi_mode}.c - use WCHAR_T/MBRTOWC/WCRTOMB. Part of a set of Windows MSVC fixes from sparrowhawk996@gmail.com builtins/{enable,hash,help}.def - enable_builtin: use sh_chkwrite after output to check for write errors 7/16 ---- arrayfunc.c - quote_compound_array_word: free SUB and VALUE after assigning from sh_single_quote(). From a coverity report from Siteshwar Vashisht bashhist.c - bash_remove_history_range: free DISCARD_LIST after freeing its elements bashline.c - bash_directory_expansion: add code to free D as a separate branch, though it's never hit in practice builtins/trap.def - showtrap: free T even if show_default == 1 if it's a non-default trap string 7/17 ---- execute_cmd.c - execute_coproc: free NAME on invalid identifier error lib/glob/glob.c - glob_vector: make sure NEXTLINK is allocated using malloc before passing it to free() - glob_filename: free RESULT before returning glob_error_return when there is only a filename print_cmd.c - indirection_level_string: make sure we free PS4 after calling decode_prompt_string if *ps4 == 0 subst.c - parameter_brace_transform: if vtype == VT_VARIABLE, we need to free a non-null VAL variables.c - assign_in_env: if NAME is not a valid shell identifier, free it after printing the error message and before returning. These are the rest of the fixes from Siteshwar Vashisht 7/22 ---- shell.c - main: set dollar_vars[0] to shell_script_filename before calling run_startup_files() in the non-interactive case. Restore it after run_startup_files returns so we can get better error messages if we can't open a script file. Suggested by several people, originally by Marc Aurèle La France back in 2/2021 (in a different form) and most recently by Tapani Tarvainen 7/28 ---- trap.c - any_signals_trapped: return that a signal is trapped only if it's not ignored. This is an additional opportunity for optimization, reported in https://bugzilla.redhat.com/show_bug.cgi?id=1981926 7/30 ---- examples/loadables/sleep.c - main: if the uconvert conversion fails, but the argument appears to contain a GNU-like interval specifier like "1m30s", return EX_DISKFALLBACK so the execute_builtin code tries to run the external version builtins/enable.def - enable_shell_builtin: if the builtin isn't found, return EX_NOTFOUND to allow the caller (enable_builtin) to react differently if it wants to - dyn_load_builtin: if the shared object isn't found, return EX_NOTFOUND, change enable_builtin to deal with it - enable_builtin: if there are no supplied options, and we attempt to enable a non-existent builtin, try modifying `enable name' to the equivalent of `enable -f name name' and return success if we successfully load a builtin from a shared object. Proposed several times, most recently by Robert Elz doc/{bash.1,bashref.texi} - enable: document new behavior of `enable NAME' when NAME is not a current shell builtin 8/3 --- lib/glob/sm_loop.c - GMATCH: check for interrupts or terminating signals each time through the loop and return FNM_NOMATCH immediately if received. Let the higher layers deal with interrupting the match and jumping out. Inspired by a report from andrej@podzimek.org 8/6 --- subst.c - {parameter_brace_remove_pattern,parameter_brace_transform, parameter_brace_substring,parameter_brace_casemod, parameter_brace_patsub}: make sure the IND argument is of type arrayind_t to handle the full range of subscripts. Reported by felix@f-hauri.ch builtins/printf.def - printf_builtin: take new format specifier: %Q. This acts like %q but applies any supplied precision to the original unquoted argument, then quotes that result and outputs the entire quoted string, preserving the original field width. Originally suggested back in 4/2020 by Sam Liddicott subst.c - char_is_quoted: check whether or not we are on the second or later line of a command with an unclosed quoted string from a previous line. If we are, see if it's closed before the character we're checking, so we don't interpret a closing quote as starting a new quoted string. Reported several times, most recently by Phi Debian in 6/2021. 8/9 --- parse.y,subst.c - locale_expand: if the variable SINGLEQUOTE_TRANSLATIONS is non-zero, single-quote the translated result of $"..." (if it's different from the untranslated string) builtins/shopt.def - noexpand_translation: new option to expose the value of SINGLEQUOTE_TRANSLATIONS 8/10 ---- doc/{bash.1,bashref.texi} - noexpand_translation: add description of new option 8/16 ---- builtins/printf.def - printf_builtin: initialize retval after parsing arguments, since we use it for the `v' option. Bug report from Keith Thompson lib/sh/unicode.c - u32tocesc: fix typo that returned \uXXXXXXXX instead of \UXXXXXXXX. From https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=992257 by 8/17 ---- siglist.c - include command.h before general.h for PARAMS and prototypes. Report from Osipov, Michael (LDA IT PLM) 8/18 ---- lib/readline/colors.c - S_ISBLK: make sure it's defined before we use it, like in complete.c lib/readline/complete.c,{bashline,variables}.c - minor changes to prep for making rl_completer_word_break_characters `const' subst.[ch],pcomplete.c - split_at_delims: the DELIMS argument is now `const'; change callers lib/readline/rlmbutil.h - _rl_wcwidth: add function prototype for inline function declaration lib/readline/bind.c - _rl_get_keyname: print keys greater than 160 (which are valid UTF-8) as octal escape sequences lib/readline/text.c - do_lowercase_version: return 99999 to prevent the linker from combining it with _rl_null_function and optimizing away the separate copy. That messes with function pointer comparisons. Part of this batch of fixes from sparrowhawk996@gmail.com 8/19 ---- complete.c,readline.c,readline.h - rl_completer_word_break_characters: now const char * like rl_basic_word_break_characters; element of readline state struct used to save it also const. THIS IS AN API CHANGE bashline.c - orig_rl_completer_word_break_characters: now const char * like rl_completer_word_break_characters 8/20 ---- bashline.c - bash_directory_completion_hook: if direxpand and dirspell are both set while trying to complete an absolute pathname as a command, don't take a spell-corrected directory name that is shorter than the original hint. https://bugzilla.redhat.com/show_bug.cgi?id=1782809 builtins/common.[ch] - sh_noassign: convenience function to print an error message when a user attempts an assignment to a "noassign" variable. Not used yet assoc.[ch] - assoc_to_kvpair_list: new function, returns a WORD_LIST containing key-value pairs as separate words array.[ch] - array_to_kvpair_list: new function, returns a WORD_LIST containing index-value pairs as separate words subst.c - string_transform: handle '@k' transform like '@K' - valid_parameter_transform: 'k' is a valid transform operator - array_transform: handle '@k' transformation by calling one of array_to_kvpair_list or assoc_to_kvpair_list and treating the result as if expanding the array depending on whether the index is `@' or `*' with the usual meanings doc/{bash.1,bashref.texi} - document new '@k' parameter transformation operator 8/27 ---- lib/readline/kill.c - rl_unix_filename_rubout: handle pathnames that consist only of one or more slashes. The old code went too far and deleted the previous word as well. From dabe@dabe.com 8/31 ---- parse.y - STRING_SAVER: now save and restore shell_input_line_len; not sure why it wasn't done before; fix push_string and pop_string accordingly - prompt_again: now takes a parameter FORCE; not used yet (every caller passes 0), needs more thought builtins/evalstring.h - open_redir_file: broke code that expands the redirection and opens the resultant filename into a new function, called from cat_file redir.h - open_redir_file: extern declaration here for now builtins/evalstring.c - parse_string: takes a new argument: COMMAND **cmdp; if non-null, saves the parsed command to *cmdp and lets the caller manage it itself. global_command is still not modified. Changed callers in parse.y 9/1 --- parse.y - parse_string_to_command: stripped-down version of xparse_dolparen that takes a string, runs it through the parser, and returns the resultant COMMAND *; uses parse_string with the new argument externs.h - parse_string_to_command: extern declaration builtins/evalstring.c - can_optimize_cat_file: new function, takes a COMMAND * argument and returns true if the command can be optimized like $(&", see if we can optimize the command and call optimize_cat_file to do it if we can. 9/2 --- configure.ac,config.h.in,lib/sh/setlinebuf.c - SETVBUF_REVERSED: remove all references configure.ac,aclocal.m4 - AC_OUTPUT: move created files and commands to AC_CONFIG_FILES and AC_CONFIG_COMMANDS; call AC_OUTPUT without any arguments - require autoconf version 2.63 - AC_HELP_STRING -> AS_HELP_STRING - replace references to $ac_cv_prog_gcc with $GCC - AC_C_LONG_DOUBLE -> AC_TYPE_LONG_DOUBLE - enable-alt-array-implementation: new configure option, placeholder for later unwind_prot.c - stddef.h: include if HAVE_STDDEF_H is defined, not STDC_HEADERS bashansi.h - memory.h: don't include any more; assume has all the necessary defines aclocal.m4 - AC_HEADER_STDC: remove AC_REQUIRE calls to this; change tests to use appropriate checks instead of STDC_HEADERS 9/3 --- configure.ac,config.h.in - AC_TYPE_SIGNAL,BASH_TYPE_SIGHANDLER: remove calls, remove mention of RETSIGTYPE and VOID_SIGHANDLER - AC_USE_SYSTEM_EXTENSIONS: use this instead of AC_ISC_POSIX (which is well and truly obsolete) and AC_MINIX (which just calls it anyway). All the right defines are already present - AC_HEADER_TIME: removed - AC_HEADER_STDC: removed - AC_DECL_SYS_SIGLIST: replaced with BASH_DECL_SYS_SIGLIST - BASH_FUNC_CTYPE_NONASCII: removed sig.h,lib/readline/signals.c - RETSIGTYPE -> void, SIGHANDLER_RETURN -> return include/posixtime.h - don't bother with TIME_WITH_SYS_TIME, just include if we have it and include unconditionally include/chartypes.h,lib/readline/chardefs.h - IN_CTYPE_DOMAIN: no longer uses STDC_HEADERS define, checks against CHAR_MAX for non-C89 systems, still a no-op on C89 systems lib/readline/chardefs.h - memory.h: don't include, no need for STDC_HEADERS; assume the mem* functions are defined in as in C89 - now that we assume IN_CTYPE_DOMAIN bounds its argument to unsigned chars, we don't need NON_NEGATIVE checks at all lib/readline/chardefs.h,util.c - ALPHABETIC -> _rl_alphabetic_p 9/8 --- lib/sh/strftime.c - include after reworking above; include unconditionally in case it defines time_t. From a report by Ori Sky Farrell 9/9 --- lib/readline/search.c - rl_history_search_internal: set rl_undo_list to NULL after calling rl_maybe_save_line, since it will be restored one way or another after the search completes, and we don't want it to be freed twice - rl_history_search_internal: leave the current history offset at the position of the last matching history entry. This means that things like ^P will start back from that history entry, ^N will move forward from there, and operate-and-get-next will work like it does with incremental searches. Reported by Vandrus Zoltán 9/10 ---- aclocal.m4 - BASH_C_LONG_LONG: removed - BASH_TYPE_LONG_DOUBLE: removed - BASH_TYPE_LONG_LONG,BASH_TYPE_UNSIGNED_LONG_LONG: removed - BASH_FUNC_CTYPE_NONASCII: removed - BASH_SYS_SIGNAL_VINTAGE,BASH_SYS_REINSTALL_SIGHANDLERS: don't require AC_TYPE_SIGNAL, use void instead of RETSIGTYPE - BASH_TYPE_SIGHANDLER: removed configure.ac,aclocal.m4,config.h.in - BASH_TYPE_LONG_LONG: removed, call AC_TYPE_LONG_LONG_INT directly, change #define to HAVE_LONG_LONG_INT - BASH_TYPE_UNSIGNED_LONG_LONG: removed, call AC_TYPE_UNSIGNED_LONG_LONG_INT directly, change #define to HAVE_UNSIGNED_LONG_LONG_INT externs.h,include/typemax.h,lib/sh/{fmtullong,snprintf,strtoimax,strtoll, strtoull,strtoumax}.c - HAVE_LONG_LONG -> HAVE_LONG_LONG_INT - HAVE_UNSIGNED_LONG_LONG -> HAVE_UNSIGNED_LONG_LONG_INT configure.ac - AC_TRY_COMPILE -> AC_COMPILE_IFELSE - BASH_CHECK_TYPE (intmax_t) -> AC_TYPE_INTMAX_T - BASH_CHECK_TYPE (uintmax_t) -> AC_TYPE_UINTMAX_T - BASH_CHECK_TYPE (sig_atomic_t): removed aclocal.m4 - BASH_FUNC_STRSIGNAL: changed to use AC_LINK_IFELSE - BASH_FUNC_LSTAT: changed to use AC_LINK_IFELSE - BASH_FUNC_SBRK: changed to use AC_LINK_IFELSE and AC_RUN_IFELSE, fixed typo - BASH_CHECK_SPEED_T: changed to use AC_COMPILE_IFELSE - BASH_TYPE_SIG_ATOMIC_T: use BASH_CHECK_TYPE instead of AC_CHECK_TYPE - BASH_STRUCT_DIRENT: new macro, like _AC_STRUCT_DIRENT but public and sets a different bash-specific shell variable - BASH_STRUCT_DIRENT_D_{INO,FILENO,NAMLEN}: call BASH_STRUCT_DIRENT with a different first argument instead of using inline code and AC_TRY_COMPILE 9/11 ---- [prayers for the victims of 9/11/2001] aclocal.m4 - BASH_CHECK_DECL: just call AC_CHECK_DECLS and use the default includes, which includes the files we used to specify - BASH_FUNC_INET_ATON: changed to use AC_LINK_IFELSE - BASH_STRUCT_WEXITSTATUS_OFFSET: changed to use AC_RUN_IFELSE - BASH_FUNC_OPENDIR_CHECK: changed to use AC_RUN_IFELSE - BASH_FUNC_ULIMIT_MAXFDS: changed to use AC_RUN_IFELSE - BASH_UNDER_SYS_SIGLIST: changed to use AC_RUN_IFELSE - BASH_SYS_SIGLIST: changed to use AC_RUN_IFELSE - BASH_FUNC_DUP2_CLOEXEC_CHECK: changed to use AC_RUN_IFELSE - BASH_FUNC_GETENV: changed to use AC_RUN_IFELSE - BASH_FUNC_GETCWD: changed to use AC_RUN_IFELSE - BASH_FUNC_FNMATCH_EXTMATCH: changed to use AC_RUN_IFELSE - BASH_FUNC_POSIX_SETJMP: changed to use AC_RUN_IFELSE - BASH_FUNC_STRCOLL: changed to use AC_RUN_IFELSE - BASH_FUNC_PRINTF_A_FORMAT: changed to use AC_RUN_IFELSE - BASH_SYS_PGRP_SYNC: changed to use AC_RUN_IFELSE - BASH_SYS_SIGNAL_VINTAGE: changed to use AC_RUN_IFELSE - BASH_SYS_NAMED_PIPES: changed to use AC_RUN_IFELSE - BASH_CHECK_RTSIGS: changed to use AC_RUN_IFELSE - BASH_CHECK_MULTIBYTE: changed to use AC_RUN_IFELSE - RL_LIB_READLINE_VERSION: changed to use AC_RUN_IFELSE - BASH_CHECK_WCONTINUED: changed to use AC_RUN_IFELSE - BASH_FUNC_SNPRINTF: changed to use AC_RUN_IFELSE - BASH_FUNC_VSNPRINTF: changed to use AC_RUN_IFELSE - BASH_FUNC_FNMATCH_EQUIV_FALLBACK: changed to use AC_RUN_IFELSE - BASH_DECL_PRINTF: changed to use AC_RUN_IFELSE - BASH_TYPE_RLIMIT: rewrote to avoid quad_t, now uses AC_COMPILE_IFELSE to determine whether rlim_t is available, otherwise determines the value of RLIMTYPE based on sizeof(rlim.rlim_cur): int, long, or long long - BASH_SIZEOF_RLIMIT,BASH_SIZEOF_QUAD_T: helper macros for BASH_TYPE_RLIMIT in the case that rlim_t is not present configure.ac - BASH_CHECK_DECL -> AC_CHECK_DECLS - quad_t: use AC_CHECK_TYPE (new style) instead of BASH_CHECK_TYPE 9/14 ---- execute_cmd.c - time_command: if we longjmp back to the top_level we saved, make sure we're in the same subshell environment before printing the timing stats. We could have longjmped back from a child process of the command or pipeline we want to time. From a report by Sergej Alikov - time_command: restore command->flags even if we longjmp back general.c - check_binary_file: if the first line of the ENOEXEC file starts with a `#!', check the rest of the buffer for NULs to determine whether this is a binary file. Since we only check the first line, a #! line followed by binary data could be confused for a shell script 9/16 ---- bashjmp.h - EXITBLTIN: new longjmp `code' value: used by the exit builtin {execute_cmd,shell,subst,trap}.c,builtins/evalstring.c - treat EXITBLTIN exactly the same as EXITPROG (for now) builtins/exit.def - exit_builtin: jump_to_top_level with value EXITBLTIN builtins/evalstring.c - parse_and_execute: EXITBLTIN has its own case, with the same contents as EXITPROG 9/17 ---- builtins/evalstring.c - parse_and_execute: change EXITBLTIN case to avoid running the unwind-protect stack to the `pe_dispose' tag (which has the effect of running all the unwind-protects installed by the commands in the string) if we're executing in a function and EXIT is trapped. This has the effect of running the EXIT trap in the function context, which is what we do when we're not in parse_and_execute (eval, bash -c, command substitution, etc.) 9/18 ---- arrayfunc.c - expand_and_quote_kvpair_word,quote_compound_array_word, expand_and_quote_assoc_word,quote_compound_array_list: make sure the value has CTLESC characters doubled even when being single- quoted, since that's what the parser does with standalone assignment statements. From https://savannah.gnu.org/support/?110538 9/19 ---- aclocal.m4 - AC_TRY_COMPILE -> AC_COMPILE_IFELSE (AC_TRY_LINK calls left to modify to avoid obsolete warnings) - BASH_STRUCT_WINSIZE: broke out checks for separate headers into new macros, call them and work with the results; needed to avoid two calls to AC_TRY_COMPILE - BASH_CHECK_KERNEL_RLIMIT: avoid multiple calls to AC_TRY_COMPILE by breaking basic test for RLIMIT_ defines into a separate test and calling the HPUX-specific _KERNEL test only if that fails m4/gettext.m4,configure.ac - AM_GNU_GETTEXT -> BASH_GNU_GETTEXT: remove deprecation warning, since [no-libtool] is how we use this; rename to avoid someone's local copy from overriding ours 9/20 ---- aclocal.m4 - BASH_CHECK_MULTIBYTE: take out code that checks for nl_langinfo(CODESET) since that's already provided by AM_LANGINFO_CODESET, which is called by BASH_GNU_GETTEXT. We AC_REQUIRE it here, though, to avoid problems from future changes - BASH_SYS_SIGNAL_VINTAGE: broke the code that checks for the different signal vintages into different AC_DEFUNed macros, changed BASH_SYS_SIGNAL_VINTAGE to stitch them together with shell code testing the cached variables - AC_TRY_LINK -> AC_LINK_IFELSE: there are no more autoconf warnings 9/21 ---- array.[ch] - array_from_argv: new convenience function array.h - num_elements: now an arrayind_t, since that's the type of indices, it makes sense to allow that many elements 9/22 ---- lib/readline/terminal.c - rl_term_kP: fix typo. Fix from Koichi Murase 9/25 ---- lib/readline/display.c - rl_clear_visible_line: call _rl_clear_to_eol with _rl_screenwidth as the argument so we clear out the entire line even if the terminal doesn't have a clear-to-eol sequence; make sure to add a call to rl_cr after that so we know we're always in column 0 - _rl_redisplay_after_sigwinch: just call rl_clear_visible_line instead of erasing the last line of the display - _rl_redisplay_after_sigwinch: if the prompt is longer than the screen width, make sure to call _rl_reset_prompt to recalculate the local_prompt_newlines array. Should fix issue from https://savannah.gnu.org/support/index.php?110543 redir.c - do_redirection_internal: if given [N]<&WORD- or [N]>&WORD- and WORD expands to null, make it identical to <&- or >&- and close file descriptor N (default 0). From a discussion back in 5/2021 9/27 ---- arrayfunc.c - expand_compound_array_assignment: since we run the string through the parser to turn it into a list (so we can make sure all shell metacharacters are properly quoted), we need to remove the CTLESC the parser uses to quote CTLESC and CTLNUL in *unquoted* words. The rest of the code assumes this has been done, and assumes that any CTLESC characters passed to expansion are part of the original word and should themselves be quoted, doubling the number of CTLESCs 9/28 ---- arrayfunc.c - expand_and_quote_kvpair_word,quote_compound_array_word, expand_and_quote_assoc_word,quote_compound_array_list: if we are single-quoting associative array subscripts and associative and indexed array values, we need to quote CTLESC characters, because that's how they come out of the parser and how the assignment statement code expects to see them. Fixes https://savannah.gnu.org/support/index.php?110538 9/29 ---- subst.c - parameter_brace_transform: invalid transformation operators are now fatal errors in non-interactive shells, as with the other word expansions. Reported by Martin Schulte in https://lists.gnu.org/archive/html/bug-bash/2020-10/msg00026.html execute_cmd.c - execute_disk_command: if we're optimizing out the fork (nofork) and not directly in a pipeline (pipe_in == pipe_out == NO_PIPE), only modify shell_level if subshell_environment says we're not already in a pipeline. Reported by Paul Smith 10/11/2020 against GNU make evalstring.c - should_suppress_fork: remove #if 1 for code that tries to suppress the fork in a process substitution subshell 9/30 ---- builtins/mapfile.def - do_chop: make sure we're comparing unsigned chars when checking whether the delim is the last character on the line. Reported by Greg Wooledge 10/1 ---- lib/readline/rltty.c - rl_deprep_terminal: if we're not echoing to the terminal (_rl_echoing_p == 0), and we just output the bracketed paste end sequence, output a newline to compensate for the \r at the end of BRACK_PASTE_FINI, since redisplay didn't do it for us. Reported by Siteshwar Vashisht shell.h - MATCH_EXPREP: new matching flag, understood only by pattern substitution; means to expand unquoted `&' in the replacement STRING to the match of PATTERN subst.c - shouldexp_replacement: uncommented - pat_subst: we expand & in the replacement string if MATCH_EXPREP appears in MFLAGS - parameter_brace_patsub: push call to shouldexp_replacement out here, after the replacement string is expanded; set MATCH_EXPREP if there is an unquoted `&' (by backslash) in the expanded replacement string doc/{bash.1,bashref.texi} - pattern substitution: overhauled the description, moved each of the possible forms to be tags in the tagged paragraph. The description now specifies the expansions that the replacement string undergoes - pattern substitution: documented new behavior of unquoted & in the replacement string 10/4 ---- shell.c - include if we're debugging malloc (MALLOC_DEBUG) so we can get an extern declaration of trace_malloc_stats 10/5 ---- subst.c - expand_subscript_string: allocate new memory for td.word, copying STRING, in case it gets freed on error by expand_word_internal. Report and fix from Koichi Murase lib/malloc/malloc.c - malloc_usable_size: return the allocated size instead of the chunk size, since writing over the bounds checking will cause fatal errors. Reported by Julien Moutinho , fix from Dominique Martinet arrayfunc.c - unbind_array_elememnt: pass (flags&VA_NOEXPAND) to skipsubscript(), instead of unconditionally passing 1 if we're operating on an associative array. This is consistent with how valid_array_reference determines the length of the subscript. Report and fix from Koichi Murase 10/6 ---- subst.c - skip_to_delim: add a new value for FLAGS. If FLAGS&2, we assume that START is one character after the opening OPEN. If not, we assume that START is at OPEN, and needs to be incremented past it. Part of fix from Koichi Murase arrayfunc.c - unbind_array_element: make sure to pass FLAGS to skipsubscript with bit 2 set, since we are passed a SUB index that's one past the opening bracket. Rest of fix from Koichi Murase - array_variable_name: make sure we pass (FLAGS&1) to skipsubscript, so we don't inadvertently pass a value with bit 2 set, which would cause an off-by-one error in subscript parsing 10/8 ---- trap.c - restore_traps: inverse of reset_signal_handlers. This understands how reset_signal_handlers changes the signal disposition while leaving the trap string in place, and knows how to restore flags and state based on that preserved trap string and whether or not the signal is "special" to the shell. builtins/exec.def - exec_builtin: instead of using restore_original_signals to completely cancel the traps, call reset_signal_handlers so the trap strings are preserved. Then if the exec fails (and we're not exiting) we can look at the trap strings and see how to restore the trap state. This calls restore_traps after reinitializing the shell signal handlers, using the trap strings saved by reset_signal_handlers. Fixes issue with not exiting after a failed exec clearing the EXIT trap reported by Mark March , using the approach suggested by Robert Elz subst.c - expand_declaration_argument: when parsing options that modify attributes that affect how the value is handled (i, c, u, etc.), make sure to create an option string and call make_internal_declare with options that start with a `+' so the attribute is off when the assignment is performed and changes how the value is expanded. From a report by Léa Gris lib/readline/complete.c - rl_display_match_list: even if _rl_completion_prefix_display_length is set to a non-zero value, pass the common prefix length to fnprint if we've turned on colored completion prefixes; passes through to fnprint via print_filename - fnprint: add the ellipsis if prefix_bytes exceeds the _rl_completion_prefix_display_length, add explicit check for prefix_bytes being longer since print_filename passes it through if colored-completion-prefix is set. This means that while completion-prefix-display-length still has precedence over colored-completion-prefix, it doesn't override it if both are set and the common prefix length is shorter than completion-prefix-display-length. From a report by Christoph Anton Mitterer 10/10 ----- parse.y - parse_dparen: if the last token is FOR, increment word_top and assign word_lineno like for other for loops. Fixes bug with LINENO after arithmetic for commands reported by Tom Coleman 10/11 ----- execute_cmd.c - shell_execve: if execve fails with ENOENT, but executable_file() succeeds, display a slightly more descriptive error message. Prompted by a report from Andrea Monaco stringlib.c - strcreplace: allow backslash to escape a backslash, since we allow it to escape a `&' 10/14 ----- pcomplib.c - COMPLETE_HASH_BUCKETS: double to 512 stringlib.c - strcreplace: the last argument is now a flags argument. 1 has its previous meaning (glob); 2 means to allow backslash to escape a backslash (as added on 10/11) subst.c - pat_subst: call strcreplace with 2 as the flags value 10/15 ----- doc/bashref.texi - updates to Installation section doc/{mkposix,mkinst,mkrbash},doc/Makefile.in - changes to allow scripts to be run outside the source directory 10/18 ----- subst.c - patsub_replacement: controls whether pattern substitution expands `&' in the replacement string. Initialized to 1 by default 10/19 ----- doc/{bash.1,bashref.texi} - document bash's WCE SIGINT behavior when job control is not enabled and the shell receives a SIGINT while waiting for the foreground command to complete. Added at Red Hat's request. A complete discussion is at https://www.cons.org/cracauer/sigint.html 10/25 ----- builtins/shopt.def - patsub_replacement: new shell option, exposes patsub_replacement variable controlling whether pattern substitution expands `&' in the replacement string. Still enabled by default. doc/{bash.1,bashref.texi} - patsub_replacement: document new shopt option 10/26 ----- lib/readline/display.c - expand_prompt: group runs of invisible characters at the right edge of the screen with the previous physical characters when setting local_prompt_newlines, since that's how update_line() expects to get it. Fix from sparrowhawk996@gmail.com. lib/readline/macro.c - rl_end_kbd_macro: make sure current_macro_index is > 0 after subtracting the key sequence length, clamp it to 0 before writing the ending NULL. From a fuzzing report by Tillmann Osswald lib/readline/isearch.c - _rl_isearch_dispatch: in the bracketed paste case, don't assume the pasted text is null-terminated, so use memcpy instead of strcpy for the length of the pasted text, then make sure the search string is null-terminated. From a fuzzing report by Tillmann Osswald lib/readline/text.c - rl_transpose_words: make sure to preserve the value of rl_end from the beginning to the end of the function. From a fuzzing report by Tillmann Osswald lib/readline/vi_mode.c - rl_vi_delete_to,rl_vi_change_to,rl_vi_yank_to: if we are redoing a command (_rl_vi_redoing == 1), save the old _rl_vimvcxt, allocate a new one, and restore the old one before returning. Prevents some pointer aliasing problems. From a fuzzing report by Tillmann Osswald 10/29 ----- arrayfunc.c - tokenize_array_reference: take valid_array_reference and add a third argument (char **SUBP), which, if non-NULL, gets the null-terminated subscript parsed from the NAME argument. If it's NULL, the caller gets the old valid_array_reference behavior. Fix from Koichi Murase - valid_array_reference: just call tokenize_array_reference with a NULL third argument - unbind_array_element: assume the caller (unset_builtin) passes a null-terminated SUB that's already been validated by a call to tokenize_array_reference so we don't need to call skipsubscript() or check VA_ONEWORD. Fix from Koichi Murase arrayfunc.h - tokenize_array_reference: extern declaration builtins/set.def - unset_builtin: use tokenize_array_reference to figure out T and pass that to unbind_array_element. Fix from Koichi Murase - unset_builtin: pass non-null argument to array_variable_part to get the length of the subscript (T), then cut off any final `]' before passing it to unbind_array_element, since that's what it now expects subst.c - expand_string_for_rhs,expand_string_for_pat: assign td.word from newly-allocated memory in case it gets freed on error during the call to call_expand_word_internal(); free it manually when that call returns 11/1 ---- findcmd.c - search_for_command: if FLAGS includes CMDSRCH_STDPATH, don't look in the hash table for the command name. Prompted by a report from Roger Morris aclocal.m4 - BASH_FUNC_POSIX_SETJMP: add a check by fetching the signal mask after the siglongjmp and making sure that SIGINT is not blocked, indicating we restored the original signal mask 11/2 ---- subst.c - expand_string_assignment: make sure to add W_TILDEEXP to the flags so expand_word_internal performs the right tilde expansion on tildes following an unquoted colon. Report from Anders Kaseorg 11/3 ---- aclocal.m4 - BASH_FUNC_POSIX_SETJMP: if cross-compiling, default to `present' if we've determined we have posix signal functions 11/4 ---- execute_cmd.c - SET_LINE_NUMBER: set line_number, but don't set line_number_for_err_trap if we're already running the ERR trap - GET_LINE_NUMBER: evaluates to line_number_for_err_trap if we're running the ERR trap and executing_line_number() otherwise - execute_function: use GET_LINE_NUMBER to push the value for the line number into the BASH_LINENO array - execute_command_internal,execute_arith_command,execute_cond_command: use SET_LINE_NUMBER to avoid overwriting line_number_for_err trap while executing the ERR trap. Tentative fix for `caller' problem reported by Quinn Grier configure.ac,patchlevel.h - set bash version to 5.2-devel, meaning shell_compatibility_level = 52, which required updating tests (array.tests, array21.sub, quotearray3.sub, new-exp10.sub, history2.sub). The first three are because bash-5.2 flushes an indexed array but does not unset the variable when given `unset a[@]' or `unset[*]' (that is, it acts like `a=()' instead of `unset a'); the other two are because the version number changed 11/8 ---- doc/Makefile.in - bash.info: use `makeinfo -o' instead of calling infopost.sh to edit the filename, since the tags table includes absolute byte offsets to nodes. From https://savannah.gnu.org/support/?110557 11/15 ----- examples/loadables/realpath.c - add -a NAME argument to put canonicalized pathnames into an indexed array instead of displaying them on stdout. Suggested by felix@f-hauri.ch lib/readline/colors.c - _rl_custom_readline_prefix: new function, looks in $LS_COLORS for a custom extension (*.readline-colored-completion-prefix) and uses that, if found, as the default color for the common prefix displayed when `colored-completion-prefix' is set. Suggested by Christoph Anton Mitterer - _rl_print_prefix_color: try the return value from _rl_custom_readline_prefix before defaulting to the C_PREFIX (C_SOCK) color. Suggested by Christoph Anton Mitterer lib/readline/doc/{readline.3,rluser.texi},doc/bash.1 - readline-colored-completion-prefix: document new custom suffix for readline's colored-completion-prefix color 11/16 ----- doc/{bash.1,bashref.texi},builtins/set.def - set: modify usage synopsis slightly, based on https://bugzilla.redhat.com/show_bug.cgi?id=2022324 builtins/set.def - unset_builtin: add shell compatibility level check <= bash-5.1 to force `unset a[@]' to unset array variable `a', like in previous versions, instead of unsetting associative array element `@' or removing all elements of an indexed array without unsetting the variable itself builtins/common.c - set_expand_once: now a no-op if the shell compatibility level is <= bash-5.1, preserving the previous versions' behavior for [[ -v (dependent on assoc_expand_once instead of forcing it) subst.c - parameter_brace_expand_rhs: make sure that the final value assigned is always returned from ${a:=b}, even when `a' has a variable attribute that modifies the value on assignment. Reported back on 1/20/2021 by oguzismailuysal@gmail.com; bash-5.2 tag removed 11/22 ----- doc/{bashref.texi,bash.1} - unset: change the description of `unset name[@]' (unset name[*]) to reflect the new behavior (unset associative array subscripts or remove all elements from an indexed array) and the difference from previous versions - set: document -r for restricted shell mode - restricted shell: change occurrences of `set +o restricted' to `shopt -u restricted_shell'. From a report from Jin Xiang - read: note that read -t0 may return 0 if a subsequent read will return EOF. From a suggestion by Dale R. Worley execute_cmd.c - execute_arith_command,eval_arith_for_expr: don't pass EXP_EXPANDED to evalexp if shell_compatibility_level > 5.1, assuming that the expression has been quoted by Q_ARITH handling test.c - test_arithcomp: if we're being called by the conditional command code and treating the arguments to the arithmetic operators as expressions, don't pass EXP_EXPANDED to evalexp if the shell compatibility level is > 5.1, assuming that the expression has been quoted by Q_ARITH handling 11/29 ----- examples/loadables/getconf.[ch] - getconf: new implementation of loadable builtin, based on glibc getconf examples/loadables/sleep.c - parse_gnutimefmt: parse GNU interval format (2m30s), returning seconds and microseconds like uconvert - sleep: use parse_gnutimefmt if uconvert can't parse the format right away and "dhms" appears in the format string. Don't return EX_DISKFALLBACK under any circumstances 11/30 ----- builtins/wait.def - wait_builtin: refer to unset_waitlist only if JOB_CONTROL is enabled. Report from Joel Ebel parse.y,builtins/set.def - changes for minimal config restrictions configure.ac,config.h.in - --enable-translatable-strings: new configuration option to allow $"..." support to be compiled in or out; not included in the minimal shell configuration shell.c - main: don't include any of the dump-strings options if TRANSLATABLE_STRINGS is not defined lib/sh/shquote.c - sh_backslash_quote_for_double_quotes: needed for translatable strings parse.y - support for translating $"..." strings now conditional on TRANSLATABLE_STRINGS locale.c - locale_expand: now conditional on TRANSLATABLE_STRINGS subst.c - expand_string_dollar_quote: support for $"..." now conditional on TRANSLATABLE_STRINGS doc/bashref.texi - --enable-translatable-strings: document new configuration option 12/2 ---- subst.c - verify_substring_values: now that the default compatibility level is 52, and the Q_ARITH code is enabled, make the EXP_EXPANDED flag for evalexp() dependent on the compatibility level - param_expand: arithmetic substitution: make EXP_EXPANDED flag for evalexp() dependent on the compatibility level - expand_word_internal: don't call expand_array_subscript if the shell compatibility level is 51 or below (Q_ARITH) test.c - test_unop: if the shell compatibility level is > 51, using [@] with an existing associative array will report on a key of `@' arrayfunc.c - array_expand_index: if the compatibility level is > 51, Q_ARITH is used and we don't pass EXP_EXPANDED to evalexp() COMPAT,doc/bashref.texi - shell compatibility mode: document effects of setting the compatibility mode to 51 12/3 ---- lib/malloc/malloc.c - mremap: only use if MREMAP_MAYMOVE is defined, since we use the Linux version of the function signature 12/6 ---- bashhist.c - bash_add_history: if we're parsing a here-document (PST_HEREDOC), only suppress adding the newline between lines if we're not at the first line of the here-document (here_doc_first_line != 0). From a report by S0AndS0 12/8 ---- lib/readline/colors.c - _rl_custom_readline_prefix: use STREQN to check for the extension string in $LS_COLORS, since it's not necessarily null-terminated. From https://savannah.gnu.org/patch/?10158 12/10 ----- variables.c - set_int_value,set_string_value: broke common code for setting int and string dynamic variable values out into separate functions; changed all callers to use them where appropriate. set_int_value takes a flags argument saying whether or not to force the integer attribute on - assign_random: store the value assigned as the variable value so things like RANDOM=42; RANDOM+=7 generate consistent sequences like in ksh93 - assign_seconds: store the value assigned as the variable value so things like SECONDS=42 ; SECONDS+=7 generate what's expected doc/Makefile.in - changes to allow man pages that include others (.so FN) to be built outside the source tree 12/13 ----- arrayfunc.c - assign_array_element_internal: take an additional argument: char **NVALP. If non-null, it gets the value eventually assigned to the array element - assign_array_element: take an additional NVALP argument; pass it to assign_array_element_internal arrayfunc.h - assign_array_element: new extern function declaration {subst,variables}.c,builtins/{common.c,declare.def} - assign_array_element: change callers subst.c - parameter_brace_expand_rhs: for the ${param:=value}, use the value returned by assign_array_element in NVALP as the return value, since it's the value ultimately assigned to the variable after possible modification (e.g., arithmetic evaluation). Reported by oguzismailuysal@gmail.com after flawed fix applied 11/16 12/14 ----- arrayfunc.h - array_eltstate_t: an object that encapsulates an array element's state (type, index, key, value) whether it's an indexed or associative array arrayfunc.c - {init,flush}_eltstate: new functions to initialize and flush any allocated memory from array_eltstate_t objects. No allocation/ deallocation functions yet; the only use is with a static instance - assign_array_element_internal: take an array_eltstate_t * instead of a char ** as the final argument, so we can return keys/indices and values depending on the type of array; populates it with the appropriate values - assign_array_element: take array_eltstate_t * as final argument instead of a char **; pass it to assign_array_element_internal {subst,variables}.c,builtins/{common.c,declare.def} - assign_array_element: change callers to modify final argument 12/15 ----- arrayfunc.c - array_value_internal: now takes an array_eltstate_t * as the final argument; there is no more `rtype' argument in favor of the `subtype' member; returns the appropriate values in its members - array_value: changed to pass array_eltstate_t to array_value_internal, saves and fetches its `ind' member into *indp; saves `subtype' member into *rtype - get_arrary_value: changed to take array_eltstate_t as third argument, passes it to array_value_internal {redir,expr}.c - get_array_value: changed callers; initializing the array_eltstate_t argument as necessary test.c - test_builtin: changed to use get_array_value, adding AV_ALLOWALL to the flags, since it didn't use any QUOTED argument. Pass array_eltstate_t * as final argument and get subtype from it (the only thing we're interested in, to deallocate memory) 12/16 ----- arrayfunc.c - array_value: now takes a final argument of array_eltstate_t *, which it passes to array_value_internal; no more rtype and indp args. Callers are responsible for marshalling values into estatep arrayfunc.h - array_value: changed function signature subst.c - get_var_and_type,parameter_brace_expand_word: changed calls to array_value to use array_eltstate_t argument and initialize it appropriately. Copy values back from it to the parameters we need to modify variables.c - assign_lineno: call set_int_value to store the value, like with RANDOM and SECONDS (from 12/10) 12/17 ----- {eval,execute_cmd}.c - when bypassing a parsed command because read_but_dont_execute is set, don't modify last_command_exit_value. From a report by Robert Elz parse.y - parse_comsub: make sure the first call to shell_getc to check whether or not it's an arithmetic expansion skips a quoted newline. From a report by Robert Elz 12/21 ----- subst.c - parameter_brace_remove_pattern,parameter_brace_patsub,parameter_brace_casemod, parameter_brace_transform,parameter_brace_substring: now take an array_eltstate_t * argument in place of the arrayind_t argument, pass it to get_var_and_type; this generalizes the indexed array behavior of expanding array subscripts once to associative arrays via an eventual call to array_value_internal with a non-null KEY member - get_var_and_type: now takes an array_eltstate_t * argument in place of the arrayind_t argument; use it in calls to array_value so we can only expand array subscripts once whether they are indexed or associative arrays - parameter_brace_expand_word: take an array_eltstate_t * argument in place of the arrayind_t * argument; pass it to array_value; use a static version (which we init and flush) if the argument passed is NULL so we can get the right state passed back and forth - parameter_brace_expand: pass a pointer to a static array_eltstate_t to parameter_brace_expand_word, and use that in the various calls to parameter_brace_XXX functions that perform specific expansions in place of the old arrayind_t argument; make sure to flush it before returning, even on errors 12/22 ----- {trap,variables}.c - internal_warning: calls changed to use translatable strings consistently error.[ch] - internal_debug: new function, prints a message like internal_warning, no-op if DEBUG is not defined - INTERNAL_DEBUG: macro that expands to internal_debug when DEBUG is defined, and nothing otherwise {jobs,trap}.c - changed some internal_warning and internal_inform calls to use internal_debug, since they were active only when DEBUG is defined parse.y - parse_comsub: add internal_debug call when a command substitution ends with unterminated here-documents builtins/common.c - number_of_args: unconditionally return posparam_count {jobs,execute_cmd,subst}.c,parse.y,builtins/{command.def,evalstring.c} - INTERNAL_DEBUG: use instead of calls to itrace protected by #ifdef DEBUG 12/26 ----- lib/glob/glob.c - glob_always_skip_dot_and_dotdot: initialize to 1 (enabled) builtins/shopt.def - globskipdots: new shell option, exposes glob_always_skip_dot_and_dotdot doc/{bash.1,bashref.texi} - globskipdots: document new shell option execute_cmd.c - fix_arrayref_words: call valid_array_reference with 0 for third arg because the words have not undergone any word expansions yet and the quotes are still present. This makes things like A=[\[]=set unset A[\[] work subst.c - word_list_split: if a word undergoes word splitting but is not changed, preserve any W_ARRAYREF flag into the new word. This makes things like rkey=']' unset A[$rkey] work because the unset builtin sees the W_ARRAYREF flag on its argument 12/29 ----- builtins/common.h - SET_VFLAGS: take a set of word flags, and set flags for use with valid_array_reference (vflags) and assign_array_element/bind_int_variable (bindflags) based on those flags and the setting of assoc_expand_once (moved from read.def and generalized) arrayfunc.c - array_variable_name: now exclusively takes AV_xxx flags; understands how to translate AV_NOEXPAND to the right flag for skipsubscript; understands AV_ONEWORD and AV_NOEXPAND - array_variable_part: just passes the FLAGS arg to array_variable_name - assign_array_element: translates ASS_ flags to AV_ flags for array_variable_name - array_value_internal: now just passes flags straight through to array_variable_part (which sends them to array_variable_name) builtins/common.[ch] - builtin_bind_var_to_int: now takes a new FLAGS third argument; passes it to bind_var_to_int builtins/printf.def - printf_builtin: use SET_VFLAGS to set flags for builtin_bind_variable (bindflags); makes things like declare -A A; key=']' ; printf -v A[$key] "foo" work without a subscript error as long as assoc_expand_once is defined builtins/read.def - read_builtin: use new common version of SET_VFLAGS instead of private version; changed to use new calling sequence; makes things like declare -A A; key=']' ; read A[$key] <<<"foo" work without a subscript error as long as assoc_expand_once is defined builtins/wait.def - wait_builtin: use SET_VFLAGS for variable name with -p option - wait_builtin: call builtin_bind_var_to_int with new bindflags third argument expr.c - expr_streval: just pass TFLAG (AV_xxx flags) straight through to array_variable_part variables.c - bind_int_variable: translate the assignment flags (ASS_xxx) to VA_xxx flags for valid_array_reference calls (ASS_ONEWORD); translate assignment flags to AV_xxx flags for array_variable_part 12/30 ----- subst.c - parameter_brace_expand: when expanding an indirect variable, extend the special case for array[@] and array[*] (set -u/no positional parameters, obeying the baroque quoting rules) to the value of the indirection. Report and fix from konsolebox 12/31 ----- parse.y - compound_list: when parsing a compound_list production, collect any pending here-documents after reading a newline, not after reading any command terminator. Fixes interactive-only prompting bug reported back in 8/2021 by Hyunho Cho 1/1/2022 -------- bashline.c - set_filename_quote_chars: break code that modifies rl_filename_quote_characters based on whether DIRNAME needs to be expanded from bash_directory_completion_hook into its own function - bash_check_expchar: break code that checks whether DIRNAME will be word expanded from bash_directory_completion_hook into its own function - bashline_reset,attempt_shell_completion: make sure complete_fullquote is set to 1 (as it is by default) in case a completion modifies it - bash_quote_filename: if we are completing (but not expanding -- direxpand is unset) and backslash-quoting a filename with expansion characters as determined by bash_check_expchar, make sure filename_bstab is set not to include the expansion char (and any following char and closer) and set complete_fullquote to 0 so sh_backslash_quote uses filename_bstab. Fixes the longstanding issue of quoting a `$', for instance, if the rest of the filename contains any characters that need quoting in filenames. This assumes that the filename is unquoted (*QCP == 0) so the word will be expanded and is not part of the filename (if needed, we can use file_exists to check whether the expansion characters are actually part of the filename) 1/7 --- examples/loadables/accept.c - new -b bindaddr option to bind to a specific IP address - move setsockopt calls before bind() to avoid TIME_WAIT issues. Fixes from Dzogovic Vehbo 1/10 ---- configure.ac - bumped version to 5.2-alpha. 1/16 ---- tests/{unicode1,glob2,intl2}.sub, tests/run-intl - minor changes to add warnings for missing locales that cause test failures 1/17 ---- parse.y - history_delimiting_chars: if we have a blank line by the time we hit the end of the tests, return a semicolon for the first blank line to avoid running lines of a multi-line command together. Fixes bug reported by Joakim Lindblad subst.c - expand_string_for_patsub: expand the replacement string for pattern substitution assuming that it will eventually be passed to strcreplace to replace `&' with the matched portion of the string. This calls expand_string_for_pat(), which leaves the string quoted and does not perform word splitting, then calls quote_string_for_repl to post-process the expanded string. - quote_string_for_repl: perform quote removal on passed string while replacing CTLESC escaping a `&' or a backslash with a backslash. The result must be passed to strcreplace - parameter_brace_patsub: call expand_string_for_patsub if patsub_replacement is set to quote 1/18 ---- subst.c - read_comsub: make istring_index a size_t to avoid overflow with very large values of istring_size. From https://savannah.gnu.org/support/index.php?110596 - expand_word_internal: make istring_index a size_t 1/20 ---- buitins/cd.def - add a description of `cd -' to the help text. Suggested by Rob Landley 1/21 ---- lib/glob/glob.c - glob_vector: if we allocate NEXTLINK using malloc, and free it due to some allocation failure, reset FIRSTMALLOC to avoid duplicate frees later on subst.[ch] - sub_append_string: the INDX parameter is now a size_t to avoid overflow parse.y - decode_prompt_string: RESULT_INDEX is now a size_t to pass to sub_append_string jobs.[ch],nojobs.c - wait_for_background_pids: now returns the number of jobs/processes reaped builtins/wait.def - wait_builtin: if -p pid supplied without -n, make sure we do something useful if no job/pid arguments are supplied and there are no jobs. Reported by Oguz builtins/read.def - read_builtin: if we have a timeout, use SIGALRM instead of select when in posix mode, since we use read instead of zread. Fixes bug reported by Andreas Schwab subst.c - expand_string_dollar_quote: handle single-quoted and double-quoted strings that might include $' and $" without attempting translation; do more error checking for unterminated $' and $" that leaves those characters unmodified. This is for use by readline's various line expansion functions (shell_expand_line) 1/23 ---- parse.y,make_cmd.c - revert change that unconditionally processes $'...' and $"..." in here-document bodies; there are only a couple of cases where they should be processed in a double-quote environment 1/24 ---- subst.c - extract_dollar_brace_string: if we see another `${' on the rhs of the operator, reset the dolbrace_state to DOLBRACE_PARAM while we read this new ${...} string - extract_heredoc_dolbrace_string: new function, variant of extract_dollar_brace_string, to process the WORD in ${PARAM OP WORD} while expanding lines of here-document data. It's complicated by the requirement to add to the result string as we go along, since we need to change the contents of the input string with ansi expansion or locale translation. - string_extract_single_quoted: take a new third argument: ALLOWESC. This allows backslash to escape an embedded single quote, needed by extract_heredoc_dolbrace_string to process $'...'; changed callers 1/25 ---- parse.y - parse_matched_pair: ansi-expand $'...' in WORD for ${PARAM OP WORD} and single-quote the result if dolbrace_state == DOLBRACE_QUOTE (posix pattern removal operators) even if extended_quote == 0 subst.c - extract_heredoc_dolbrace_string: add logic to align with parse.y: parse_matched_pair and its $'...' expansion, including handling extended_quote 1/27 ---- builtins/evalstring.c - should_optimize_fork: broke conditions for optimizing away the fork for a simple command out of optimize_fork into new function, call from should_suppress_fork and optimize_subshell_command. Call from optimize_fork if (subshell_environment & SUBSHELL_PAREN), relying on fact that CMD_TRY_OPTIMIZING is only set in a couple of specific conditions - optimize_fork: call should_suppress_fork only if startup_state == 2; it does the extra checks for that specific case - optimize_fork: call should_optimize_fork if we're in a (list) subshell (subshell_environment & SUBSHELL_PAREN) - optimize_subshell_command: set CMD_TRY_OPTIMIZING on the right side of a `&&', `||', or `;' list as long as it's a simple command so we can check with optimize_fork() when it's time to execute it execute_cmd.c - execute_in_subshell: call optimize_subshell_command for (list) subshells to either set CMD_NO_FORK for simple commands or set CMD_TRY_OPTIMIZING for likely candidates for later optimization builtins/common.h,builtins/evalstring.c - optimize_fork: renamed to optimize_connection_fork; changed callers 1/31 ---- include/shmbutil.h - COPY_CHAR_I,SCOPY_CHAR_I: add check for locale_utf8locale and (c & 0x80) as in other macros lib/sh/shquote.c - sh_backslash_quote_for_double_quotes: rewrote to use array indexing and COPY_CHAR_I to make it easier to drop in future calls to charvis() to make `unsafe' characters visible if FLAGS == 1 2/1 --- parse.y - parse_comsub: if we are currently expanding aliases, temporarily turn off alias expansion if we are not in posix mode so we defer alias expansion until command_substitute(). Fixes double-expansion bug reported by Martijn Dekker and aligns with https://www.austingroupbugs.net/view.php?id=1342 - xparse_dolparen: turn off alias expansion entirely while running the parser: either we do it in parse_comsub (posix mode) or in command_substitute (default mode) - parse_string_to_command: ditto subst.c - command_substitute: if we are expanding aliases, temporarily turn off alias expansion if we are in posix mode, since we already performed it in parse_comsub() and are using the command string reconstituted from the parse result doc/bashref.texi - bash posix mode: add description of alias expansion and command substitution parsing and execution 2/4 --- lib/readline/rltty.c - rl_deprep_terminal: set _rl_last_c_pos to 0 after outputting BRACK_PASTE_FINI, since the last character in that is \r. Partially address issue raised by Markus Schwarzenberg in https://lists.gnu.org/archive/html/bug-bash/2022-02/msg00056.html 2/5 --- doc/{bash.1,bashref.texi} - minor typo fixes from Helge Kreutzmann 2/7 --- {arrayfunc,variables}.c - ARRAY_EXPORT: changes to encode array and assoc variables using a scheme similar to shell functions so we can export arrays and differentiate them from scalar variables and differentiate array and assoc variables. Still not enabled by default. variables.c - mk_env_string: third argument is now the variable's attributes or 0, with all the attributes we can export arrays lib/readline/bind.c - active-region-start-color,active-region-end-color: new bindable string variables, one to set the active region color (instead of standout mode) and one to turn it off (instead of the "se" terminal capability). They set _rl_active_region_start_color and _rl_active_region_end_color variables via functions lib/readline/display.c - putc_face: if setting standout mode, check for both the start color and end color variables and output the start color string. If turning off standout mode (normal face), output the end color string. Both variables must be set lib/readline/{readline.c,rlprivate.h} - declarations for _rl_active_region_start_color and _rl_active_region_end_color 2/8 --- bashline.c - initialize_readline: add bindable name `vi-edit-and-execute-command' shell.c - subshell_exit: make sure to set last_command_exit_value before calling the exit trap. Fixes bug reported by Greg Edwards 2/9 --- lib/readline/{terminal.c,rlprivate.h} - _rl_region_color_on,_rl_region_color_off: functions to output the _rl_active_region_start_color and _rl_active_region_end_color - _rl_reset_region_color: function to encapsulate setting the region color to an arbitrary string, including doing memory management lib/readline/display.c - putc_face: call _rl_region_color_on and _rl_region_color_off instead of _rl_standout_on and _rl_standout_off lib/readline/terminal.c - _rl_init_terminal_io: initialize _rl_active_region_start_color and _rl_active_region_end_color from _rl_term_so and _rl_term_se, respectively; reset every time the terminal is changed - _rl_init_terminal_io: turn off the active region for a dumb terminal lib/readline/bind.c - sv_region_{start,end}_color: call _rl_reset_region_color with the appropriate value for WHICH 2/10 ---- lib/readline/doc/{rluser.texi,readline.3},doc/bash.1 - active-region-start-color,active-region-end-color: documented new bindable readline variables - enable-active-region: document using active-region-start-color to highlight the text in the region 2/11 ---- parse.y - read_token,read_token_word: make sure characters read by shell_getc are protected by appropriate calls to MBTEST when testing for shell metacharacters and operators 2/14 ---- builtins/shopt.def - set_compatibility_level: if the current compatibility level is outside the range of the compatNN options, just leave it alone when unsetting one of the options (which by definition was already unset). Fixes issue reported by Mihai Moldovan 2/16 ---- lib/readline/search.c - rl_history_search_{pos,len,flags}: rename to have a leading `_' - _rl_history_search_pos: no longer static so other parts of readline can see it lib/readline/rlprivate.h - _rl_history_search_pos: extern declaration lib/readline/readline.c - readline_internal_teardown: don't run the undo list against the current history entry if the non-incremental search functions have set _rl_history_search_pos to it, since it doesn't reflect the current contents of the line buffer. Fixes issue reported by Andreas Schwab lib/readline/misc.c - _rl_start_using_history: initialize _rl_history_search_pos to something invalid so it doesn't match where_history() 2/17 ---- lib/readline/callback.c - rl_callback_read_char: make sure _rl_eof_found is set to the value of eof before calling the deprep terminal function, so it can do different things based on whether the input code read EOF (or the user entered the EOF character). From a gdb discussion started by Andrew Burgess (still more to do, since this is not part of the public API) 2/18 ---- lib/readline/readline.h - RL_STATE_EOF: new readline state value; set when readline reads an EOF character on an empty line or a read returns an error - rl_eof_found: new public variable lib/readline/rprivate.h - _rl_eof_found: renamed to rl_eof_found, so not declared here lib/readline/{callback,readline}.c - RL_STATE_EOF: set appropriately when readline gets an EOF. Suggested by Andrew Burgess - RL_STATE_EOF: make sure it's not set when readline starts - rl_eof_found: set appropriately when readline gets an EOF lib/readline/{callback,readline,rltty}.c - rl_eof_found: new name for _rl_eof_found lib/readline/doc/rltech.texi - RL_STATE_EOF: document 2/19 ---- parse.y - parse_comsub: turn off parser state flags we don't want to inherit into this call to the parser (PST_REGEXP, PST_EXTPAT, PST_CONDCMD, PST_CONDEXPR for now). Fixes bug reported by konsolebox 2/23 ---- findcmd.c,builtins/hash.def - replace calls to is_directory with file_isdir, which only performs a stat and doesn't do the eaccess call to check for an executable file findcmd.c - find_in_path_element: takes a new RFLAGSP argument, an int * where the status flags for the returned pathname are returned; saves additional calls to stat/eaccess - search_for_command: get the returned flags from find_user_command_in_path so we don't need any additional calls to file_status after we find the command in $PATH 2/24 ---- doc/{bash.1,bashref.texi} - FUNCTIONS: some small changes to the description of local variables and dynamic scoping, with emphasis on how that affects `unset' behavior. Inspired by a discussion with Christoph Anton Mitterer 2/25 ---- examples/loadables/realpath.c - renamed -s option to -q to align with other versions - perform array assignment for `-a varname' even if -q option supplied - renamed -S option to -s for Linux compatibility 2/28 ---- lib/readline/misc.c - _rl_free_saved_history_line: call rl_free_undo_list, saving and setting rl_undo_list to the saved history line's data, so the right call to _hs_replace_history_data happens and we don't end up with a pointer aliasing problem. Fixes core dump reported by Andreas Schwab , but does not make his scenario equivalent to incremental search 3/1 --- lib/readline/search.c - make_history_line_current: save the current line before replacing it with the found history entry using rl_maybe_save_line - noninc_dosearch: we don't want the saved history line, so free it after calling make_history_line_current - _rl_history_search_internal: call rl_maybe_replace_line after making changes to the line buffer with make_history_line_current so we can save the undo list we constructed before we set the history position 3/2 --- lib/readline/display.c - expand_prompt: add missing piece to patch from 10/26/2021: if we are recalculating the number of invisible characters on the first line of the prompt, we need to update INVFL, even if we already set it when we hit the number of physical characters. This ends up being assigned to prompt_invis_chars_first_line, and is used in several subsequent calculations. Reported by Andreas Schwab lib/readline/doc/{readline.3,rluser.texi},doc/bash.1 - enable-bracketed-paste: add some language making it clearer that bracketed paste prevents the pasted text from being interpreted as editing commands. Suggested by Karl O. Pinc 3/4 --- make_cmd.c - make_here_document: perform quote removal on the here-doc delimiter only if it's marked as quoted, which prevents quotes from inside a command substitution from being removed (they're supposed to begin a new quoting context) when the word itself isn't flagged as quoted (which means the body of the here-document gets expanded). You can't perform quote removal *and* expand the here-document lines. From an austin-group discussion back in early February lib/sh/strvis.c - charvis -> sh_charvis; change caller - sh_charvis: now take an additional SLEN argument to avoid having to compute the string length every time; change callers - sh_charvis: add a utf-8 locale-specific check before calling COPY_CHAR_I (in practice, doesn't make any real difference) 3/10 ---- arrayfunc.c - convert_var_to_array: if we're being asked to create an associative array (flags & 2), and we have an existing variable that is not an assoc array (and not an existing indexed array), call convert_var_to_assoc to make it one 3/11 ---- jobs.c - wait_for: don't call get_tty_state() if readline is dispatching (RL_STATE_DISPATCHING) with the terminal settings changed (RL_STATE_TERMPREPPED), the same way we don't if we are running a command for programmable completion. Fixes bug with SIGINT reverting to the saved readline terminal settings reported by Markus Napierkowski parse.y - decode_prompt_string: make sure the expansion of \w, \W, and \s are all run through sh_strvis before calling sh_backslash_quote_for_double_quotes or just through sh_strvis if we're not running the prompt string through word expansions. Fixes issue reported by Josh Harcome back in mid-January 3/16 ---- bashline.c - bash_quote_filename: if we have a word to complete that contains characters that introduce a word expansion, make sure the passed string does *not* exist as a filename before removing those characters from the set that must be backslash-quoted. See change from 1/1/2022 3/18 ---- lib/readline/search.c - make_history_line_current: don't free rl_undo_list or _rl_saved_line_for_history; don't unconditionally save the history line. This reverts some of the changes to support setting the history position in history-search-backward - rl_history_search_internal: only free the saved history line if we were the ones who created it 3/21 ---- lib/readline/nls.c - xmalloc.h: include for systems without setlocale(), so xfree has a prototype. Report and fix from András Kucsma lib/readline/search.c - _rl_history_search_internal: use previous-history/next-history to move to the found history line instead of directly calling history_set_pos. This makes the behavior more similar to incremental search - rl_history_search_internal: make sure to set rl_undo_list to the current history undo list around the calls to rl_get_previous_history or rl_get_next_history, in order to fool the call to maybe_replace_line they make lib/readline/readline.c - _rl_executing_func: the currently-executing readline command function lib/readline/rlprivate.h - _rl_executing_func: extern declaration lib/readline/search.c - _rl_history_search_internal: removed (commented out) code that sets the current history entry to the found history entry, too much assumes that the current undo list should be applied to the current history entry (where_history()) 3/23 ---- subst.c - parameter_brace_expand_word: if we have double-quoted ${*} or ${@}, make sure we are setting W_HASQUOTEDNULL in the flags we return to the caller if we are returning QUOTED_NULL(word) - parameter_brace_expand_word: if we have a double-quoted associative array reference using `*' or `@', make sure we are setting W_HASQUOTEDNULL in the flags we return to the caller if we are returning QUOTED_NULL(word) - parameter_brace_expand: if we're using the `[:]+' word expansion operator, we need to note a quoted null string and pass the W_QUOTEDNULL flag back to the caller - expand_word_internal: make sure to return a QUOTED_NULL (word[0] == CTLNUL) back to the caller if HAD_QUOTED_NULL is set, regardless of whether or not we see a quoted dollar at. Fix for bug reported by Andreas Luik arrayfunc.c - array_value_internal: fix typo and set estatep->type to ARRAY_INDEXED for indexed arrays 3/31 ---- lib/readline/{history.c,histlib.h} - _hs_at_end_of_history: convenience function to tell whether or not the current history position is at the end of the history list 4/1 --- lib/readline/search.c - make_history_line_current: don't free rl_undo_list if it is equal to _rl_saved_line_for_history->data, since we will need to restore it later if we got it from a history entry. Fixes issue dating back to 7/2021 and changes to _rl_free_saved_line_for_history, current issue reported by Andreas Schwab 4/5 --- lib/readline/{complete,histfile,histsearch,isearch,terminal}.c - xfree: use instead of free 4/7 --- configure.ac - bumped version to bash-5.2-beta [bash-5.2-beta frozen] 4/8 --- lib/readline/input.c - _rl_orig_sigset: need extern declaration if HAVE_SELECT is defined. From https://savannah.gnu.org/support/?110634 examples/loadables/seq.c - PRIdMAX: redefine if PRI_MACROS_BROKEN is defined. From https://savannah.gnu.org/support/index.php?110635 4/11 ---- configure.ac - BASH_FUNC_STRTOIMAX: replace strtoimax if the system doesn't provide a declaration in a standard header file. Uses new m4/strtoimax.m4. From https://savannah.gnu.org/support/index.php?110633 builtins/printf.def - getdouble: new function, parses string into `double' using strtod - printf_builtin: check for the `L' length modifier and use long doubles for the floating point conversion specifiers. If it's not supplied, use `double' when in posix mode (as posix specifies) and long double (if it's available, double if not) in default mode. From a report from Paul Eggert 4/12 ---- lib/sh/oslib.c - bzero: update function signature to modern BSD version 4/14 ---- lib/sh/oslib.c - bcopy, gethostname, mkfifo: update function signatures to modern versions 4/15 ---- jobs.c,nojobs.c - wait_for_single_pid: if the pid or job argument is invalid -- isn't a child of this shell -- return 257, which is out of the range of valid 8-bit status values execute_cmd.c - execute_pipeline: if wait_for_single_pid returns > 256, set it to 127 (invalid process) jobs.c - wait_for_background_pids: if wait_for_single_pid returns > 256, set the status we return in PS to 127 (what it was before) builtins/wait.def - wait_builtin: if wait_for_single_pid returns > 256, treat it as an error and set pstat.pid to NO_PID - wait_builtin: if -p supplied, and we get to the end of the argument list with PSTAT.PID != NO_PID (which we assume means that the return value is set from PSTAT.STATUS), set the variable name to PSTAT.PID. From a report by Robert Elz - wait_builtin: for compatibility with the netbsd sh, leave the variable name specified with `-p' unset if there are no PID arguments. From a report by Robert Elz 4/17 ---- parse.y - xparse_dolparen: if (flags & SX_NOLONGJMP), don't call jump_to_top_level() on errors bashline.c - bash_quote_filename: don't call quote_word_break_chars() unless we have word break chars initialized. Fixes bug reported by Sam James 4/18 ---- pcomplete.c - gen_globpat_matches: call glob_filename with the GX_GLOBSTAR flag if the `globstar' shell option is enabled. From a report by Steve lib/malloc/malloc.c - internal_free: remove the GLIBC21 code (!) - internal_free: make the code that tests against memtop and calls lesscore depend on USE_LESSCORE being defined, which it is by default lib/malloc/imalloc.h - USE_LESSCORE: define parse.y,shell.h - token_buffer_size and its corresponding saved value in the shell's parser state are now size_t instead of int stringlib.c - strsub,strcreplace: use size_t instead of int for local length and indexing variables lib/sh/zmapfd.c - zmapfd: use size_t instead of int for local length and indexing variables lib/sh/zgetline.c - zgetline: use size_t instead of int for local length and indexing variables 4/20 ---- pcomplete.c - init_itemlist_from_varlist: free VLIST after assigning it from *SVFUNC and after we get the variable names and values out of it. Report from Robert E. Griffith 4/25 ---- redir.c - here_document_to_fd: if the shell compatibility level is bash-5.0 or earlier, use tempfiles for all here-documents and here-strings. From a bug-bash discussion started by Sam Liddicott 4/26 ---- parse.y - parse_comsub: non-interactive shells exit on a syntax error while parsing the command substitution - parse_comsub: unset additional PARSER_STATE flags before calling yyparse(). Inspired by https://bugs.gentoo.org/837203; unsetting PST_COMPASSIGN is the fix for that bug - parse_string_to_word_list: use save_parser_state/restore_parser_state instead of saving pieces of the shell state in individual variables - parse_compound_assignment: use save_parser_state/restore_parser_state instead of saving pieces of the shell state in individual variables - parse_compound_assignment: unset additional PARSER_STATE flags before calling read_token(); set esacs_needed_count and expecting_in_token to 0 like in parse_comsub() since read_token can use them 4/27 ---- lib/sh/strvis.c - sh_charvis: changes to handle being compiled without multibyte support 4/29 ---- lib/readline/callback.c - rl_callback_read_char: don't set rl_eof_found unless eof is > 0, since it can be -3 if we need to read more input in a multi-key sequence. Report from Andrew Burgess examples/loadables/Makefile.sample.in - new file, containing the rules to build the example shared object - includes Makefile.inc from wherever it's installed. Suggested by Robert E. Griffith examples/loadables/Makefile.inc.in - remove rules that create the example shared object 4/30 ---- builtins/evalstring.c - parse_and_execute: check for terminating signals before returning, after any longjmp, to improve responsiveness and fix the -c code path before running any exit trap. Report from Emanuele Torre 5/17 ---- builtins/suspend.def - suspend_builtin: the -f option now forces a suspend even if job control is not enabled. Inspired by a discussion with Robert Elz doc/{bash.1,bashref.texi} - suspend: updated description to include expanded -f behavior 5/25 ---- builtins/mkbuiltins.c - -includefile: new argument, specifies extern filename to insert into the #include statement in the structfile (builtins.c) and the filename in the comment in the extern file (builtext.h). From Alexander Kanavin via https://savannah.gnu.org/patch/?10210 builtins/Makefile.in - builtins.c: change call to mkbuiltins to add -includefile option - builtins.c: change recipe to run all the commands in the same shell invocation - builtins.c: change recipe to specify new filenames in the call to mkbuiltins and move them onto builtins.c/builtext.h if the new ones are different; make the new filenames use the current make recipe shell pid in the filename ($$RECPID). Inspired by Alexander Kanavin via https://savannah.gnu.org/patch/?10210 6/2 --- builtins/common.c - builtin_find_indexed_array: new function, factored common code out of mapfile and read builtins to find an in-scope indexed array or create one builtins/common.h - builtin_find_indexed_array: extern declaration builtins/{mapfile,read}.def - change callers to use builtin_find_indexed_array variables.c - unbind_global_variable, unbind_global_variable_noref: new functions that remove variables from the global_variables table lib/sh/shmatch.c - sh_regmatch: use unbind_global_variable_noref to make sure we act on the copy of BASH_REMATCH in the global scope all the time, ignoring any local variables that might exist. Tentative fix for memory leak report from Emanuele Torre doc/{bash.1,bashref.texi} - BASH_REMATCH: add caveat about making it a local variable 6/6 --- print_cmd.c - print_redirection: if the redirectee for r_duplicating_output_word (r_duplicating_input_word) is 1 (0), don't print it; only print a non-default file descriptor number - print_redirection_list: remove the code that tries to temporarily translate a >&word redirection to >&word now that we won't print a non-default file descriptor number. Fixes issue with `declare -f' and function export reported by Namikaze Minato 6/13 ---- configure.ac - bumped version to bash-5.2-rc1 [bash-5.2-rc1 released] 6/15 ---- parse.y - parse_string_to_word_list: save the parser state before any state- changing functions like bash_history_disable(). Reported by Clark Wang 6/16 ---- doc/bash.1 - play tricks with the value of the zZ number register to refer to `bash(1)' instead of `above' or `below' when creating the builtins man page 6/17 ---- doc/{bash.1,bashref.texi} - wait: note that wait will return > 128 if interrupted by a signal. Reported by AA execute_cmd.c - {execute_cond_node,execute_arith_command,eval_arith_for_expr}: make sure to reset this_command_name after running any DEBUG trap so the DEBUG trap doesn't overwrite it. Reported by Emanuele Torre . - execute_select_command: set this_command_name to NULL after running any DEBUG trap like execute_for_command does 6/23 ---- test.c - three_arguments: when given [ ! ! arg ], make sure to advance POS after calling two_arguments to avoid a `too many arguments' error. Report from Steffen Nurpmeso 6/27 ---- subst.c - expand_word_internal: when expanding backquoted command substitution, call string_extract with the SX_REQMATCH flag (closing backquote required) only if the word flags don't contain W_COMPLETE, indicating that we're doing this for completion, probably to determine whether or not to append something to the word. Fixes bug reported by Emanuele Torre . 7/5 --- execute_cmd.c - execute_connection: treat a connector of '\n' the same as ';' print_cmd.c - print_comsub: new function, sets flag noting we are printing a command substitution and calls make_command_string - make_command_string_internal: add '\n' to the ';' case; print command list with newline connector appropriately parse.y - parse_comsub: call print_comsub instead of make_command_string - list1 production (part of compound_list): if a list is separated by newlines, and the parser is parsing a command substitution, make the connection command with a '\n' connector. Makes the text output of parse_comsub closer to the original source text. From a report from Martijn Dekker 7/6 --- doc/bash.1,lib/readline/doc/rluser.texi - complete: add note about arguments passed to command specified by `complete -C'; suggested by Mark Chandler builtins/setattr.def - show_local_var_attributes: special-case `local -', since there is no `declare -' equivalent. Reported by Emanuele Torre . - show_all_var_attributes: use `local -' when printing a variable named `-' at the current non-zero variable context parse.y - shell_getc: if we are at the end of an alias, returning a space, make sure we mark the previous character as single-byte by modifying shell_input_line_property so the space we return is properly recognized. This would fail before if the last character of the alias was a multi-byte character. Reported by Vangelis Natsios 7/12 ---- lib/readline/isearch.c - rl_display_search: don't call rl_redisplay_function before returning; rl_message already calls it. Reported by Frédéric Moulins configure.ac - bumped version to bash-5.2-rc2 7/18 ---- jobs.c - set_job_control: don't bother calling tcgetpgrp if shell_tty < 0, since it will just fail variables.c - reset_local_contexts: new function, delete all context tables associated with shell functions and set variable_context to 0. Called when we want to stop executing in a shell function without going through the pop_context chain with its side effects variables.h - reset_local_contexts: extern declaration builtins/evalstring.c - parse_and_execute: call reset_local_contexts instead of setting variable_context to 0 eval.c - reader_loop: call reset_local_contexts in cases where the shell has longjmped for a fatal error and errexit is enabled (ERREXIT), but not for other cases, and especially not for the exit builtin, instead of just setting variable_context to 0. Fixes issue originally reported by Robert Stoll subst.c - pat_subst: implement sed-like behavior when presented with a null pattern that's anchored at the start or end of the string, or when presented with a null string: process the replacement string for `&' and `\&' and substitute in the result as before. Patch from Koichi Murase 7/20 ---- [bash-5.2-rc2 frozen] 7/27 ---- parse.y - reset_parser: set need_here_doc, esacs_needed_count, expecting_in_token all to 0, since jumping back to a top-level parse needs that - parse_comsub: make sure to reset expand_aliases and shell_eof_token if we're not going to exit immediately out of this function 7/28 ---- parse.y - parse_comsub: if the compatibility level is <= 51, set extglob while parsing the command substitution, so bad pattern errors can still be caught but valid patterns are let through and can be evaluated at runtime, when extglob may have been set. If it isn't set, it will still be a parser error when the command substitution is executed. Fixes report from Sam James about gentoo scripts. - reset_parser: set extended_glob from global_extglob if the parser state includes PST_CMDSUBST - xparse_dolparen: set global_extglob but don't modify extended_glob, so parse errors can be caught before forking a child for command substitution, as part of word expansion, but after extglob may have been set by command execution (e.g., in a shell function) 8/5 --- [bump version to bash-5.2-rc3] 8/6 --- trap.c - run_pending_traps: move code from evalstring() so we call parse_and_execute() directly and handle any `return' invocations so we can restore the value of running_trap. Otherwise, if we longjmp past this function, we will think we're running a trap after we finish. Prompted by post from Koichi Murase 8/9 --- lib/readline/nls.c - _rl_current_locale: private variable, stores the value of the LC_CTYPE locale category, as determined by _rl_init_locale; set to allocated memory in _rl_init_locale() - _rl_set_localevars: new function, code from _rl_init_eightbit that checks the current locale (passed as an argument) and sets the various locale-aware variables based on it. It accepts a second argument: FORCE. If non-zero, it means to restore the default "C" locale values if the locale is "C" or "POSIX", now that this function can be called multiple times - _rl_init_eightbit: now just calls _rl_init_locale and _rl_set_localevars - _rl_reset_locale: new function, checks whether our the locale has changed since we last called _rl_init_locale to set our internal idea of its value. If it has changed, call _rl_set_localevars with the new locale and a FORCE argument of 1 to change the locale-dependent variables. lib/readline/rlprivate.h - _rl_reset_locale: extern declaration lib/readline/readline.c - rl_initialize: call _rl_reset_locale instead of _rl_init_locale so the internal readline variables get set when we move from a non- multibyte locale ("C") to a multibyte one ("en_US.UTF-8"). Report from Alan Coopersmith 8/16 ---- lib/sh/setlinebuf.c - sh_setlinebuf: allocate buffers for line-buffering stdout and stderr only once, the first time it is requested. Only allocate memory if we're using setvbuf (we usually are). Double the buffer size to 2016 if we're using the bash malloc. Otherwise, let stdio handle it. 8/17 ---- builtins/exec.def - exec_builtin: make sure to initialize orig_job_control in case the command is not found by search_for_command. Report and fix from Xiami [bash-5.2-rc3 frozen] 8/27 ---- parse.y - parse_comsub: restore extended_glob to a local copy (local_extglob) only if we changed it; a safer way to do it. Fixes extglob change issue reported by Kerin Millar - cond_term: restore extended_glob to a local copy; safer than using global_extglob, which we will reserve for error recovery 8/30 ---- parse.y - parse_comsub: don't clear the pushed string list; we might need it to consume additional input to satisfy this command substitution. When we restore the parser state, don't restore the pushed string list in case we used it. From https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1018727 - parse_comsub: don't modify extended_glob if parser_state includes PST_EXTPAT, in which case we've already set extended_glob and global_extglob appropriately. Only matters in compatibility mode. 8/31 ---- subst.c - parameter_brace_transform: make sure we return an error if *xform is '\0'. Report from Ivan Kapranov 9/7 --- [bump version to bash-5.2-rc4] 9/8 --- [bash-5.2-rc4 frozen] 9/20 ---- lib/readline/history.c - replace_history_entry: check for a NULL timestamp before trying to copy it. Report from nov.ondrej@gmail.com 9/23 ---- [bash-5.2 frozen] 9/26 ---- [bash-5.2 released] 9/27 ---- mailcheck.c - time_to_check_mail: don't bother if we're in the middle of readline completion or command dispatching. From https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1019462 lib/readline/display.c - rl_forced_update_display: use memset to clear out visible_line instead of trying to use a loop; line_size is the size. Report from print_cmd.c - indirection_level_string: use MBRLEN and an initialized mbstate_t object to avoid altering an internal mbstate_t. From Koichi Murase subst.c - string_extract_verbatim: use an explicit mbstate_t object and pass it to MBRLEN and mbrtowc to avoid altering the state that ADVANCE_CHAR is using. From Koichi Murase - setifs: use an explicit mbstate_t object and pass it to MBRLEN so to avoid altering an internal mbstate_t. From Koichi Murase builtins/printf.def - asciicode: use an explicit mbstate_t object and pass it to mbrtowc to avoid using an incorrect internal mbstate_t From Koichi Murase lib/sh/{mbscasecmp,mbscmp}.c - mbscasecmp,mbscmp: use different mbstate_t objects for the different strings so they don't affect each others' intermediate mbstate. From Koichi Murase 9/28 ---- execute_cmd.c - execute_command_internal: if executing a (command) subshell, restore the value of line_number from save_line_number before returning early due to being a non-terminal pipeline element. From https://savannah.gnu.org/support/index.php?110714 9/30 ---- subst.c - expand_array_subscript: make sure to pass a non-NULL first argument to sh_backslash_quote. Report from Emanuele Torre , patch from Koichi Murase 9/30 ---- jobs.c - cleanup_dead_jobs: delete dead foreground jobs we won't notify the user about (those not killed by a signal or killed by SIGINT/SIGPIPE). From a report from Koichi Murase - wait_for_any_job: never return a foreground job, even when requested by pid, if it's in the jobs list Report and fix from Koichi Murase m4/strtoimax.m4 - BASH_FUNC_STRTOIMAX: fix logic inversion of result; we should be replacing the function if the tests show we *don't* have a working version. Report from Emanuel Haupt 10/4 ---- lib/readline/nls.c - _rl_init_locale: cope with setlocale returning NULL, make sure we set up a default value for LC_CTYPE. Report from https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1021109 lib/readline/{input.c,rlprivate.h} - fixes for compiling on w64-mingw32 10/5 ---- parse.y - xparse_dolparen: save and restore extended_glob using local_extglob in the same way that parse_comsub does. Fixes compat issue reported in https://bugs.gentoo.org/873931 and by Kerin Millar 10/6 ---- subst.c - extract_dollar_brace_string: keep a stack of dolbrace_state values to handle nested expansions. Report from Antoine 10/7 ---- builtins/{shopt.def,common.h} - extglob_flag: new variable to hold the state of the `extglob' option; initialized to EXTGLOB_DEFAULT like extended_glob - shopt_set_extglob: new function to set extended_glob to the right value for the rest of the shell parse.y - reset_parser: set extended_glob to extglob_flag, which is only changed by shopt, remove global_extglob - parse_comsub,parse_cond_command,xparse_dolparen: no longer use global_extglob execute_cmd.c - execute_cond_node: reset extended_glob to the value of extglob_flag, since we're executing a command here 10/8 ---- parse.y - save_dstart: when we set the value of function_dstart, save the old value in save_dstart (read_token, read_token_word); restore it in the grammar production after calling make_function_def. This gives you correct line numbers for one level of function nesting. Report from Daniel Castro 10/10 ----- execute_cmd.c - shell_execve: rearrange code so that we check for a bad interpreter before printing a generic ENOENT error message. Report from Kirill Elagin 10/12 ----- lib/readline/callback.c - CALLBACK_READ_RETURN: add an inlined call to RL_CHECK_SIGNALS so we can handle any signals that arrived before we restored the calling application's signal handlers. From a gdb dicussion with Simon Marchi 10/12 ----- variables.c - pop_var_context: flag an internal error for shell_variables not pointing to a function context only if we haven't already flushed all the local contexts and reset variable_context. This can happen if errexit is enabled and we're going to be exiting the shell, but we're running unwind-protects on our way out. Report from Xavier Delaruelle 10/13 ----- sig.c - kill_shell: broke the code that resets the signal disposition to the default and sends a terminating signal to the shell into a separate function - termsig_handler: set handling_termsig to terminating_signal and make it file-scope so other functions know we're handling a terminating signal and are about to exit, and which signal it is (latter not used yet) - termsig_sighandler: if we get a fatal signal while we're handling a fatal signal, kill ourselves with the second fatal signal immediately. Fixes issue reported by Andreas Schwab execute_cmd.c - execute_case_command: call CHECK_TERMSIG after the call to strmatch, since gmatch will return FNM_NOMATCH if there's a pending terminating signal and we don't want incorrect results subst.c - pat_subst: make sure REP is non-NULL before calling savestring on it. Report from Justin Wood (Callek) 10/14 ----- builtins/{shopt.def,common.h} - expand_aliases: split into a variable that holds the current state of alias expansion (expand_aliases) and a variable that reflects the global option value (expalias_flag), make sure expand_aliases is set appropriately by shopt shell.c,execute_cmd.c,general.c - expand_aliases: make sure expand_aliases and expalias_flag always agree parser.h - PST_STRING: new parser flag, set when parsing a string to a command or word list parse.y - reset_parser: if we're parsing a command substitution or a string, and need to restore expand_aliases, make sure it's set to the value of expalias_flag. Fixes SIGINT during interactive command substitution parsing bug reported by feng xiangjun - parse_string_to_word_list,parse_string_to_command: make sure to set PST_STRING in parser_flags since we're resetting expand_aliases 10/17 ----- print_cmd.c - print_function_def, named_function_string: if the function doesn't have any here-documents, unset was_heredoc after printing the closing brace so we can add a separator if there is another command following the function definition. Fixes bug with multiple function definitions and here-documents reported by D630 - make_command_string_internal: if we're printing a command substitution, make sure to preserve newlines in the comsub so we avoid problems with re-parsing function defs or command substitutions containing multiple compound commands that need to be separated by newlines, but don't double them up parse.y - P_ARITH: new parse_matched_pair flags value; used to jump to parse_comsub if we see a $( inside a $(( )), $[ ], or (( )) - parse_matched_pair,parse_comsub,parse_arith_command,read_token_word: make sure to call parse_matched_pair with P_ARITH if we're parsing an arithmetic command or expansion - parse_matched_pair: if we see $( inside a call with P_ARITH in the flags argument, jump to parse_dollar_word and call parse_comsub to parse it. The rest of the steps assume that parse_comsub has expanded aliases (or not) appropriately. Fixes bug 2 from https://bugzilla.redhat.com/show_bug.cgi?id=2134307 10/18 ----- builtins/evalstring.c - parse_and_execute: we play tricks in the parser and in command_substitute to turn expand_aliases on and off depending on the comsub parser pass and whether or not we're in posix mode. Since this should only matter for parsing, when we're processing a command substitution, we set expand_aliases to the global flag value after parsing the command and before executing it, and restore it if the global flag value isn't changed during execution. Fixes bug 1 from https://bugzilla.redhat.com/show_bug.cgi?id=2134307 command.h - PF_BACKQUOTE: new flag value, only used by command_substitute to differentiate between `` and $() forms. The former does not have aliases expanded by parse_comsub in posix mode, so needs to expand them in parse_and_execute like default mode subst.c - command_substitute: don't modify expand_aliases if PF_BACKQUOTE is included in flags, since aliases haven't been expanded in the command string (it was run through parse.y:parse_matched_pair()) - expand_word_internal: pass PF_BACKQUOTE to command_substitute() if expanding a `` command substitution 10/20 ----- parse.y - yylex: return YYUNDEF as current_token if read_token returns < 0. Fixes parser reset issue reported by Todd Stein in https://savannah.gnu.org/support/index.php?110745 subst.c - ARITH_EXP_CHARS: chars that are special and trigger expansion in arithmetic expressions, EXP_CHARS without `<' and `>'. Fixes bug reported by Glenn Jackman execute_cmd.c - retain_fifos: replace executing_list (which is still present) as the indicator of whether or not to save and restore the FIFO list around a call to execute_command_internal/execute_command; it's more descriptive and can be used elsewhere - execute_for_command: set retain_fifos so we don't unlink the fifo list until the for command completes. Fixes issue reported in https://savannah.gnu.org/support/index.php?110743 lib/readline/xmalloc.c - memory_error_and_abort: add `const' qualifiers to the argument. Fix from Markus Elfring 10/24 ----- general.c - check_binary_file: check the first two lines for NULs if the first line begins with a `#!'; otherwise check the first. From a discussion and patch in https://savannah.gnu.org/support/?110744 (larsh@apache.org) parse.y - parse_matched_pair: set PST_NOERROR if we read to EOF without finding a closing match and call parser_error; avoids redundant error message lib/sh/uconvert.c - uconvert: RETURN: if ipart (integer part) is 0, but upart (fractional part) is non-zero, multiply upart by mult so we don't lose the sign for values in the range (-1, 0]. From a report by izabera execute_cmd.c - execute_in_subshell: check for terminating signals before we return to our caller, which will immediately exit, and before running any exit trap (since termsig_handler will run any exit trap). Fixes bug reported by Andrew Neff 10/26 ----- lib/readline/complete.c - rl_filename_completion_function: if the application doesn't supply any directory hook functions, we need to handle the case where we dequoted users_dirname and also tilde-expanded dirname. We choose to tilde expand users_dirname rather than call the application dequoting function again. Report and patch from Stefan H. Holek 10/27 ----- execute_cmd.c - execute_in_subshell: call procsub_clear in addition to clear_fifo_list, since none of these process substitutions are children of this new subshell jobs.c - wait_for_background_pids: call procsub_waitpid on the last procsub created as long as it's the same as $!, then call reap_procsubs to clean up the procsub list. Don't call procsub_waitall. Report from Oguz İsmail Uysal 10/31 ----- expr.c - expr_streval: explicitly permit `@' and `*' as associative array subscripts if the shell compatibility level is greater than 51. Prompted by report from Corey Hickey subst.c - string_list_pos_params, param_expand: make sure that $* is treated the same as "$*" in here-document bodies that are being expanded. From https://lists.gnu.org/archive/html/bug-bash/2022-09/msg00002.html 11/1 ---- lib/readline/readline.c - readline_internal_char: save and restore the value of _rl_top_level around our use of setjmp, even though it doesn't matter in most cases. Report from sparrowhawk996@gmail.com that it causes crashes in some obscure callback-mode cases lib/readline/isearch.c - _rl_isearch_dispatch: if we end the search and stuff characters back to be read again, decrement the current key sequence index by 2 so we don't have duplicate characters in the sequence lib/readline/misc.c - rl_digit_argument: call _rl_del_executing_keyseq after rl_execute_next 11/3 ---- builtins/setattr.def - show_localname_attributes: special-case "-" local variable, since there is no `declare -' equivalent. Reported by Emanuele Torre . 11/4 ---- lib/sh/tmpfile.c - sh_mktmpname,sh_mktmpfd: use get_urandom32() instead of random() if we're not using mktemp or mkstemp and the system provides the appropriate support trap.c - trap_variable_context: new variable, set to variable_context every time a trap string runs (that is, every time running_trap is set to a value > 0) in _run_trap_internal, run_exit_trap, run_pending_traps trap.h - trap_variable_context: extern declaration builtins/common.c - get_exitstat: if the `return' builtin is running, we are running a trap (but not the DEBUG trap), and the return would cause the trap string to complete (variable_context == trap_variable_context, so we haven't executed another shell function), use the last command exit value as the return status. POSIX interp 1602, from https://www.austingroupbugs.net/view.php?id=1602 11/4 ---- lib/readline/display.c - local_prompt_invis_chars: new array variable, similar to local_prompt_newlines, that keeps track of the number of invisible characters on each line of the prompt string. Use in WRAP_OFFSET macro with eye to using in W_OFFSET as well. Used for the case where the last line of the prompt is not the last line with invisible characters. Can use this in more calculations replacing wrap_offset and prompt_invis_chars_first_line going forward. Inspired by report from https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1018851 - update_line: use local_prompt_invis_chars to set _rl_last_c_pos correctly when on a non-terminal prompt line that contains invisible characters 11/7 ---- subst.c - get_var_and_type: allocate new memory for *valp and return it if the type is VT_ARRAYMEMBER. The callers -- all of which treat this case the same as VT_VARIABLE -- will free appropriately. Fixes problems with using parameter operations like ## on dynamic array variables, reported by Ivan Kapranov back on 8/31/2022 expr.c - change some error messages to make it clearer these are arithmetic syntax errors. These will require new translations. test.c - integer_expected_error: change the message to remove `expression', since this is only called for the `test' builtin, where TEST_ARITHEXP is not in FLAGS. This will require a new translation. Reported by Emanuele Torre back on 6/15/2022 11/10 ----- builtins/read.def - reset_timeout: when a timer expires, make sure to cancel any readline timeout as well with rl_clear_timeout(). From https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1023776 11/11 ----- builtins/reserved.def - !: add short help text lib/readline/complete.c - make_quoted_replacement: if readline found quote characters anywhere in the filename, but is not completing on an unclosed quoted string, and there is an application filename quoting function, let that function quote the filename if necessary bashline.c - COMPLETE_DQUOTE2: new completion quoting style define - bash_expand_filename: new function, runs a filename passed to it through word expansion, suppressing command and process substitution and returns the result (in new memory if it's different than the filename argument) - bash_quote_filename: if the return value from bash_check_expchar is `$' or ``', expand the filename and if it contains characters that should be quoted, set the quoting style to COMPLETE_DQUOTE2 - bash_quote_filename: if the completion quoting style is COMPLETE_DQUOTE2, use sh_mkdoublequoted to put double quotes around the entire unexpanded filename instead of using sh_double_quote, which will backslash-quote any `$'. This is a start at preserving user-supplied quotes if the (expanded) completed filename requires them. From reports by Peng Yu in 5/2020 and Pedro Gimeno in 6/2021 11/14 ----- parse.y - report_syntax_error: if word_top is set and we reach EOF, report the starting line number in the error message. From a patch from L A Walsh and Manuel Reiter back in 9/2018 - MAX_CASE_NEST: doubled to 256, renamed to MAX_COMPOUND_NEST - read_token_word: add `if', `while', and `until' to the set of compound commands that use WORD_TOP and WORD_LINENO (with corresponding changes to grammar productions) - [grammar]: changed check to decrement WORD_TOP to >= 0 since we start at -1 and we want to decrement back to -1 when all loops are closed builtins/jobs.def - jobs_builtin: call notify_and_cleanup after displaying the status of jobs to implement POSIX requirement that `jobs' remove terminated jobs from the jobs list https://pubs.opengroup.org/onlinepubs/9699919799/utilities/jobs.html#tag_20_62 11/15 ----- lib/sh/input_avail.c - include signal.h unconditionally, we need it for HAVE_SELECT and HAVE_PSELECT - nchars_avail: make sure we declare and use readfds and exceptfds if we have pselect or select available lib/readline/input.c - rl_gather_tyi,rl_getc: need to declare and use readfds and exceptfds if HAVE_PSELECT or HAVE_SELECT is set. Report from Henry Bent , fixes from Koichi Murase 11/16 ----- lib/sh/ufuncs.c - quit.h: include unconditionally for declaration of sigemptyset even if HAVE_SELECT is not defined lib/sh/timers.c - USEC_PER_SEC: make sure it's defined even if HAVE_SELECT is not lib/glob/sm_loop.c - BRACKMATCH: if an equivalence class does not match, and the next character following the class is a `]', treat that as the end of the bracket expression. Report and fix from Koichi Murase - GMATCH: if the current character in the string is a `/' and the current element in the pattern is a bracket expresion, and the FLAGS include FNM_PATHNAME, return FNM_NOMATCH immediately. A bracket expression can never match a slash. Report and fix from Koichi Murase - BRACKMATCH: if we encounter a in a bracket expression, either individually or as part of an equivalence class, nullify the bracket expression and force the `[' to be matched as an ordinary character 11/17 ----- lib/glob/sm_loop.c - BRACKMATCH: if a slash character appears as the first character after a non-matching character class or equivalence class, treat the bracket as an ordinary character that must be matched literally - BRACKMATCH: if a slash character appears as the second character of a range expression, treat the bracket as an ordinary character - BRACKMATCH: if a slash character appears in the portion of a bracket expression that already matched, treat the bracket as an ordinary character Updates from Koichi Murase - BRACKMATCH: if a range expression is incomplete (no end char), treat the bracket as an ordinary character 11/18 ----- doc/bashref.texi - Reporting Bugs: add mention of the Savannah project page. Suggested by Loïc Yhuel 11/20 ----- builtins/common.h - SEVAL_NOOPTIMIZE: new flag for parse_and_execute: means don't try to optimize forks out of any simple or conditional commands builtins/evalstring.c - parse_and_execute: if FLAGS includes SEVAL_NOOPTIMIZE, don't try to call can_optimize_connection to optimize away forks from AND_AND or OR_OR commands builtins/eval.def,trap.c,parse.y,jobs.c - parse_and_execute: include SEVAL_NOOPTIMIZE in any calls to parse_and_execute. Fixes bug reported by Frode Nordahl 11/21 ----- lib/readline/readline.c - readline_initialize_everything: use xmalloc to initialize rl_executing_keyseq, since we use xrealloc to reallocate it and don't check it for NULL anywhere subst.c - parameter_brace_expand_word: make sure to update *ESTATEP and pass ES back to the caller even if array_value returns NULL. Fixes memory leak referencing unset associative array elements reported by Ralf Oehler ; fix from Koichi Murase 11/22 ----- lib/glob/sm_loop.c - BRACKMATCH: if a backslash appears before a slash in a bracket expression, treat the whole thing as literal and try to match the open bracket - BRACKMATCH: if a bracket expression ends in an unquoted backslash, try to match the left bracket literally and go on. More discussion with Koichi Murase execute_cmd.c - execute_in_subshell: don't try to optimize away forks in the command if the subshell command is being timed or its return value is being inverted. Report from Dabrien 'Dabe' Murphy 11/23 ----- lib/readline/complete.c - __WIN32__: remove uses of this preprocessor symbol, use _WIN32 lib/readline/rldefs.h - Don't define NEW_TTY_DRIVER on _WIN32. Fix from sparrowhawk996@gmail.com include/posixselect.h - include "posixtime.h" instead of so the right guards are present builtins/read.def - read_builtin_timeout: make sure we don't try to call shtimer_select if HAVE_SELECT isn't define; it won't be compiled in include/stdc.h - restrict: make sure we have a definition if __STDC__ is not defined include/posixtime.h,lib/sh/gettimeofday.c - gettimeofday: make sure the function signature matches POSIX by adding `restrict' 11/25 ----- include/posixselect.h - don't try to define HAVE_SELECT at all if _WIN32 is defined lib/readline/kill.c - define WIN32_LEAN_AND_MEAN for the windows-specific paste function lib/readline/rlprivate.h - _rl_timeout_select: only include declaration if HAVE_SELECT or HAVE_PSELECT are defined 11/28 ----- lib/readline/input.c - rl_gather_tyi: call rl_input_available_hook after FIONREAD test, let the application hook into whether or not there are characters to read execute_cmd.c - execute_function: if the maximum function nesting level is exceeded, run the unwind-protects before jumping back to top level. Report from felix variables.c - find_variable_nameref_for_create,find_variable_nameref_for_assignment: handle find_variable_nameref returning &nameref_maxloop_value, return NULL and print a warning message in that case - find_variable_nameref: return &namref_maxloop_value instead of NULL if the number of resolution cycles exceeds NAMEREF_MAX; change callers to print a new error message in this case; no other change in behavior. From a 4/2020 suggestion from Andrej Podzimek - find_variable_nameref_context, find_nameref_at_context, find_variable_last_nameref_context: make sure to return &nameref_maxloop_value consistently and handle getting it as a return value from other functions 11/29 ----- lib/readline/display.c - _rl_update_final: if there is only one line (_rl_vis_botlin == 0) and that line is empty (botline_length == 0), assume there is no prompt and the line has no contents, so output the CR/LF to indicate that the newline has been read. From a report from Kevin Pulo 11/30 ----- builtins/evalstring.c - parse_and_execute: if we are executing the eval builtin, and the return status from the command is suppressed (builtin_ignoring_errexit), set CMD_IGNORE_RETURN in the parsed command's flags. From a report from Tycho Kirchner 12/1 ---- lib/glob/sm_loop.c - PARSE_SUBBRACKET: a reworked PARSE_COLLSYM, generalized to handle [:, [=, and [. special bracket expressions - BRACKMATCH: change to use PARSE_SUBBRACKET consistently to parse [:, [=, and [. special bracket expressions - PATSCAN: takes new FLAGS arg, changed all callers to just pass the flags they get through to PATSCAN - PATSCAN: call PARSE_SUBBRACKET for [:, [=, and [. to treat them consistently (uses FLAGS argument here) - PATSCAN: handle FNM_NOESCAPE appearing in FLAGS argument This set of fixes contributed by Koichi Murase lib/glob/{glob,gmisc}.c - PATSCAN: change all callers to add extra flags arg of 0 doc/{bash.1,bashref.texi} - trap description: replace ARG with ACTION - clarify the language describing what trap without arguments and trap -p without arguments display - clarify that trap in non-interactive shells does not allow signals that were ignored at shell start to be trapped; interactive shells allow this - trap: add missing command types to the description of when the DEBUG trap is executed builtins/trap.def - LONG_DOC: replace ARG with ACTION - make it clear that trap and trap -p without arguments display a information about specified signals, but in a form that may be reused as shell input 12/5 ---- pcomplete.c - compgen_compspec: new flag saying whether or not we're running via the compgen builtin or via programmable completion: set to 0 every time gen_progcomp_completions runs, set to 1 in gen_compspec_completions if this_shell_builtin == compgen_builtin - gen_shell_function_completions: call rl_clear_signals and rl_set_signals so the shell's signal handlers are active when running shell functions to generate completions. We try to do it only when programmable completion, not the compgen builtin, is active. Idea from Koichi Murase , fixes bug reported in 9/22 sig.c - sigint_sighandler, termsig_sighandler: since these can now be called during programmable completion, we need to set the readline signal event hook if the readline state includes RL_STATE_DISPATCHING or RL_STATE_COMPLETING, since we can't rely on RL_STATE_SIGHANDLER being set and we want to call throw_to_top_level here - throw_to_top_level: if the shell is interactive, before calling bashline_reset (which zeroes rl_signal_event_hook), call rl_cleanup_after_signal if we think readline hasn't called it already 12/9 ---- configure.ac - add "LOCAL_LDFLAGS = -rdynamic" for freebsd/openbsd/netbsd. Reported by Christian Weisgerber 12/12 ----- subst.c - parameter_brace_expand_length: handle namerefs with values that are valid length expansion expressions but invalid identifiers. From ed7-aspire4925@hotmail.com via https://savannah.gnu.org/support/?110799 12/13 ----- subst.c - extract_heredoc_dolbrace_string: fix off-by-one error after calling extract_command_subst and extract_process_subst that caused it to copy one too many parsed characters. Fix for bug reported by Norbert Lange 12/14 ----- execute_cmd.c - execute_cond_node: if a regular expression fails to compile, print an error message. Report from Emanuele Torre back on 6/15/2022 trap.c - trap_variable_context -> trap_return_context, initialize from funcnest + sourcenest instead of variable_context so we handle shell function execution and `./source', both of which can use `return'. Idea from Koichi Murase builtins/common.c - get_exitstat: compare trap_return_context against funcnest+sourcenest, since that's how it's initialized now lib/readline/readline.c - readline_internal_charloop: if we're not using the callback interface, don't restore _rl_top_level from olevel, since we will just be going around the loop again and will potentially need to use it multiple times. Report from Emanuele Torre 12/15 ----- shell.h - EX_UTILERROR: new generic special builtin return status to indicate a POSIX utility error that should cause a non-interactive shell to abort execute_cmd.c - builtin_status: translate EX_UTILERROR to EXECUTION_FAILURE builtins/set.def - unset_builtin: return EX_UTILERROR if posix_utility_error is set; set it when trying to unset a non-identifier (variable) or a non-unsettable or readonly variable 12/16 ----- subst.c - de_backslash: now takes a second argument with the current quoting flags - de_backslash: if the quoting flags include Q_HERE_DOCUMENT and the shell is in posix mode, remove backslashes quoting double quotes subst.h - de_backslash: update extern declaration lib/readline/histexpand.c - history_expand_internal,get_history_word_specifier,get_subst_pattern, hist_error,history_find_word,hist_string_extract_single_quoted: now take const char * string arguments lib/readline/mbutil.c - _rl_get_char_len,_rl_adjust_point,_rl_find_next_mbchar_internal, _rl_find_next_mbchar,_rl_find_prev_mbchar,_rl_find_prev_mbchar_internal, _rl_test_nonzero,_rl_find_prev_utf8char,_rl_is_mbchar_matched, _rl_compare_chars,_rl_char_value: take const char * string arguments 12/17 ----- builtins/umask.def - parse_symbolic_umask: add missing POSIX pieces: o `action' of ugo, meaning to copy portions of initial mask o multiple `op' specs as part of the action string (`u=r-w') (resets perm) o missing perm characters Xst in action string o default `who' equivalent to `a' instead of fixing up later 12/19 ----- builtins/type.def - if -a and -P are both supplied, look in the command hash table but continue and perform a $PATH search even if the NAME is hashed. From a report by Adam Vodopjan doc/{bash.1,bashref.texi} - type: update description to fix it to what the code actually does builtins/set.def - set_edit_mode: don't run with_input_from_stdin or with_input_from_stream unless command_execution_string is NULL. Report from Harald van Dijk 12/20 ----- braces.c - array_concat: add a check for interrupts and terminating signals into the inner loop jobs.c - without_job_control: set original_pgrp == NO_PID since we don't want to be messing with the terminal pgrp if we call end_job_control. From a report from ks1322 ks1322 12/21 ----- include/{posixtime.h,systimes.h,shmbutil.h,shtty.h} - PARAMS: remove 12/22 ----- builtins/evalfile.c - _evalfile: add SEVAL_NOOPTIMIZE to the parse_and_execute flags (maybe overkill, but safer). From a report by Dan Church and Oguz lib/sh/gettimeofday.c - gettimeofday: added code for _WIN32 version 12/27 ----- doc/{bash.1,bashref.texi},builtins/printf.def - specify the valid format characters as a subset of what printf(3) specifies, note that the %n format specifier takes a variable name as an argument. Suggestion from Emanuele Torre 12/28 ----- {xmalloc,variables,unwind_prot,trap,test,subst,sig,siglist,shell,redir, pcomplete,pathexp,make_cmd,mailcheck,jobs,input,hashcmd,hashlib,general, flags,findcmd,externs,execute_cmd,error,dispose_cmd,command,bashline, bashhist,assoc,arrayfunc,array,alias}.h {version,version2}.c - PARAMS: remove for ANSI C-ification builtins/{getopt,common,bashgetopt}.h builtins/{mkbuiltins,gen-helpfiles,psize,psize-posix}.c builtins/{evalstring,bashgetopt,common,evalfile,getopt}.c builtins{alias,bind,break,cd,command,complete,declare,enable,exit,fc, fg_bg,getopts,hash,help,history,jobs,kill,mapfile,printf,pushd,read,set, shopt,source,suspend,trap,type,ulimit,umask,wait}.def - PARAMS: remove for ANSI C-ification lib/malloc/{watch,table,shmalloc,mstats,imalloc}.h lib/malloc/{malloc,stats,table,trace}.c - PARAMS: remove for ANSI C-ification lib/glob/{glob,strmatch}.h lib/glob/{glob,glob_loop,gmisc,sm_loop,smatch,strmatch,xmbsrtowcs}.c - PARAMS: remove for ANSI C-ification lib/intl/{eval-plural,plural-exp,loadinfo,hash-string,gettextP}.h, lib/intl{plural.y,plural.c}, lib/intl/{finddomain,bindtextdom,dcigettext,l10nflist,loadmsgcat,localealias,plural-exp}.c - PARAMS: remove for ANSI C-ification - STATIC -> static - internal_function -> inline lib/sh/{casemod,clock,eaccess,fnxform,fpurge,makepath,mktime,netopen,pathphys,random,shquote,snprintf,spell,strtod,strtoimax,strtoumax,times,timeval,tmpfile,unicode,winsize,zcatfd,zgetline,zmapfd}.c - PARAMS: remove for ANSI C-ification support/{bashversion.c,mkversion.sh,mksignames.c} - PARAMS: remove for ANSI C-ification lib/sh/{clock,strchrnul,getcwd,times,rename,inet_aton,strdup,strerror,strnlen strpbrk,strtod,strtoimax,strtol,strtoumax,wcsdup,wcsnwidth,wcswidth,strcasecmp, strcasestr}.c - C89-style function declarations lib/malloc/{alloca,malloc,stats,table,trace,xmalloc}.c - C89-style function declarations 12/29 ----- lib/glob/{glob,glob_loop,gmisc,sm_loop,smatch,strmatch,xmbsrtowcs}.c - C89-style function declarations lib/sh/{zcatfd,zgetline,zmapfd,zread,zwrite,winsize,vprint,uconvert,ufuncs, unicode,utf8}.c lib/sh/{stringlist,stringvec,strtrans,strvis,timeval,tmpfile}.c lib/sh/{setlinebuf,shmatch,shmbchar,shquote,shtty,snprintf,spell}.c lib/sh/{random,pathcanon,pathphys,oslib,netconn,netopen}.c lib/sh/{mailstat,makepath,mbscasecmp,mbschr,mbscmp,mktime}.c lib/sh/{input_avail,itos,getenv,fmtulong,fnxform,eaccess,dprintf,casemod,clktck}.c - C89-style function declarations examples/loadables/{accept,asort,basename,bcat,cat,csv,cut,dirname,dsv}.c examples/loadables/{fdflags,finfo,head,hello,id,lcut,ln}.c examples/loadables/{mkdir,mkfifo,mktemp,mypid,necho,pathchk}.c examples/loadables/{print,printenv,push,realpath,rm,rmdir,seq}.c examples/loadables/{setpgid,stat,strftime,sync,tee,template,truefalse,tty}.c examples/loadables/{uname,unlink,whoami}.c - C89-style function declarations xmalloc.c, array2.c, list.c, dispose_cmd.c, copy_cmd.c, bracecomp.c, mksyntax.c - PARAMS: remove for ANSI C-ification - C89-style function declarations 12/30 ----- hashlib.c,assoc.c,input.c,hashcmd.c,alias.c,alias.h,array.c,flags.c,pcomplib.c, pcomplib.h,locale.c variables.c,arrayfunc.c,bashhist.c,bashline.c,error.c,pathexp.c,nojobs.c, stringlib.c,eval.c,findcmd.c,mailcheck.c,make_cmd.c,shell.c,print_cmd.c, general.c,expr.c - PARAMS: remove for ANSI C-ification - C89-style function declarations - add `const' and `inline' where it makes sense 12/31 ----- test.c,execute_cmd.c,parse.y,pcomplete.c,trap.c,subst.c,braces.c,jobs.c,sig.c builtins/{evalstring.c,gen-helpfiles.c,mkbuiltins.c} builtins/{alias.def,bind.def,break.def,builtin.def,caller.def,cd.def,colon.def} builtins/{command.def,complete.def,declare.def,echo.def,enable.def,eval.def} builtins/{exec.def,exit.def,fc.def,fg_bg.def,hash.def,help.def,history.def} builtins/{jobs.def,kill.def,let.def,mapfile.def,printf.def,pushd.def,read.def} builtins/{return.def,set.def,setattr.def,shift.def,shopt.def,source.def} builtins/{suspend.def,test.def,times.def,trap.def,type.def,ulimit.def} builtins/{umask.def,wait.def} - PARAMS: remove for ANSI C-ification - C89-style function declarations - add `const' and `inline' where it makes sense 1/2/2023 -------- lib/sh/{stringvec.c,inet_aton.c,strnlen.c,spell.c,netopen.c,stringlist.c} lib/sh/{strtrans.c,tmpfile.c,uconvert.c,zcatfd.c} builtins/{common.c,common.h} mailcheck.c,bashline.c,pathexp.c,stringlib.c,locale.c,pcomplete.c,redir.c test.c,trap.c,variables.c,bashhist.c,expr.c,arrayfunc.c,assoc.c,variables.c subst.c,findcmd.c arrayfunc.h,variables.h,findcmd.h,bashline.h,trap.h,pathexp.h,externs.h, pcomplete.h,bashhist.h,assoc.h,general.h,subst.h - more `const' changes, remove `register' 1/3 --- parse.y - more `const' changes, remove `register' variables.h - VARLIST: list_len and list_size now size_t variables.c - vlist_alloc: take size_t parameter, changed callers - vlist_realloc: take size_t parameter, changed callers lib/sh/stringvec.c,externs.h - functions that take int size params now take size_t - strvec_len: returns size_t lib/sh/stringlist.c,externs.h - STRINGLIST: list_size and list_len now size_t - functions that take int size params now take size_t variables.c,bashline.c,builtins/bind.def - change strvec_ int parameters to size_t where it makes sense - change strlist_ int parameters to size_t where it makes sense builtins/shopt.def - reset_shopt_options: set glob_always_skip_dot_and_dotdot to 1, since that's the current default in lib/glob/glob.c 1/4 --- execute_cmd.c,parse.y,general.c,make_cmd.c,print_cmd.c,variables.c,expr.c jobs.c,subst.c,input.c,unwind_prot.c,pathexp.c,pathexp.h,alias.c,bashline.c bashhist.c,lib/sh/shquote.c,externs.h,stringlib.c,locale.c,findcmd.c builtins/common.c,builtins/declare.def,builtins/enable.def,builtins/fc.def lib/sh/zmapfd.c,builtins/help.def,builtins/mapfile.def,builtins/read.def builtins/shopt.def,lib/glob/glob.c,lib/glob/smatch.c,lib/glob/sm_loop.c lib/glob/xmbsrtowcs.c,lib/glob/gmisc.c,lib/glob/gm_loop.c lib/sh/getenv.c,lib/sh/makepath.c,lib/sh/pathphys.c,lib/sh/tmpfile.c lib/sh/stringlib.c,lib/sh/spell.c,lib/sh/strtrans.c,lib/sh/zgetline.c lib/sh/mbscasecmp.c,lib/sh/utf8.c lib/readline/readline.c,lib/readline/rlprivate.h,lib/readline/vi_mode.c lib/readline/complete.c,lib/readline/bind.c,lib/readline/isearch.c lib/readline/util.c,lib/readline/kill.c,lib/readline/macro.c lib/readline/text.c,lib/readline/histexpand.c,lib/readline/histfile.c lib/readline/mbutil.c,lib/tilde/tilde.c - size_t changes - remove some unused variables - remove `register' 1/5 --- builtins/getopt.c - sh_getopt_restore_state: make sure argv[sh_curopt] is non-null before using it parse.y - CHECK_FOR_RESERVED_WORD: make sure to set word_top appropriately if the last token was IF, WHILE, or UNTIL, since these tokens can follow those reserved words Makefile.in - ubsan: new target, to build with gcc/clang UBSan undefined behavior sanitizer 1/6 --- execute_cmd.[ch],locale.c,parse.y,expr.c,bracecomp.c,bashline.c,externs.h stringlib.c,general.[ch] builtins/{subst.c,array.c,echo.def,printf.def,enable.def,help.def,set.def} lib/sh/{strtrans.c,stringvec.c,shquote.c} lib/glob/sm_loop.c lib/readline/{vi_mode.c,histfile.c,funmap.c,complete.c,display.c,bind.c,isearch.c} lib/tilde/tilde.c - final code cleanups for ANSI C and lint/ubsan/asan runs 1/7 --- lib/readline/complete.c - rl_complete_internal: when computing nontrivial_lcd and the length of the text from the line buffer, dequote the text from the line buffer before comparing it against the common prefix of the matches (matches[0]), so we don't get spurious mismatches if the quoted text from the line is longer than the unquoted match. Report from Martin Castillo input.[ch] - get_buffered_stream(fd): new function, return the BUFFERED_STREAM * corresponding to file descriptor FD, if it exists 1/8 --- parse.y - shell_getc: if yy_getc returns EOF but the buffered stream input indicates an error state, set shell_input_line_terminator to READERR. Currently treated the same as EOF. 1/10 ---- lib/sh/random.c - intrand32: use % operator instead of (mathematically equivalent) subtraction and multiplication, which can cause signed 32-bit overflow. Report from Sam James lib/readline/input.c - rl_getc: if the application hasn't defined a signal event handler, and readline is in callback mode, make sure SIGINT breaks out of operations like incremental and non-incremental searches and digit arguments by calling _rl_abort_internal, since the callback code expects the various contexts that rl_callback_sigcleanup() resets to be valid - rl_getc: if there is a pending signal when we enter the while loop in callback mode, handle it in the same way we would if a signal arrived while we were waiting in select/pselect. That forces callback mode to call _rl_abort_internal in the right place. From a report from Andrew Burgess 1/13 ---- trap.c - SIG_ASYNCSIG: new value for sigmodes; means that the signal is ignored because an async command is being executed - set_signal: if the original signal disposition is SIG_IGN only return if SIG_ASYNCSIG isn't set. This implements POSIX interp 751 for setting a trap - ignore_signal: if we're ignoring a signal that already has SIG_IGNORED set, make sure we set SIG_TRAPPED in case of SIG_ASYNCSIG (POSIX interp 751) - reset_signal,restore_signal: if we want to reset or restore a signal that has SIG_ASYNCSIG set, make sure we reset to SIG_DFL if the signal is trapped to maintain POSIX semantics - set_signal_async_ignored: new function to set original_signals to SIG_IGN and set the SIG_IGNORED and SIG_ASYNCSIG sigmode flags; used by setup_async_signals - signal_is_async_ignored: return true if the SIG argument is being ignored because it's part of an async list trap.h - set_signal_async_ignored,signal_is_async_ignored: extern declarations sig.c - initialize_terminating_signals: if a signal is ignored because it's part of an async list, don't bother trying to initialize or set the hard ignored flag execute_cmd.c - execute_simple_command: if we're executing a builtin or function asynchronously, call reset_terminating_signals like we do in execute_in_subshell() - execute_subshell_builtin_or_function: only call set_sigint_handler if async is 0, so we don't undo the work done by setup_async_signals. If the old handler is SIG_IGN and the signal has SIG_ASYNCSIG set and isn't trapped, restore the old handler - setup_async_signals: call set_signal_async_ignored to set the right flags 1/16 ---- parse.y - read_token_word: if we're in posix mode, and we find a reserved word before checking a token for aliases, turn off the `check next word for aliases' flag (PST_ALEXPNEXT). This is for compatibility with other shells. From a report by - shell_getc: make sure the shell is reading a buffered stream before checking the default buffered stream for errors if yy_getc returns 0 1/17 ---- subst.c - split_at_delims: if the cursor is at whitespace just preceding a word, and we're doing this for completion (e.g., to build COMP_WORDS), force the creation of an empty word and set the current word to it. This is a change from previous behavior, but a useful one: it makes COMP_WORDS[COMP_CWORD] and $2 to a completion function consistent and the same as the word readline wants to complete. From a report from Naim Favier back in 6/2022 parse.y - read_token: turn off PST_ALEXPNEXT if we read a shell metacharacter: we will expand aliases anyway if the token puts the shell in a command word position and we took pains to turn it off for redirection tokens; remove special cases that turned it off - pop_string: don't turn off PST_ALEXPNEXT if it's on; this is what prevents it from being enabled if an alias expansion of more than two aliases ends with a space. Fixes issue reported by 1/18 ---- jobs.c - wait_for: if we are reaping a background job, don't bother trying to give the terminal back to shell_pgrp -- we never set the terminal's pgrp to that job in the first place. Fixes issue reported by Steffen Nurpmeso 1/19 ---- lib/readline/{histexpand.c,history.h} - history_expand: first argument is now `const char *'. Request from Simon Marchi lib/readline/doc/{history.3,hstech.texi} - history_expansion: change description to note new const first arg 1/21 ---- lib/readline/input.c - rl_gather_tyi: move call to rl_input_available_hook to the top, giving it first shot. There's still no way for it to set chars_avail. Conditionalize the rest of the input methods based on whether the result == -1. lib/readline/rl_private.h - move the decision making about RL_TIMEOUT_USE_SELECT/RL_TIMEOUT_USE_SIGALRM here from input.c 1/23 ---- trap.c - restore_default_signal: if we have a signal that's not trapped, but is a signal that is set to SIG_ASYNCSIG, POSIX interp 751 requires the shell to allow it to be reset to the default, even though it wasn't already trapped subst.c - do_assignment_internal: don't allocate new memory for NAME, just modify and restore it in place 1/25 ---- lib/readline/histfile.c - history_write_slow: a fallback function that uses stdio to write the history list to a supplied file descriptor - history_do_write: call history_write_slow if ftruncate/mmap/malloc fail; hope the simpler approach works 1/27 ---- general.[ch] - valid_function_name: new function, returns non-zero if the name can be used as a function name without the `function' prefix print_cmd.c - named_function_string,print_function_def: use valid_function_name so the rules about when to use the `function' reserved word are consistent builtins/trap.def - new option -P, which prints only the action associated with each signal spec argument doc/{bash.1,bashref.texi} - trap: document new -P option, add note that trap -p or trap -P in a subshell can print the parent's traps 1/28 ---- parse.y - shell_getc: make sure references to shell_input_line_property are protected by #ifdef HANDLE_MULTIBYTE. From https://savannah.gnu.org/patch/?10309 print_cmd.c - named_function_string,print_function_def: don't copy the function COMMAND * before printing it as text; unwind-protect any redirects attached to the function body parse.y - PST_CMDBLTIN: new parser state, means the previous token word was `command' in a command position - read_token_word: in posix mode, "command" followed by a declaration utility preserves the declaration utility status of the following word, so we mark that state and then check a word for being an assignment builtin if in that state on the next call. This makes compound assignments to builtins that accept them (e.g., `declare') work as if `command' were not present. This is a POSIX issue 8 requirement - read_token: turn off PST_CMDBLTIN state as appropriate subst.c - use locale_mb_cur_max instead of MB_CUR_MAX consistently 2/1 --- builtins/printf.def - printf_builtin: %q and %Q use the `alternate form' flag to force single quoting instead of backslash quoting. $'...' is still preferred if there are characters that require it - getwidestr, getwidechar: functions to take a possibly multibyte character string and convert to a wide character string or a wide character, respectively - convwidestr, convwidechar: functions to take a wide character string or single character, apply any precision, convert back to a multibyte character string, and return the result for printing - printf_builtin: consistently print an error message if printstr() returns < 0, since it will only do that if ferror(stdout) is true; let the PRETURN macro do that if appropriate, so we don't duplicate code 2/2 --- builtins/printf.def - printwidestr: wide-character version of printstr; understands fieldwidth and precision as characters instead of bytes - printf_builtin: interpret %ls and %lc as referring to wide strings and characters, respectively; use printwidestr to print the wide- character string obtained from the (presumably multibyte) argument. Fieldwidth and precision are characters, not bytes 2/3 --- builtins/set.def - unset_builtin: since tokenize_array_reference modifies its NAME argument, we need to restore the original word if there is no array variable found with that name, so we can go ahead and try to unset a function with that wonky name. Inspired by a report from Emanuele Torre parse.y - read_token_word: don't set the W_HASDOLLAR flag in the returned WORD for <(command) and >(command). That way they can be used in function names. execute_cmd.c - execute_simple_command: if in posix mode, don't perform function lookup for command names that contain a slash builtins/enable.def - dyn_load_builtins: don't allow dynamically loaded builtins to have slashes in their name 2/4 --- general.c - valid_function_word: new function, used to check whether a word can be used as a function name. Replaces call to check_identifier(); incorporates POSIX interp 383 check for special builtin so we can move it out of execute_intern_function(). - valid_function_name: use the FLAGS argument instead of POSIXLY_CORRECT; only reject reserved words if FLAGS&1 execute_cmd.c - execute_intern_function: call valid_function_word() instead of check_identifier(); remove check for special builtin that valid_function_word performs print_cmd.c - print_function_def,named_function_string: pass POSIXLY_CORRECT to valid_function_name as the FLAGS argument 2/6 --- subst.c - expand_subscript_string: don't set W_NOTILDE in td.flags so we perform `normal' tilde expansion (expansion at the start of the word) on the array subscript - expand_oneword,expand_compound_assignment_word: change second arg to ATYPE, which better reflects its use arrayfunc.c - assign_assoc_from_kvlist,assign_compound_array_list,expand_and_quote_assoc_word: use expand_assignment_string_to_string instead of expand_subscript_string for the rhs of the assignment or the second word in the kvpair to be more consistent with how assignment statements behave (e.g., tilde expansion). Fixes issue reported by maroloccio@gmail.com 2/7 --- parse.y - discard_until: now returns the last character read. If we get an EOF while processing a comment, return EOF (rarely happens; shell_getc returns EOF only under certain circumstances) - read_token: if discard_until returns EOF, return yacc_EOF input.h - bclearerror: new macro, clears the B_ERROR flag in the buffered stream BP input.c - b_fill_buffer: try making errors `sticky': if the B_ERROR flag is set for the buffered stream, return EOF immediately. This is similar to what stdio does builtins/fc.def - fc_builtin: in posix mode, if there are extra arguments supplied to -s (e.g. pat=rep first last), throw an error doc/{bash.1,bashref.texi} - printf: document new altform for %q/%Q and %ls/%lc 2/8 --- bashline.c - set_up_new_line: if the new line doesn't change rl_line_buffer, just return right away without changing rl_point. Affects alias-expand-line, history-expand-line, and history-and-alias-expand-line. Inspired by a report from Addison Brendtro builtins/declare.def - declare_internal: make multiple calls to `local -' at the same context have no effect after the first one. That way we don't keep overwriting the saved option set. Report and fix from Emanuele Torre 2/10 ---- execute_cmd.[ch] - bind_lastarg: now a public function parse.y - execute_variable_command: call bind_lastarg instead of bind_variable to avoid exporting $_ if allexport is set. Fix from Emanuele Torre - history_delimiting_chars: if we're parsing some kind of delimited construct that's *not* a quoted string, don't bother adding an extra newline to the history if the current history entry already ends in a newline. From https://savannah.gnu.org/support/?110838 via Ganapathi Kamath 2/13 ---- bashline.c - set_up_new_line: add some more heuristics to try and usefully position rl_point 2/16 ---- lib/readline/complete.c - complete_fncmp: fix for case mapping `-' and `_' if the bytes do not form a valid multibyte character lib/readline/util.c - new functions: _rl_strcaseeqn, like strcasencmp but with a flag to handle character mapping (-_) like completion-map-case wants, and _rl_charcasecmp, to do the same thing for two characters. Inspired by a patch from Brennan Vincent lib/readline/mbutil.c - new functions: _rl_mb_strcaseeqn and _rl_mb_charcasecmp, multibyte versions of the above Inspired by a patch from Brennan Vincent lib/readline/{rlprivate.h,rlmbutil.h} - extern declarations for the above functions lib/readline/complete.c - complete_fncmp: use _rl_mb_strcaseeqn and _rl_strcaseeqn as appropriate, remove duplicated inline code Inspired by a patch from Brennan Vincent lib/readline/{isearch.c,rlprivate.h} - _rl_search_case_fold: new variable, will be used to indicate case- insensitive incremental and non-incremental searches lib/readline/bind.c - search-ignore-case: new bindable variable, tied to _rl_search_case_fold Inspired by a patch from Brennan Vincent lib/sh/mbsncmp.c - mbsncmp(3) replacement lib/readline/isearch.c - _rl_isearch_dispatch: honor `search-ignore-case' variable and perform case-insensitive incremental searches if it's set Inspired by a patch from Brennan Vincent lib/readline/histlib.h - CASEFOLD_SEARCH: new flag for history searching, means to search strings case-insensitively lib/readline/histsearch.c - history_search_internal: anchored searches honor CASEFOLD_SEARCH lib/readline/histsearch.c - history_search_internal: searches now honor CASEFOLD_SEARCH - _hs_history_search: new function, just external interface to history_search_internal, allows the rest of the library to specify search flags lib/readline/histlib.h - _hs_history_search: extern declaration lib/readline/search.c - noninc_search_from_pos: use _hs_history_search, add CASEFOLD_SEARCH to flags if _rl_search_case_fold is non-zero doc/bash.1,lib/readline/doc/{readline.3,rluser.texi} - search-ignore-case: document new bindable variable 2/17 ---- config-top.h - PATSUB_REPLACE_DEFAULT: macro set to 1 or 0 to set the default value of `patsub_replacement'. Can be overridden by -DPATSUB_REPLACE_DEFAULT=0 in the `make' command builtins/shopt.def - reset_shopt_options: reset patsub_replacement to PATSUB_REPLACE_DEFAULT subst.c - patsub_replacement: initialize to PATSUB_REPLACE_DEFAULT parse.y - shell_getc: only call clearerr(stdin) when yy_stream_get returns EOF if the shell is currently interactive (interactive != 0) - shell_getc: if the shell is not interactive (interactive_shell == 0) and using stdio to read the script (bash_input.type == st_stream), set shell_input_line_terminator to READERR if ferror(stdin) is true after yy_stream_get returns EOF 2/20 ---- doc/{bash.1,bashref.texi} - update aliases description based on a bug-bash discussion - update description of word splitting behavior when IFS is unset builtins/evalfile.c - _evalfile: if reading the entire file doesn't return the same number of bytes as requested (the file size), treat this as a read error 2/21 ---- parse.y - yylex: if read_token returns < 0, return YYEOF if EOF_Reached = 1. Don't return yacc_EOF because that allows commands to be executed. - grammar: add a production to handle YYEOF, treating it the same as yacc_EOF. Based on a report from Eduardo A. Bustamante López jobs.c - wait_for: if we're checking for window size changes, allow checks during trap commands while readline is active or `bind -x' command execution. Fix from Koichi Murase 2/22 ---- jobs.c - wait_for: check for window size changes during programmable completion 2/23 ---- execute_cmd.c,input.c,jobs.c,nojobs.c,parse.y,redir.c,shell.c builtins/{exec.def,read.def} input.h - BUFFERED_INPUT: preprocessor #define is gone, this code is now unconditional 2/24 ---- builtins/declare.def - declare_internal: set the att_propagate flag on a variable that is marked as att_tempvar only if it's *not* marked att_local. The effect is that local variables with the same name as variables in the temporary environment are not propagated to the previous scope. From a report by Voldemar Lazarev that prompted a bug-bash discussion builtins/setattr.def - set_var_attribute: don't set local variables for which we are setting export or readonly to propagate back to the previous context, but make sure we preserve posix semantics parse.y - read_token_word: if we're parsing the words of a compound assignment (parser_state & PST_COMPASSIGN), and we have a valid assignment statement as a word ([sub]=value), set W_NOBRACE in the word so we don't try and perform brace expansion on it. This makes ( [sub]=word ) closer to name[sub]=word. From a report by Ilkka Virta back in July, 2020: https://lists.gnu.org/archive/html/bug-bash/2020-07/msg00133.html 2/27 ---- lib/readline/display.c - rl_redisplay: if HANDLE_MULTIBYTE is defined, do the special META_CHAR handling if wc_bytes == wc_width == 1 execute_cmd.c - eval_arith_for_expr: make sure we don't call make_word with a NULL string if expand_arith_string returns NULL. Report from F G arrayfunc.c - assign_assoc_from_kvlist: added code, currently disabled, to perform all expansions, including word splitting, on the kv-pair word list 2/28 ---- lib/readline/search.c - rl_history_search_reinit: change check against history_string_size to account for history_string_size now being a size_t. Report and fix from Grisha Levit 3/2 --- lib/readline/history.c - _hs_search_history_data: search the history list for an entry with `data' matching the argument; return the index - _hs_replace_history_data: search backwards through the history list from the end lib/readline/histlib.h - _hs_search_history_data: extern declaration lib/readline/search.c - make_history_line_current: make sure rl_undo_list doesn't appear in any history entries (it should be a private search string undo list) before freeing it lib/readline/misc.c - _rl_free_saved_history_line: make sure the undo list in _rl_saved_line_for_history isn't rl_undo_list and doesn't appear in any history list entries as `data' before freeing it. Fixes core dump after SIGINT reported by Grisha Levit parse.y,print_cmd.c - changes for size_t when computing new amount for xrealloc 3/3 --- parse.y - set_word_top: new inline function to check and set word_top; called from the appropriate places - parse_dparen: call set_word_top before parsing an arith command or nested subshell. Fixes underflow issue reported by Grisha Levit examples/loadables/kv.c - kv: new loadable builtin, reads key-value pairs from stdin and assigns them to an associative array 3/7 --- builtins/trap.def - trap_builtin: if trying to restore SIGQUIT to its default disposition, and the shell is in posix mode, set to SIG_DFL if the shell is running as an async command and signal_is_async_ignored (SIGQUIT) is true. Posix conformance issue 751 3/8 --- builtins/alias.def - print_alias: now returns int. Check for write errors using sh_chkwrite and return EXECUTION_FAILURE if it fails - alias_builtin: if print_alias returns EXECUTION_FAILURE, return EXECUTION_FAILURE immediately (write error). POSIX test conformance 3/9 --- builtins/cd.def - change_to_directory: if we're not in physical mode and are in posix mode, add step 9 of the posix cd algorithm, which essentially tries the pathname the user supplied if it's shorter than PATH_MAX and the length of the canonicalized pathname is longer than PATH_MAX. This is basically what default bash mode does without the length checks. POSIX test conformance. subst.c - strip_trailing_ifs_whitespace: use ifs_whitespace(*s) instead of spctabnl(*s) like word splitting and get_word_from_string. POSIX test conformance jobs.c - wait_for: tweak change from 1/18 to make sure a J_ASYNC job isn't in the foreground (J_FOREGROUND) as it would be if it had been continued in the foreground with `fg'; if it is, we want to give the terminal back to shell_pgrp 3/10 ---- input.c - fd_berror: returns true if the file descriptor passed as an arg corresponds to a buffered stream with an error condition parse.y - shell_getc: use fd_berror instead of inline code - shell_getc: if yy_getc returns EOF, immediately set line[0] to NULL and set shell_input_line_terminator to READERR. Then exit immediately with status 128 if index == 0, line[index] == NULL, and READERR. Posix interp 1629 config-top.h - FATAL_READERROR: now defined to 1 by default lib/readline/display.c - rl_expand_prompt: allocate an empty string for local_prompt if prompt is the NULL or empty string. Fixes seg fault with rl_message and empty prompt reported by Grisha Levit lib/readline/histexpand.c - history_expand: if hist_extract_single_quoted returns an index at or past the end of the string, correct the index and break the loop, which will cause a return. Report and fix from Grisha Levit 3/13 ---- lib/readline/vi_mode.c - rl_domove_motion_callback: set the MOVE_FAILED flag in the context M if rl_dispatch returns non-zero and other movement-command-specific conditions are met. Right now, an unsuccessful `f' or `F' command is failure. - _rl_vi_domove_motion_cleanup: if the command is `c' or `C', we don't enter insert mode if the MOVE_FAILED flag is set in the context M. This is how standalone vi works. lib/readline/kill.c - rl_unix_filename_rubout: add some checks for rl_point > 0 before looking at rl_line_buffer[rl_point - 1]. Report and fix from Grisha Levit lib/readline/isearch.c - _rl_isearch_dispatch: reset cxt->sline_index to 0 if it's past the end of the history string (sline_len). Fixes asan bug reported by Grisha Levit lib/readline/histexpand.c - history_tokenize_word: don't increment the string index past the NULL if the line ends in a backslash.Fixes asan bug reported by Grisha Levit lib/readline/vi_mode.c - rl_vi_yank_to,rl_vi_change_to,rl_vi_delete_to: if these are called with an existing valid vimvcxt, save it and allocate a new one for the current function call to use, then restore it before returning. This prevents referencing vimvcxt after freeing it in the case that these functions call each other recursively (e.g., y1y1, c1d1, etc.). Fixes asan bug reported by Grisha Levit 3/14 ---- bashhist.c - bash_add_history: make sure curlen is large enough before checking curlen - 1 and curlen - 2, otherwise you can get buffer underflow. Fixes asan bug reported by Grisha Levit bashline.c - bash_spell_correct_shellword: don't bother trying to correct empty words. Fixes asan bug reported by Grisha Levit lib/sh/shquote.c - sh_mkdoublequoted: make sure to allocate enough memory for the pathological case: where every character in the string needs to be backslash-escaped because it's special within double quotes. Fixes asan bug reported by Grisha Levit doc/{bash.1,bashref.texi} - here documents: make sure the summary specifies the here-document is terminated by `delimiter'; specify what the delimiter is if the word is not quoted. Prompted by a help-bash post from goncholden 3/15 ---- doc/bash.1,lib/readline/doc/rluser.texi - shell-expand-line: enumerate the specific expansions performed. From a request by Alex Bochannek findcmd.c - path_value: new function, returns normalized version of $PATH. Takes a single argument saying whether or not to look in the temporary environment first. Normalizes null $PATH (PATH=) into ".". From a bug-bash discussion started by Moshe Looks - _find_user_command_in_path: use path_value() - user_command_matches: use path_value() - search_for_command: if PATH[0] == 0, set PATH = "." like path_value(). This has the side effect of calling command_not_found_handle when PATH is the empty string and the command isn't in the current directory - search_for_command: if path == 0, but we found an executable in the current directory, make sure to set dot_found_in_search before adding it to the command hash table findcmd.h - path_value: extern declaration bashline.c - command_word_completion_function: use path_value() builtins/enable.def - dyn_load_builtin: use path_value for BASH_LOADABLES_PATH; it's cleaner though it doesn't change the behavior 3/17 ---- bashline.c - bashline_reset: reset rl_filename_quoting_function, since there is a SIGINT code path where it could remain reset by glob completion. Reported by Grisha Levit - command_word_completion_function,command_subst_completion_function, glob_complete_word: set some static variables to NULL after freeing them so they don't potentially get freed twice after a SIGINT during completion Reported by Grisha Levit parse.y - read_token_word: when deciding how to quote translated strings, don't free ttok before comparing it to ttrans Reported by Grisha Levit - parse_matched_pair: same thing for freeing nestret - read_token: if we return a newline, make sure we call set_word_top() if necessary. Fixes underflow reported by Grisha Levit lib/glob/sm_loop.c - EXTMATCH: don't bother with FNM_PATHNAME checks; this function isn't called with a pathname. Fixes underflow reported by Grisha Levit lib/readline/misc.c - rl_maybe_replace_line: if we replace a history entry's data with rl_undo_list, we need to set rl_undo_list to 0, since we should not have rl_undo_list pointing to something from a history entry Fixes asan error reported by Grisha Levit 3/20 ---- lib/readline/histexpand.c - postproc_subst_rhs: fix off-by-one error when reallocating new string for case where we're *not* substituting for `&' Fixes asan error reported by Grisha Levit jobs.c - terminate_current_pipeline: add a SIGKILL call in there in case the SIGTERM doesn't work - terminate_current_pipeline: add a version that works when job control isn't enabled (pipeline_pgrp == shell_pgrp) subst.c - skip_to_delim: pass completeflag (SX_COMPLETE) to extract_dollar_brace_string in case it includes a command substitution, since we want to inhibit error messages in this case parse.y - no_semi_successors: add DOLPAREN to list since something like $(\n shouldn't have a semicolon added there. From a report by Grisha Levit 3/21 ---- pcomplete.c - gen_globpat_matches: set noglob_dot_filenames here since we're not calling shell_glob_filename to do it From a report by Grisha Levit 3/22 ---- lib/readline/misc.c - rl_maybe_replace_line: if we're replacing a history entry with the current line buffer and undo list, then zero out any undo list in _rl_saved_line_for_history, since we're not going to use it any more Fixes asan error reported by Grisha Levit lib/readline/undo.c - rl_free_undo_list: if we're freeing rl_undo_list, it can't be a valid value for data in _rl_saved_line_for_history Fixes asan error reported by Grisha Levit lib/readline/complete.c - display_matches: don't call rl_display_match_list if we received a signal during get_y_or_n and didn't jump out for some reason From a report by Grisha Levit - rl_display_match_list: return if we received a signal during _rl_internal_pager (via get_y_or_n) and didn't jump out From a report by Grisha Levit 3/23 ---- lib/readline/history.c - history_truncate_file: don't bother trying to read the existing history file if we're just truncating it to 0 lines 3/24 ---- lib/readline/histfile.c - history_truncate_file: fix off-by-one error that resulted in the timestamp associated with the first history entry not being written to the truncated history file From a report by Grisha Levit parse.y - decode_prompt_string: do something rational if localtime() returns NULL (extraordinarily rare) From a report by Paul Eggert examples/loadables/stat.c - stattime: ditto examples/loadables/strftime.c - strftime_builtin: ditto lib/readline/examples/histexamp.c - main: ditto include/posixtime.h - getnow(): new inline function, calls gettimeofday and returns tv_sec; replacement for time(0). From Paul Eggert general.h - NOW: use a call to getnow() to standardize on gettimeofday for time information lib/readline/history.c - include #posixtime.h for getnow() - hist_inittime: use getnow() instead of calling time(0) parse.y - posixtime.h: include instead of - decode_prompt_string: use getnow() instead of calling time(0) 3/27 ---- parse.y - count_all_jobs: add extern definition if JOB_CONTROL is not defined. From Emanuele Torre bashhist.c - bash_add_history: make sure the current history line is not empty before checking what ended it. From a report by Grisha Levit lib/readline/histexpand.c - history_tokenize_word: break out of loop for incomplete command and process substitution or extended globbing patterns From a report by Grisha Levit bashline.c - hostnames_matching: fix size_t issue when checking whether to realloc From a report by Grisha Levit lib/sh/random.c - genseed: use uintptr_t instead of u_bits32_t to silence compiler warnings. From Paul Eggert aclocal.m4 array2.c builtins/mkbuiltins.c,builtins/psize-posix.c,version.c builtins/getopts.def lib/readline/text.c,lib/readline/bind.c support/recho.c,support/mksignames.c,support/zecho.c,support/xcase.c support/printenv.c,support/signames.c,support/bashversion.c support/memtest.c,support/endian.c lib/glob/gm_loop.c - some cleanups for function prototypes and other stdc requirements From Paul Eggert lib/readline/doc/{rltech.texi,hstech.texi} - add prototypes and stdc requirements to the examples builtins/complete.def - compgen_builtin: save and restore readline variables affecting how filename completions are treated in case someone uses compgen in the foreground. From a report by Grisha Levit 3/28 ---- aclocal.m4 builtins/enable.def,builtins/gen-helpfiles.c,builtins/help.def,builtins/printf.def general.h,hashlib.h include/ansi_stdlib.h,include/ocache.h lib/glob/collsyms.h,lib/glob/glob.c lib/readline/history.h,lib/readline/xmalloc.h xmalloc.c,xmalloc.h lib/sh/strtrans.c lib/tilde/tilde.c mksyntax.c support/endian.c - don't bother with __STDC__ if we are just using C89 features like const, function prototypes, etc. From Paul Eggert externs.h - prototype the list functions that take GENERIC_LIST * From Paul Eggert builtins/common.c,examples/loadables/getconf.c execute_cmd.c,pcomplete.c,subst.c lib/sh/stringlist.c,lib/sh/stringvec.c - changes and casts now that the generic list functions are prototyped From Paul Eggert 3/29 ---- aclocal.m4,configure.ac builtins/common.c,builtins/mkbuiltins.c,builtins/printf.def include/memalloc.h,include/stdc.h general.h,jobs.h pcomplete.c,print_cmd.c m4/iconv.m4 lib/readline/bind.c,lib/readline/complete.c,lib/readline/display.c lib/readline/funmap.c,lib/readline/util.c,lib/readline/parens.c lib/readline/histlib.h lib/readline/readline.h,lib/readline/rldefs.h,lib/readline/rlstdc.h,lib/readline/rlprivate.h lib/sh/dprintf.c,lib/sh/snprintf.c,lib/sh/vprintf.c lib/malloc/shmalloc.h,lib/malloc/imalloc.h,lib/malloc/xmalloc.c lib/termcap/termcap.h - stdarg: since we now assume C89 for prototypes at least, always prefer stdarg to old-style varargs - SH_VA_START: now use plain old va_start - don't bother with extern declarations for standard C89 functions From Paul Eggert config-bot.h - PREFER_STDARG: remove builtins/mkbuiltins.c execute_cmd.c,unwind_prot.c unwind_prot.h - start on ANSI/ISO C changes for the unwind-protect framework - introduce new unwind-protect function pointer type to replace Function From Paul Eggert 3/30 ---- builtins/command.def,builtins/fc.def,builtins/read.def,builtins/source.def builtins/evalfile.c,builtins/evalstring.c dispose_cmd.c,execute_cmd.c,pcomplete.c,print_cmd.c,subst.c,jobs.c,variables.c dispose_cmd.h,general.h,trap.h unwind_prot.c,unwind_prot.h - rest of work to convert unwind-protect framework to ANSI/ISO C. Convert unwind-protects to use new shim functions that obey the new unwind-protect function type From Paul Eggert general.h - STREQ, STREQN: now static inline functions to take advantage of arg type conversion pcomplete.c,subst.c lib/readline/histexpand.c,lib/readline/mbutil.c - fix size_t subtractions in comparisons to avoid unsigned underflow externs.h execute_cmd.c builtins/times.h - add prototypes for functions taking clock_t and timeval, make their declaration in externs.h conditional on NEED_CLOCK_FUNCS_DECL and NEED_TIMEVAL_FUNCS_DECL, respectively 3/31 ---- subst.c - expand_string_dollar_quote: check for zero-length return from string_extract_{single,double}_quoted and avoid size_t underflow Fixes asan error reported by Grisha Levit 4/3 --- builtins/evalstring.c - open_redir_file: treat the case of failing to open the file as a non-fatal expansion error instead of a (fatal) redirection error. Report from Zev Weiss 4/4 --- lib/readline/vi_mode.c - _rl_domove_motion_cleanup: d/D require the same special case as c/C so that a motion command that doesn't move the cursor consumes/deletes at least one character. From Emanuele Torre - _rl_domove_motion_cleanup: ditto for y/Y - rl_domove_motion_callback: if t/T/;/, fail (return non-zero without moving point), flag the motion command as having failed (MOVE_FAILED) 4/6 --- subst.c - command_substitute: save the return status of the child process in last_command_subst_status; don't assign to last_command_exit_value in posix mode (posix interp 1150) execute_cmd.c - execute_null_command: a simple command without a command word but with command substitutions now returns last_command_subst_status. It may or may not have already modified $? depending on posix mode redir.c - redir_varvalue: legal_number validity check should be == 0, not < 0 From a report by Grisha Levit 4/10 ---- subst.c - parameter_brace_expand_rhs: check for namerefs in the variable name part of the ${name=word} expansion so we can go back and implement the POSIX semantics of returning "the final value of parameter." From a report by Grisha Levit sig.c,sig.h - sigpipe_handler: clean up, set $?, and throw to top level on receipt of a SIGPIPE - termsig_handler: if the variable builtin_catch_sigpipe is set (it's not set anywhere yet), call sigpipe_handler instead of terminating the shell. Still need to make sure a sighandler is installed for SIGPIPE even if initialize_terminating_signals isn't called 4/14 ---- lib/sh/tmpfile.c - sh_mktmpfd: add support for MT_UNLINK flag, unlinks created filename before returning fd, fails if the unlink fails for some reason 4/15 ---- pathexp.c,pathexp.h - globsort: support for various glob sorting criteria as defined by the GLOBSORT variable: name, size, blocks, mtime, atime, ctime. Default is ascending sort by name, as is traditional. Based on an idea from Evan Gates - setup_globsort: parse the value of the GLOBSORT variable, find sorting type and whether or not it's ascending or descending - noquote_glob_filename: glob_filename for programmable completion, doesn't call quote_string_for_globbing and honors GLOBSORT, dotglob, and globstar pcomplete.c - gen_globpat_matches: call noquote_glob_filename, let that handle all the variables and GLOB_FAILED variables.c - GLOBSORT: support for special variable, call setup_globsort on set or unset doc/bash.1,doc/bashref.texi - GLOBSORT: document variable and its effects on pathname expansion 4/17 ---- builtins/complete.def - build_actions: take an additional char ** arg VNAMEP; if it is non- null, accept a -V VARNAME option that names an indexed array variable in which to store the completions. - compgen_builtin: if the -V option is supplied, store the possible completions into VARNAME instead of printing them on stdout. From a patch from Grisha Levit , idea originally from konsolebox back in 2/2022 doc/bash.1,lib/readline/doc/rluser.texi - compgen: document new -V option 4/18 ---- lib/readline/vi_mode.c - rl_domove_read_callback: call _rl_arg_init before reading digit argument so prompt gets saved, both callback and non-callback modes From a report by Grisha Levit - rl_vi_arg_dispatch: if we return 0 to end the digit argument and turn off NUMERICARG, call rl_restore_prompt lib/readline/histexpand.c - history_expand_internal: make sure subst_lhs is not null before we call postproc_subst_rhs. From a report by Grisha Levit lib/readline/display.c - rl_redisplay: remove redundant code that ends up incrementing lpos too much for meta characters displayed as octal escape sequences. From a report by Grisha Levit 4/19 ---- bashline.c - attempt_shell_completion: handle case where s > e after calls to find_cmd_start and find_cmd_end, respectively. From a report by Grisha Levit lib/readline/display.c - update_line: fix setting of `wsatend' when computing ols and nls in the multibyte case; this caused redisplay to print more characters than it needed - update_line: fix calculation of visible_wrap_offset when moving up to line 0, trying to figure whether or not the visible prompt has any invisible characters. Fixes C locale issue reported by Grisha Levit 4/20 ---- lib/readline/text.c - _rl_rubout_char: update erase-at-end-of-line hack to deal with the unlikely event that a character in a non-multibyte locale is displayed using an octal representation. From a report by Grisha Levit execute_cmd.c,execute_cmd.h,builtins/evalstring.c - interrupt_execution: replaces `executing_list' - interrupt_execution: set while executing while, for, until, select, and arithmetic for loops jobs.c - wait_for: if a process dies due to a SIGINT while the shell is executing a loop or a list, act as if the shell got SIGINT. This is no longer dependent on the compatibility level From a report by Grisha Levit bashline.c - name_is_acceptable: allow FIGNORE suffixes to match the entire pathname like tcsh does instead of requiring a non-empty prefix. From Emanuele Torre 4/23 ---- lib/readline/misc.c - _rl_unsave_line: new function, takes a HIST_ENTRY * as an argument, generalized rl_maybe_unsave_line - _rl_alloc_saved_line: allocate a new HIST_ENTRY *, populate it like rl_saved_line_for_history - _rl_free_saved_line: free a saved HIST_ENTRY * - _rl_free_saved_history_line: call _rl_free_saved_line - _rl_previous_history_internal: new function, internals of rl_get_previous_history; change rl_get_previous_history to call it lib/readline/history.h - HIST_ENTRY_DEFINED: define sentinel if HIST_ENTRY struct defined lib/readline/rlprivate.h - _rl_alloc_saved_line,_rl_free_saved_line,_rl_unsave_line: new extern function definitions - _rl_free_saved_search_line: new extern definition lib/readline/search.c - _rl_saved_line_for_search: use private storage to save current line, avoid clashes with rl_saved_line_for_history - _rl_free_saved_search_line,_rl_unsave_saved_search_line: new functions to manage _rl_saved_line_for_search - noninc_dosearch,_rl_nsearch_abort: clean up _rl_saved_line_for_search instead of rl_saved_line_for_history - rl_history_search_internal: allocate private _rl_saved_line_for_search; unsave or free it as appropriate - _rl_history_search_reinit: free up any _rl_saved_line_for_search just in case lib/readline/vi_mode.c - rl_vi_search: free _rl_saved_line_for_search, since that's what the non-incremental search functions use now - rl_domove_read_callback: if we read a numeric argument, then a motion character, make sure the motion character is valid before returning it and going on From a report by minipython <599192367@qq.com> 4/27 ---- lib/readline/display.c - _rl_move_cursor_relative: when checking to see whether data is within the invisible line, make sure to stay within the invisible line line break boundaries From a report by minipython <599192367@qq.com> lib/readline/search.c - dispose_saved_search_line: new function, either unsave the saved search line (if we're not at the end of the history) or free it (if we are). Needs to be immediately followed by call to make_history_line_current, which resets rl_undo_list - make_history_line_current: change strategy: rely on callers to manage _rl_saved_line_for_search and rl_undo_list; just copy the data from the passed history entry into the line buffer and reset rl_undo_list - noninc_dosearch,rl_history_search_internal: call dispose_saved_search_line before calling make_history_line_current lib/readline/misc.c - _rl_start_using_history: free the undo list associated with _rl_saved_line_for_history, if one exists. Have to watch out for this causing pointer aliasing problems, maybe add a call to _hs_search_history_data() 4/28 ---- pathexp.c - sh_globsort: a sort specifier of `nosort' disables sorting completely doc/bash.1,doc/bashref.texi - GLOBSORT: document `nosort' sort specifier bashline.c - attempt_shell_completion: attempt completion with the empty command for a null command word on a line with only whitespace following any optional assignment statements - attempt_shell_completion: make sure that the start of the command word is not after rl_point before calling the programmable completion for the initial word. From a report by Grisha Levit lib/readline/histexpand.c - history_expand: don't read past the end of the string if it ends with an incomplete multibyte character From a report by Grisha Levit 5/1 --- lib/readline/text.c - _rl_insert_char: make sure to return 1 in all cases where there are pending bytes accumulating for a multibyte character - _rl_insert_char: if the COUNT argument is 0, insert any pending bytes read so far. The idea is that these bytes were accumulating to compose a possible multibyte character, and the count of 0 indicates that the current character doesn't contribute to that because it's bound to an editing command. - rl_insert, rl_quoted_insert: make sure to call _rl_insert_char with a 0 count as appropriate to insert any accumulating bytes From a patch by Grisha Levit sig.c - top_level_cleanup,throw_to_top_level: set variable_context to 0, just to make sure execute_cmd.c - function_misc_cleanup: new function, runs make_funcname_visible and unlink_fifo_list at the end of execute_function, replaces inline code - uw_function_misc_cleanup: unwind-protect function, just calls function_misc_cleanup. Installed by execute_function. Partial fix for FUNCNAME bug reported by Grisha Levit - execute_builtin: run unwind-protects if evalnest exeeds max - execute_builtin: run unwind-protects if sourcenest exeeds max parse.y - parse_string_to_word_list: set current_token = '\n' on error - parse_string_to_word_list: if we're executing and not running in parse_and_execute, call top_level_cleanup before we jump to the top level. Rest of fix for FUNCNAME bug - parse_compound_assignment: set last_read_token and current_token to '\n' on parse error. Fix for alias expansion bug reported by Grisha Levit - parse_compound_assignment: if we're executing and not running in parse_and_execute, call top_level_cleanup before we jump to the top level 5/2 --- jobs.c - run_sigchld_trap: don't freeze and unfreeze the jobs list; move it to the callers who are running this on SIGCHLD receipt and cannot deal with jobs disappearing from the list while they're running - waitchld: if we are running the wait builtin, continue to run the SIGCHLD trap immediately upon signal receipt, but freeze and unfreeze the jobs list around the call to run_sigchld_trap(). Fixes issue with `jobs' in a SIGCHLD trap not removing jobs from the jobs list reported by Guldrelokk 5/4 --- input.c - fd_to_buffered_stream: make sure buffers[fd2] is non-null before turning on SHAREDBUF. From Emanuele Torre eval.c - reader_loop: save and restore EOF_Reached around expanding $PS0, setting it to 0 for the expansion so it doesn't affect things like command substitution. From a report by Grisha Levit parse.y - parse_comsub: handle ${Ccommand; }, where C == space, tab, newline, or '|' - read_token_word: understand ${Ccommand; } and call parse_comsub to parse the contents - parse_matched_pair: understand ${Ccommand; } inside another paired construct - new `funsub' production, like comsub but for ${Ccommand;} nofork command substitution execute_cmd.c - execute_simple_command: if we fork because of pipes in or out, clear the fifo list, since subshell aren't supposed to inherit fifos 5/8 --- parser.h - PST_FUNSUBST: new parser state, parsing the command in ${Ccommand;}. Primarily used to have read_token_word return `}' as a separate token even if it's not delimited like a reserved word subst.h - SX_FUNSUB: new flag, tell xparse_dolparen we are parsing the command in a ${Ccommand;} nofork command substitution parse.y - xparse_dolparen: handle parsing ${Ccommand;} nofork command substitution like parse_comsub - read_token: if we're parsing a foreground command substitution, treat a word that begins with an unquoted `}' as a `}' token to terminate the ${Ccommand;} construct without requiring that the close brace be delimited like a reserved word - no_semi_successors: add DOLBRACE - reserved_word_acceptable: add DOLBRACE subst.c - extract_function_subst: like extract_command_subst, but for nofork command substitutions; calls xparse_dolparen with the SX_FUNSUB flag to differentiate - string_extract_double_quoted,extract_delimited_string, extract_heredoc_dolbrace_string,extract_dollar_brace_string, param_expand: call extract_function_subst as appropriate - function_substitute: like command_substitute, but for nofork command substitutions. Just a stub for now 5/11 ---- jobs.[ch] - pipeline_saver: already_making_children now saved and restored as part of the struct; no longer need a separate variable - save_pipeline,restore_pipeline: save already_making_children to struct parse.y - parse_comsub,xparse_dolparen: handle `(' as the first char of a funsub (ksh93 supports that) subst.c - function_substitute: complete implementation; handle both funsubs and valsubs (need different terminology); dummy up a fake shell function so local variables and `return' work; do the right thing with process groups - executing_funsub: flag for the rest of the shell to know we're expanding a funsub; used by parse_and_execute builtins/evalstring.c - parse_and_execute: check executing_funsub to do the right thing with alias expansion and posix mode 5/12 ---- doc/bash.1,doc/bashref.texi - function substitution (or lambda substitution, or nofork command substitution): document 5/15 ---- subst.c - function_substitute: make sure to decrement variable_context in the unwind-protects since we don't unwind_protect it individually. Report from Oguz 5/17 ---- configure.ac - i370 machines cannot use the bash malloc Patch from Igor Todorovski support/signames.c - initialize_signames: add support for z/OS signals. Patch from Igor Todorovski lib/glob/smatch.c - collseqcmp: new name for collequiv[_wc], that function apparently exists in libc on z/OS Patch from Igor Todorovski jobs.c - mark_all_jobs_as_dead: make sure to reset js.j_ndead to the right number of jobs in the array execute_cmd.c - execute_while_or_until: move call to REAP() after the action executes, like other loop functions shell.c,parse.y,lib/intl/gettextP.h - add some __MVS__ checks for z/OS systems 5/18 ---- jobs.c - wait_for_any_job: catch jobs that can terminate and be reaped when we call UNBLOCK_CHILD between the loop where we check the jobs list for dead jobs and the loop where we check for running jobs and poll for them with wait_for(). Fixes bug reported by Aleksey Covacevice 5/20 ---- subst.c - read_comsub: quote CTLNUL if it's in $IFS and we're parsing the RHS of an assignment statement - comsub_quote_string: ditto - comsub_shouldquote: break code that decides whether to CTLESC escape a character read from a command substitution or $REPLY into its own inline function; change read_comsub/comsub_quote_string callers 5/21 ---- subst.c - quote_escapes_internal: use ifs_cmap instead of iterating through ifs_value to set skip_ctlesc and skip_ctlnul 5/23 ---- parser.h - remove `(' from FUNSUB_CHARS parse.y - parse_comsub: don't bother with `(' as one of the FUNSUB_CHARS doc/bash.1,doc/bashref.texi - command substitution: remove mention of `(' as a valid character following the open brace in nofork command substitution 5/24 ---- parse.y - parse_comsub: avoid adding `;' to an empty ${ command substitution - parse_comsub: if open == '{', push back a newline we read into peekc to get PS2 prompting right. From a report by Grisha Levit subst.c - skip_double_quoted: pass the SX_COMPLETE flag down to extract_dollar_brace_string in case it contains an unclosed command substitution. From a report by Grisha Levit 5/25 ---- shell.c - make sure shell_initialized is set to 1 before running -c command make_cmd.c - make_function_def: initialize source_file to something other than `main' or `environment' if the shell is not interactive. From a report from Emanuele Torre parse.y - save_token_state,restore_token_state: save and restore token_to_read - read_token_word: if we read EOF in an interactive shell, set token_to_read to yacc_EOF and EOF_Reached to 1 so the token will be delimited but the next call to read_token returns yacc_EOF - handle_ignoreeof: broke the code that checks eof_encountered and possibly exceeds the ignoreeof limit out into a separate public function, called from handle_eof_input_unit eval.c - reader_loop: if EOF_Reached is set in an interactive shell after parse_command and execute_command, and ignoreeof has been enabled, check for ignoreeof with handle_ignoreeof and reset EOF_Reached back to 0 if we should not exit. This can occur when EOF delimits a simple command (simple_command_terminator) From a report by Grisha Levit 5/27 ---- lib/readline/mbutil.c - _rl_get_char_len: if SRC is "", make sure to return 0 lib/readline/isearch.c - _rl_isearch_callback: if _rl_search_getchar returns EOF, return 1 to abort search; if it returns with RL_STATE_ISEARCH unset, return 1 (handling a signal can turn it off via _rl_isearch_cleanup) 5/29 ---- lib/readline/text.c - _rl_insert_char, rl_quoted_insert: minor fixes from Grisha Levit builtins/printf.def - bexpand: use the same calculation for the size of the return buffer as ansicstr(). From a report by Grisha Levit locale.c - locale_setblanks: work around problem with macOS where isblank(x) incorrectly returns true for characters 0x80-0xff in a UTF-8 locale From a report by Grisha Levit - set_default_locale_vars,set_locale_vars,reset_locale_vars: make sure we call locale_setblanks after setting/resetting locale_utf8locale, since we use that in locale_setblanks lib/sh/smatch.c - charseqcmp,is_cclass,FOLD: make sure we don't treat chars >= 0x80 as valid in a UTF-8 locale, since they are multibyte characters. From a report by Grisha Levit with a full explanation in http://www.openradar.me/FB9973780 5/30 ---- builtins/enable.def - enable_builtin: don't call dlopen with a null pathname, it will succeed. Report from Wiley Young lib/readline/readline.c - _rl_dispatch_subseq: if a user binds do-lowercase-version to something that isn't an uppercase character, flag an error rather than recurse. From a report by Grisha Levit 5/31 ---- builtins/enable.def - enable_builtin: don't try to load a builtin that's not found if the -n flag is supplied. From a report from Emanuele Torre parse.y - parse_comsub: use was_word code again to make sure we add a closing `;' if the last token parsed before the closing ';' or '\n' was a word, no matter what character ends it. From a report by Grisha Levit aclocal.m4,configure.ac,findcmd.c - IBM z/OS changes from Igor Todorovski 6/1 --- parse.y - shell_getc: if the shell is not interactive and reading from a buffered stream or stdin, and not expanding an alias, add a backslash to a line ending with like we do when reading from a string. This prevents a backslash-newline from being discarded when we are removing backslash-newlines from the input, since we will add a newline to shell_input_line in this case. From a report by Rob Landley parse.y - parse_compound_assignment: check and compensate for an alias being popped out from underneath this function by read_token() (e.g., alias L='m=("x")'. Since we don't push a new input source, we should never restore pushed_string_list from the saved parser state, but we check and only do this if we were expanding an alias when this function was called. From a report by Wiley Young 6/2 --- print_cmd.c - print_if_command: make sure to print any pending here-documents after printing the test. From https://bugzilla.redhat.com/show_bug.cgi?id=2211214 builtins/enable.def - if -n is supplied with -f, attempt to load the builtin but mark it as disabled after loading. Suggested by Robert Elz builtins/evalfile.c - _evalfile: increment retain_fifos, so we don't delete any FIFOs or pipes we inherited before sourcing this file; restore original value before we return. From https://savannah.gnu.org/support/index.php?110883 6/3 --- subst.c - expand_string_dollar_quote: if singlequote_translations is set, there is a chance for a use-after-free of `t'. From a report by Grisha Levit 6/7 --- jobs.c,nojobs.c,jobs.h - job_control_active_p: new function, returns 1 if job control is currently enabled and we have given away the terminal builtins/evalstring.c - should_suppress_fork: don't optimize away the fork if job control is active because we need to call end_job_control and give the terminal back to the original process group when this command finishes. Call job_control_active_p() to check this. Report from Andrew Hamon 6/8 --- shell.c - ssh_run_startup_files: new variable, set by run_startup_files if the shell decides it's being run by ssh and runs bashrc (only if SSH_SOURCE_BASHRC is defined). Non-zero while the startup files are being run. Not used by any other part of the shell yet. 6/12 ---- arrayfunc.h - new inline functions to convert between assignment flags (ASS_*), array value flags (AV_*), and valid array flags (VA_*) shell.h - include arrayfunc.h after subst.h and variables.h so all the values for the new inline functions are available variables.c - bind_int_variable: use convert_assign_flags_to_validarray_flags and convert_assign_flags_to_arrayval flags instead of inline assignments builtins/common.c - builtin_bind_variable: use convert_assign_flags_to_validarray_flags instead of inline assignments arrayfunc.c - assign_array_element: use convert_assign_flags_to_arrayval_flags instead of inline code - assign_array_element_internal: use convert_assign_flags_to_arrayval_flags to pass the right values to array_expand_index - assign_array_var_from_string: since expand_compound_array_assignment performs one round of expansion on the subscripts, make sure to add ASS_NOEXPAND (if assoc_expand_once is set) and ASS_ALLOWALLSUB to the flags we pass to assign_compound_array_list - unbind_array_index: use convert_validarray_flags_to_array_value_flags to pass the right values to array_expand_index - array_expand_index: uncomment code tagged for bash-5.3 that checks AV_NOEXPAND and suppresses call to expand_arith_string if it's set. This set of changes extends assoc_expand_once to indexed arrays arrayfunc.[ch],execute_cmd.c,expr.c,test.c,builtins/common.c,builtins/mkbuiltins.c builtins/declare.def,builtins/set.def,builtins/shopt.def - array_expand_once: new name for internal assoc_expand_once variable 6/13 ---- builtins/shopt.def - array_expand_once: new option, replaces assoc_expand_once doc/bash.1,doc/bashref.texi - array_expand_once: document new shopt option 6/15 ---- lib/readline/complete.c - rl_full_quoting_desired: new application-settable variable, if set to non-zero by an application completion function, all completions will be quoted as if they were filenames, subject to the value of rl_filename_quote_characters. Initialized to 0 and reset to 0 by set_completion_defaults. From a contribution by Grisha Levit - QUOTING_DESIRED: new macro to check whether we should quote completions lib/readline/readline.h - rl_full_quoting_desired: extern declaration lib/readline/doc/rltech.texi - rl_full_quoting_desired: document configure.ac - bumped version to 5.3-devel pcomplete.h - COPT_FULLQUOTE: new compspec option, for rl_full_quoting_desired pcomplete.c - pcomp_set_readline_variables: set rl_full_quoting_desired according to COPT_FULLQUOTE builtins/complete.def - fullquote: add new complete/compopt option value doc/bash.1,lib/readline/doc/rluser.texi - fullquote: document new complete/compopt option lib/sh/timeval.c - timeval_to_secs: sfp argument is now a long *; there is a new maxval argument (always 10^6 for now) that determines how to round; returns tv_usec in *sfp if maxval == 10^6; otherwise return value is basically tv_usec/maxval with rounding - print_timeval: change argument type in call to timeval_to_secs lib/sh/clock.c - clock_t_to_secs: sfp argument is now a long * - print_clock_t: change argument type in call to clock_t_to_secs externs.h - timeval_to_seconds, clock_t_to_seconds: change prototypes execute_cmd.c - mkfmt,print_formatted_time: fractional seconds arguments are now longs - mkfmt: accept precisions up to 6; since sec_fraction is in usecs, perform any necessary rounding for precisions < 6 - print_formatted_time: allow precision specifiers up to 6; default is still 3 - time_command: fractional seconds variables are now long; add new argument to timeval_to_secs (always 1000000) - time_command: if we call clock_t_to_secs, multiply the return fractional seconds value by 1000 to convert it from msec to usec doc/bash.1,doc/bashref.texi - TIMEFORMAT: document that the max precision is now 6 From https://savannah.gnu.org/support/?110343 6/16 ---- variables.c - get_monoseconds: use clock_gettime to return the value of the system's monotonic clock, if available - BASH_MONOSECONDS: a new dynamic variable that returns the value of get_monoseconds From a suggestion by William Kennington doc/bash.1,doc/bashref.texi - BASH_MONOSECONDS: document trap.c - save_bash_trapsig,set_bash_trapsig,restore_bash_trapsig: functions to manage value of BASH_TRAPSIG - run_pending_traps: save and restore the value of BASH_TRAPSIG; set it to the numeric signal number of the trap being run while it's executing From a feature request by Eric Marceau - _run_trap_internal,run_exit_trap: save and restore BASH_TRAPSIG; make sure it's set appropriately doc/bash.1,doc/bashref.texi - BASH_TRAPSIG: document builtins/declare.def - declare_internal: if declare is supplied -f and an argument that looks like an assignment statement, fail only if there is not a function with that name already defined. From a report from Emanuele Torre back in 12/22 jobs.c - wait_for: make checkwinsize work in subshell commands started from interactive shells New feature request by Kerin Millar print_cmd.c - semicolon: don't return immediately unless a preceding `&' is itself preceded by a space, as the rest of the printing code prints asynchronous commands. Report and patch by Grisha Levit 6/17 ---- arrayfunc.c - arrayvar_copyval: copy an array or hash table from shell variable V1 to V2 arrayfunc.h - arrayvar_copyval: extern declaration subst.h - ASS_NOTEMPENV: new flag for bind_variable and assignments variables.c - bind_variable: if ASS_NOTEMPENV is in FLAGS, don't bother calling bind_tempenv_variable builtins/declare.def - declare_internal: if we find the variable in the temporary environment, call bind_variable with the ASS_NOTEMPENV flag since we are already dealing with a variable from the temporary environment - declare_internal: if we bind a new global variable because we're modifying a variable from the temporary environment, make sure to call arrayvar_copyval to copy array variables from the temporary environment to the global scope variables.c - push_temp_var: if we are pushing an array variable, make sure to call arrayvar_copyval on the new variable Final fix for bug reported by Wiley Young - push_posix_tempvar_internal: call arrayar_copyval instead of using inline code 6/19 ---- lib/sh/shmatch.c - sh_regmatch: use builtin_find_indexed_array to allow a local copy of BASH_REMATCH. From https://lists.gnu.org/archive/html/bug-bash/2022-05/msg00052.html execute_cmd.c - execute_in_subshell: tentative change to clear procsub list after performing redirections, so nothing in a redirection can affect what's in the subshell. From https://lists.gnu.org/archive/html/bug-bash/2022-10/msg00114.html test.c - binary_operator: in posix mode, the < and > operators compare strings using the current locale (strcoll). From https://www.austingroupbugs.net/view.php?id=375 bashline.c - bash_default_completion: make sure that command completion adds the trailing `/' suffix for directories even if the directory name is the same as the common prefix of matches. From https://lists.gnu.org/archive/html/bug-bash/2022-11/msg00034.html lib/readline/kill.c - rl_vi_yank_pop: make repeated invocations of vi-yank-pop cycle through the kill ring just like in emacs mode, even when point is at the end of the line. From https://lists.gnu.org/archive/html/bug-bash/2022-08/msg00105.html parse.y - parse_comsub: set variables if the last token (last_read_token) is `;', `\n', or `&', in addition to was_word - parse_comsub: if the last read token was a newline (was_newline), terminate the comsub with a newline instead of a semicolon; keeps the line numbers more consistent. From a report by Grisha Levit lib/sh/anonfile.c - anonopen: use memfd_create if it is available, fall through to traditional Unix/POSIX implementation if it fails 6/20 ---- parse.y - yy_string_unget: don't push EOF back to a string, like the other unget functions 6/21 ---- parse.y - read_a_line: don't push an EOF back into the string or line if the line ends in a backslash. From a report by Rob Landley 6/22 ---- lib/readline/input.c - rl_gather_tyi: make sure result is initialized lib/readline/kill.c - _rl_read_bracketed_paste_prefix: make sure key is initialized builtins/read.def - read_builtin: make sure pass_next and saw_escape are initialized before a possible goto - read_builtin: make sure to initialize prevset for very short timeouts subst.c - function_substitute: make sure tflag is initialized even when we don't call read_comsub, since we return it in ret->flags All from a report by Grisha Levit builtins/complete.def - compgen_builtin: use array_from_argv to assign elements of the stringlist to the array variable. Update from Grisha Levit builtins/cd.def - bindpwd: fix seg fault from an unlikely set of circumstances From a report by Grisha Levit arrayfunc.h - convert_validarray_flags_to_arrayval_flags: initialize avflags From a report by Grisha Levit lib/sh/anonfile.c - anonopen: set *fn if memfd_create is used From a report by Grisha Levit 6/23 ---- arrayfunc.c - bind_assoc_var_internal: free key to fix small leak if assign_func used - quote_compound_array_word: free value to fix small leak builtins/evalstring.c - parse_and_execute: free parsed command on failed function definition import - open_redir_file: free FN if we're not passing it back to the caller subst.c - param_expand: free TEMP1 in code paths that don't do it now bashline.c - bash_command_name_stat_hook: if we modify *NAME, free the old value examples/loadables/{kv,stat}.c - bind_assoc_variable is caller-free VALUE, so free when needed and don't allocate a new copy if not All from a report by Grisha Levit 6/24 ---- lib/readline/complete.c - rl_menu_complete: use _rl_free_match_list instead of just freeing MATCHES if we have too many possible completions to display. From a report by Grisha Levit 6/25 ---- eval.c - reader_loop: make sure to allow exit builtin (code == EXITBLTIN) to exit the shell when ignoreeof is set From a report by Grisha Levit 6/26 ---- lib/readline/complete.c - complete_sigcleanarg_t: new struct to hold match list and saved line buffer for cleanup on receipt of SIGINT - _rl_complete_sigcleanup: use new sigcleanarg_t struct and free both members - rl_complete_internal: whenever we display the match list, set up to clean the matches and saved line buffer in the event of a SIGINT From a report by Grisha Levit lib/readline/signals.c - _rl_handle_signal: if readline is compiled to include callbacks, only call rl_echo_signal_char if we're not in callback mode or the application has set rl_persistent_signal_handlers, leaving any application signal handler to call it (or not) otherwise. From a discussion with Andrew Burgess lib/glob/glob.c - glob_filename: if ARRAY == TEMP_RESULTS, make sure to free TEMP_RESULTS after copying the filenames out of ARRAY, since we either assigned it directly or glob_dir_to_array returned it because the dirname was the empty string. From a report by Grisha Levit lib/readline/rltypedefs.h - Function: only compile in these obsolete typedefs if WANT_OBSOLETE_TYPEDEFS is defined print_cmd.c - xtrace_print_assignment,xtrace_print_word_list: prioritize checking for characters that need $'...' printing over shell metacharacters so that strings containing both get the $'...' treatment From a report by Grisha Levit 6/27 ---- subst.c - skip_double_quoted: make sure to call extract_function_subst with the SX_NOALLOC flag From a report by Grisha Levit trap.c - _run_trap_internal: if we're running a trap but shell_eof_token is set, this means we're somehow running a trap while parsing a command substitution (weird set of circumstances). Call reset_parser and unset shell_eof_token if shell_eof_token is set From a report by Wiley Young builtins/setattr.def - set_var_attribute: don't set the att_propagate attribute unless the variable whose attribute is being modified is in the temporary environment for this builtin, not any previous temporary environments (like the temp env for a function call, or a special builtin like `.') variables.c - push_posix_temp_var: if we have an array variable we're trying to push up here, use arrayvar_copyval to copy the value correctly arrayfunc.c - arrayvar_copyval: make sure that V2 ends up being the same type of array as V1 (att_array or att_assoc), since we're copying the value. Fixes issue reported by Grisha Levit , but with the previous fix to set_var_attribute, it's not needed - assign_compound_array_list: check for integer overflow if the max index of the array is already INT_MAX. From a report from Emanuele Torre - assign_array_var_from_word_list: ditto 6/28 ---- pcomplete.c,bashline.c - uw_restore_parser_state,uw_rl_set_signals: move to bashline.c so the general readline support can use them bashline.h - uw_restore_parser_state,uw_rl_set_signals: extern declarations bashline.c - unset_readline_variables,uw_unset_readline_variables: function to unset READLINE_{LINE,POINT,MARK,ARGUMENT} and its unwind-protect counterpart - bash_execute_unix_command: add unwind-protect to free up allocated memory Report and patch by Grisha Levit doc/bash.1,doc/bashref.texi - BASH_ARGC,BASH_ARGV,BASH_SOURCE,BASH_LINENO: note that these variables may not be assigned to or unset 6/29 ---- builtins/declare.def - declare_internal: attempting to assign to a noassign variable is no longer an assignment error; the attempt is just ignored arrayfunc.c - find_or_make_array_variable: take a new flag value, 4: means to return noassign variables as themselves instead of NULL without performing the assignment so the caller can handle them differently. - assign_array_from_string: pass flags & 4 to find_or_make_array_variable so we can simply ignore attempted assignments to noassign array variables instead of treating them as assignment errors From a report by Kerin Millar expr.c - exppower: only report negative exponent error if noeval == 0. Fixes issue reported by Steffen Nurpmeso 6/30 ---- lib/sh/anonfile.c - anonopen: add support for using shm_open or shm_mkstemp to create an anonymous file using POSIX shared memory configure.ac - if --enable-static-link is supplied, add -static to LDFLAGS on Linux if opt_profiling isn't enabled From a report from Emanuele Torre print_cmd.c - make_command_string_internal: if we're recursively printing a list (connection) with more than two elements, don't print any deferred here-documents after the make_command_string_internal on the right side of the connection unless we're at the end of the list (printing_connection == 1). This way the caller gets to add the appropriate connector before printing the deferred here-documents. From a report by Grisha Levit - make_command_string_internal,print_group_command: after we call PRINT_DEFERRED_HEREDOCS and follow it with a closing `)' (subshell) or `}' (group command), we can set was_heredoc to 0 because we are no longer printing a here-document From a report by Grisha Levit 7/3 --- lib/readline/complete.c - rl_username_completion_function: simplify things by just skipping the function body if HAVE_GETPWENT is not defined. From a report by Grisha Levit lib/sh/ufuncs.c - fsleep: broke into three functions: nsleep, which uses nanosleep, ssleep, which uses select/pselect, and ancientsleep, which uses sleep. Prefer nanosleep, use the others depending on what's defined. From a report and patch by Grisha Levit arrayfunc.c - assign_compound_array_list: now behaves more like a series of assignment statements (a=(one two) is more like a[0]=one a[1]=two) in terms of error handling -- break after the first invalid assignment -- and return value -- now return int, 0 on failure and 1 on success so the callers can throw assignment errors - assign_compound_array_list: if we have an explicit [subscript]=value assignment, use the negative-indices-count-back-from-the-end strategy; otherwise use last_ind and throw an error on overflow. This is like standalone assignment statements - assign_array_var_from_string: catch new return value from assign_compound_array_list; return NULL if assign_compound_array_list returns failure subst.c - do_compound_assignment: if assign_compound_array_list or assign_array_var_from_string return error (0 or NULL, respectively), treat as an assignment error similar to assigning to a readonly variable Inspired by a report from Emanuele Torre arrayfunc.h - assign_compound_array_list: new return type 7/5 --- builtins/enable.def - enable_builtin: make sure dynamic builtin loading and unloading updates the `enabled' and `disabled' programmable completion lists. From a report and patch by Grisha Levit doc/bash.1,doc/bashref.texi - hash: add note that assigment to PATH clears the hash table. Report by Sebastian Carlos lib/readline/complete.c - rl_filename_completion_function: make sure the results of *rl_filename_rewrite_hook are freed consistently if the function returns newly-allocated memory From a report and patch by Grisha Levit back in May 7/6 --- builtins/cd.def - cd_builtin: if $OLDPWD is set by the user or script to something that's not a full pathname, allow it to use $CDPATH. From a report by Dustin Boyd 7/7 --- shell.h - user_info: add members for saved uid and saved gid shell.c - uidget: if we have setresuid/setresgid, get the saved uid and saved gid so we can set them if we disable privileged mode - disable_priv_mode: only call setuid/setresuid and setgid/setresgid if the euid (egid) != uid (gid). If we have setresuid/setresgid, add a check whether the saved uid (gid) isn't the same as the real uid (gid). Potentially saves a couple of system calls. Fixes Android issue, patch by Grisha Levit test.c - unary_operator: only support historical handling of -t and its optional argument when not in posix mode - unary_test: print an error if the argument to -t is not a number - unary_operator: only make the argument to -t optional if the next argument is -a or -o and we are using the historical algorithm (argc >= 5), otherwise force it to be a number and print an error message if it's not From a report by Stephane Chazelas doc/bash.1,doc/bashref.texi - added note about test/[ sorting using the current locale with the `<' and `>' operators when in posix mode - added note about the integer argument to test -t being required when in posix mode - added note recommending against the use of test with 5 or more arguments in favor of combining multiple instances of test with && or || From a report by Stephane Chazelas 7/10 ---- jobs.c - start_job: turn off (on) the J_ASYNC flag depending on whether the job is being started in the foreground (background). It matters now that we use IS_ASYNC to determine whether to give the terminal back to the shell's process group. From a report by Grisha Levit bashline.c - command_word_completion_function: if we have a glob pattern that expands to multiple words, dispose of the list before returning NULL. From a report by Grisha Levit variables.c - makunbound: if we're unsetting a local variable, preserve the export attribute even if the variable's not in the temporary environment Tentative fix, compatible with other POSIX shells (except the BSD variants of ash) 7/12 ---- lib/sh/unicode.c - include if HAVE_LANGINFO_CODESET is defined, use nl_langinfo (CODESET) if HAVE_LANGINFO_CODESET is define, since we no longer check for nl_langinfo. This is what locale.c does. 7/15 ---- lib/readline/display.c - rl_redisplay: rudimentary code to handle some cases where the line buffer consumes more than the physical number of screen lines. It's not a screen editor, but it handles some of the common cases. From https://bugs.launchpad.net/ubuntu/+source/bash/+bug/2024628 7/17 ---- builtins/printf.def - getwidestr: make sure to force mbsrtowcs to convert the closing NULL so the wide-character string is NULL-terminated. From a report by Grisha Levit lib/sh/shmatch.c - strregerror: new function to return regerror(3) error message text in allocated memory - sh_regmatch: add an additional argument to hold the error message from regerror(3), in newly-allocated memory; generated by call to strregerror() on regcomp(3) failure From a patch by Grisha Levit externs.h - sh_regmatch: update extern declaration execute_cmd.c - execute_cond_node: pass pointer to buffer to return any error message; if sh_regmatch returns 2, use a non-NULL error buffer in the error message. lib/readline/bind.c - _rl_get_string_variable_value: return the values for active-region-start-color (_rl_active_region_start_color) and active-region-end-color (_rl_active_region_end_color) if they are set. Fixes bind -v issue reported by Sebastian Carlos - _rl_get_string_variable_value: if the history is stifled, make sure we report the history size as < 0 so the output of bind -v can be reused as input. Fixes bind -v issue reported by Sebastian Carlos 7/20 ---- pathexp.c - quote_string_for_globbing: don't bother backslash-quoting multibyte characters. It matters for macOS and doesn't make a difference for other systems, and those aren't going to be special globbing chars. From a report by Grisha Levit lib/glob/glob.c - glob_vector: normalize the pattern with fnx_fromfs so we make sure it's also in NFC on macOS (other systems don't care). This is a POSIX violation, but part of correcting the longtime HFS+ (and Finder) misfeature lib/readline/complete.c - rl_filename_completion_function: if the application sets a filename rewrite hook, use it to rewrite the filename we're matching as well as filenames read with readdir(). [SUPERSEDED BY CHANGE OF 10/4/2023] From a report by Grisha Levit 7/21 ---- pathexp.c - quote_string_for_globbing: see if we can get away without backslash- quoting single-byte characters that aren't CTLESC or characters that are special to EREs or BREs, even if they're quoted bashline.c - bashline_set_filename_hooks: convenience function to set the directory hook and the filename rewrite and stat hooks; call from other places in the file so these are set consistently bashline.h - bashline_set_filename_hooks: extern declaration builtins/complete.def - compgen_builtin: make sure to set the directory and filename hooks so compgen called from a script works more like compgen called from a word completion context. Inspired by a report by Grisha Levit doc/Makefile.in - bash.pdf: use groff -T pdf to create, instead of creating postscript and then using ghostscript 7/24 ---- jobs.c - start_job: unconditionally send SIGCONT to the job we just started, even if we think it was already running, to reduce the window for a potential race condition, and only change the state to RUNNING if the SIGCONT succeeds or if the process has terminated so we can clean it up later. Report from Earl Chew make_cmd.c - make_simple_command,make_bare_simple_command: now take line number as an argument; used instead of line_number to set the line associated with the simple command make_cmd.h - make_simple_command,make_bare_simple_command: change prototype parse.y,builtins/command.def,builtins.jobs.def - make_simple_command,make_bare_simple_command: change callers 7/25 ---- parse.y - read_token_word: try to keep track of whether we are parsing the first word of a simple command and set simplecmd_lineno if we are. We start a simple command when we return ASSIGNMENT_WORD or WORD in a command token position. This gives better values for $LINENO in multi-line simple commands - redirection: set simplecmd_lineno if it's -1, meaning this is a simple command with a leading redirection or a null command with only a redirection - pass simplecmd_lineno to make_simple_command if we are starting a simple command (the first ELEMENT) - parse_comsub,xparse_dolparen: save and restore simplecmd_lineno - simple_command,function_def production: reset simplecmd_lineno to -1 after parsing a simple command or a function declared as WORD() 7/26 ---- doc/bash.1,doc/bashref.texi - xpg_echo: add description of how this can be used to suppress echo option interpretation Inspired by report from Zachary Santer 7/27 ---- builtins/printf.def - printf_erange: make out-of-range errors conversion errors even if the conversion function fully consumes the argument. From a report by thomas@habets.se 7/31 ---- lib/readline/rltypedefs.h - rl_macro_print_func_t: typedef for a function to print macro key bindings lib/readline/bind.c - rl_macro_display_hook: hook function to call if the application wants to display a key sequence bound to a macro - rl_macro_dumper: if rl_macro_display_hook is non-NULL, call it to display the macro value (after `untranslating' it to add back any required backslashes) lib/readline/doc/rltech.texi - rl_macro_display_hook: add description pathexp.c - glob_char_p: add the characters that are special within BRE bracket expressions. Inspired by a report by Grisha Levit bashline.c - bind_keyseq_to_unix_command: allow whitespace to separate the keyseq and the command string, but require the command string to be surrounded by double quotes if it is. If we get a whitespace separator, call rl_macro_bind so we get the string `translated' according to how readline does it (backslash-escape sequences processed, etc.) - print_unix_command: function to print a key sequence and a unix command bound to it from cmd_xmap, using a space separator - print_unix_command_map: make sure to set rl_macro_display_hook to print_unix_command before calling rl_macro_dumper to print the bound command strings 8/1 --- builtins/read.def - read_builtin: saw_escape is now a count of CTLESCs we added to the input - read_builtin: since we increment saw_escape when we add a CTLESC, decrement it when we remove one from the input (\\\n, \\\0) - read_builtin: remove the CTLESC we added if we have a backslash- escaped NULL character ('\0') - read_builtin: if the input ends with an unescaped newline, remove the CTLESC we added From a report by Grisha Levit 8/2 --- builtins/history.def - history_builtin: if history -[anrw] is used without a filename argument, and HISTFILE is unset or null, return success immediately From a report by Grisha Levit doc/{bash.1,bashref.texi},lib/readline/doc/hsuser.texi - history: document that if filename is not supplied, and HISTFILE is unset or null, the [-anrw] options have no effect 8/3 --- pcomplete.c - gen_progcomp_completions: if the actual compspec that's run is something different from what's typed (full pathname, alias, etc.), set pcomp_curcmd to the word corresponding to the actual compspec, instead of what was typed. This is what eventually gets used by the compopt builtin's output. From a report by Grisha Levit execute_cmd.c - execute_coproc: in parent, close /dev/fd FIFOs only if we're not executing a shell function, like other calls to unlink_fifo_list(). Report from Hal Blackburn in https://savannah.gnu.org/support/index.php?110910 subst.c - do_compound_assignment: don't attempt to convert a global associative array to an indexed array with declare -g. From a report and patch by Grisha Levit 8/7 --- lib/readline/bind.c - rl_print_keybinding: new function, prints a the key bindings for a single function specified by name - rl_function_dumper: call rl_print_keybinding for every function name lib/readline/readline.h - rl_print_keybinding: extern declaration lib/readline/bind.c - _rl_untranslate_macro_value: make sure characters betweeen 128 and 159 (metafied control characters) are printed using the \M-\C- notation instead of directly writing the control character. From a report and patch by Grisha Levit - rl_get_keyname: if a key binding shadows a function or macro, print the ANYOTHERKEY binding as a null string From a report and patch by Grisha Levit - rl_untranslate_keyseq: make changes analogous to _rl_untranslate_macro_value so that we don't print control characters directly - rl_invoking_keyseqs_in_map: if the key corresponds to a keymap, use _rl_get_keyname to get its text representation, so this function returns consistent results - rl_translate_keyseq: if convert-meta is on, don't convert meta chars that are, e.g., \M-\C-@ (128) into \e followed by a NUL, resulting in truncated key sequences. Just don't honor convert-meta in this case 8/8 --- bashhist.c - bash_history_no_expand_chars: add the rest of the shell metacharacters so we don't get empty history events; tcsh does more or less the same thing Inspired by a report from Emanuele Torre 8/9 --- parse.y - grammar: change bare `time' and `!' productions to use nullcmd_terminator, which is a list_terminator or a `&' to allow things like `time &' and `! &' which the original version of POSIX interp 267 (the one I implemented) says we should support 8/11 ---- m4/bison.m4,m4/flexmember.m4,m4/locale_h.m4 - new m4 files from gettext-0.21.1 m4/gettext.m4,m4/intl.m4 m4/{codeset,extern-inline,fcntl-o,glibc2,glibc21,host-cpu-c-abi,iconv,intdiv0}.$ m4/{intl-thread-locale,intlmacosx,intmax,inttypes-pri,inttypes,inttypes_h}.m4 m4/{lcmessage,lib-ld,lib-link,lib-prefix,locale_h,lock,nls,po,printf-posix}.m4 m4/{progtest,pthread_rwlock_rdlock,size_max,stdint_h,strtoimax,threadlib}.m4 m4/{uintmax_t,ulonglong,visibility,wchar_t,wint_t,xsize}.m4 - updated m4 files from gettext-0.21.1 configure.ac,config.h.in - updated autoconf files with changes for gettext-0.21.1 MANIFEST - updated distribution manifest with new files from gettext-0.21.1 lib/sh/unicode.c - u32cconv: prefer nl_langinfo to locale_charset like locale.c: locale_isutf8() execute_cmd.c - uw_restore_lineno: an unwind-protect fuction to restore a saved line_number - execute_command: add an unwind-protect to restore line_number in case execute_simple_command longjmps back to top_level or a return context. Side effect of https://savannah.gnu.org/support/index.php?110919 - execute_for_command, execute_select_command, execute_case_command: add unwind-protect to restore line_number (could also do it for if, while, until but those don't modify line_number) - execute_command_internal: add_unwind_protect to restore line number for arith, cond, function def commands 8/14 ---- execute_cmd.c - execute_simple_command: don't call savestring on the_printed_command_except_trap if it's NULL. From a report by Grisha Levit 8/15 ---- lib/readline/rltty.c - prepare_terminal_settings: replace USE_XON_XOFF macro with private variable _rl_use_tty_xon_xoff (initially set to 1); if it's set to 0 disable the tty start and stop characters. Prep for making it a bindable variable setting builtins/read.def - edit_line: now takes a third argument saying whether or not to set rl_attempted_completion_function to NULL to use readline's default filename completion (the default) - read_builtin: new option -E to use readline and use the bash default completion (that is, leave rl_attempted_completion_function unchanged) From a suggestion by konsolebox back in 5/2021 doc/bash.1,doc/bashref.texi - read: document new -E option 8/17 ---- aclocal.m4 - BASH_CHECK_LIB_TERMCAP: add a check for bash_cv_termcap_lib == "libcurses"; set TERMCAP_LIB=-lcurses in this case 8/18 ---- subst.c - bash_variable_assignment_error: new function, implements default mode behavior for variable assignment errors - posix_variable_assignment_error: new function, implements posix mode behavior for variable assignment errors - parameter_brace_expand_rhs,expand_declaration_argument, do_assignment_statements: call posix_variable_assignment_error or bash_variable_assignment_error as appropriate 8/21 ---- lib/readline/misc.c - _rl_arg_dispatch: add the digits or other characters to rl_executing_keyseq if we're not calling _rl_dispatch. From a report by Grisha Levit print_cmd.c - print_simple_command: make sure that the_printed_command[0] == '\0' if we're printing a null simple command with no words Patch by Grisha Levit 8/22 ---- builtins/read.def - read_builtin: make `delim' variable local, pass to edit_line to set the last character of the returned line; change edit_line prototype. From a suggestion by Grisha Levit 8/23 ---- parse.y - comsub and funsub productions: don't reset eof_encountered to 0 in the action subst.c - function_substitute: unwind-protect eof_encountered so ignoreeof doesn't keep getting reset to 0 if PS1 includes a ${ ...;} command. From a report by Grisha Levit 8/25 ---- doc/{bash.1,bashref.texi} - clarify the cases where arrayname[@] doesn't refer to all the elements of an associative array (the unset builtin and test/[/[[ -v for now) Inspired by https://savannah.gnu.org/support/index.php?110924 8/28 ---- execute_cmd.c - execute_for_command,execute_select_command,execute_arith_command, execute_case_command: fix up order in which BASH_COMMAND is set, the DEBUG trap is run, and PS4 is printed. From a report by Grisha Levit lib/readline/colors.c - RL_COLOR_PREFIX_EXTENSION: make sure the custom suffix has a leading `.' Report and fix from Stefan Klinger lib/readline/parse-colors.c - free_color_ext_list: new function, called on parse error, frees the list of extensions - rl_reparse_colors: new public function, frees old extension list and reparses (presumably new) value of LS_COLORS; some minimal checking to make sure the value has really changed 9/1 --- parse.y - time_command_acceptable: make sure `time' is recognized as a reserved word after DOLPAREN and DOLBRACE. Report from Dima Korobskiy builtins/printf.def - printf_builtin: make 'C' equivalent to 'lc'; 'S' equivalent to 'ls' as POSIX specifies 9/6 --- eval.c - parse_command: don't run pending traps if we're parsing a command substitution From a report from Emanuele Torre 9/8 --- jobs.c - stop_pipeline: don't have the parent set the terminal pgrp; rely on child processes to do it to avoid potential race conditions. From a discussion beginning at https://lists.gnu.org/archive/html/bug-bash/2023-08/msg00087.html builtins/printf.def - bexpand: when parsing the string to be expanded, consume entire multibyte characters at a time, in case their encoding includes a '\\' - printf_builtin: when parsing the format string, consume entire multibyte characters at a time unless they begin with '\\' or '%', to avoid characters whose encoding contains those characters 9/11 ---- builtins/hash.def - hash_builtin: if the -p option is supplied without a name to hash, it's an error hashcmd.c - phash_remove: if the hash table is empty, return 1 like unalias does when the alias isn't in the hash table. Report from from Mike Jonkmans 9/13 ---- test.c - binary_test,unary_test: now static - cond_test: new function, calls binary_test or unary_test as appropriate and handles longjmps from test_syntax_error test.h - cond_test: extern declaration execute_cmd.c - execute_cond_node: call cond_test instead of binary_test or unary_test. From a report and patch from Grisha Levit doc/bash.1,doc/bashref.texi - hash: add some text about the mutual exclusivity of the -t, -p, and -d options 9/14 ---- subst.c,subst.h,externs.h - string_extract,string_extract_verbatim,string_extract_double_quoted, string_extract_single_quoted,skip_single_quoted,skip_double_quoted, extract_delimited_string,extract_heredoc_dolbrace_string, extract_dollar_brace_string,parameter_brace_expand,param_expand, extract_command_subst,extract_function_subst,extract_arithmetic_subst, extract_process_subst,extract_array_assignment_list, expand_array_subscript change function signatures to take size_t * instead of int * for arguments that are string indices or lengths, change callers - unquoted_substring,unquoted_member,string_extract, string_extract_double_quoted,skip_double_quoted, string_extract_single_quoted,skip_single_quoted, string_extract_verbatim,extract_delimited_string, extract_heredoc_dolbrace_string,extract_dollar_brace_string, parameter_brace_expand,param_expand,skip_matched_pair,skip_to_delim, skip_to_histexp,list_string,expand_string_dollar_quote, parameter_brace_expand_rhs,expand_array_subscript,shouldexp_replacement, expand_array_subscript,string_quote_removal: change variables used for string indexes to size_t [merge from size_t branch] arrayfunc.c - expand_compound_array_assignment: use size_t for string indices braces.c - brace_gobbler: use size_t for string indices parse.y - xparse_dolparen: change function signature to use size_t * instead of int * for string index, use size_t for string indices general.c, general.h - bash_tilde_find_word: change function signature to use size_t * instead of int * for string index lib/readline/readline.h - rl_reparse_colors: extern declaration lib/readline/doc/rltech.texi - rl_reparse_colors: document as public function 9/21 ---- pcomplete.c - it_init_aliases: move the free (alias_list) inside #ifdef ALIAS. Report and fix from Mike Jonkmans redir.c - redirection_error: set W_NOPROCSUB in the word redirection_error creates to call redirection_expand builtins/bind.def - bind_builtin: use CASE_HELPOPT so the long doc gets printed when --help is supplied builtins/getopts.def - getopts_builtin: change to use no_options() 9/22 ---- execute_cmd.c - execute_select_command: in posix mode, an invalid selection variable name is a fatal error, just like with `for'. This is compatible with ksh93 and mksh 9/25 ---- subst.c - parameter_brace_expand_length: rearrange the code slightly to reduce the number of find_variable calls. This matters if the variable is dynamic and produces a new value each time (e.g., RANDOM). 9/26 ---- builtins/shift.def - shift_builtin: if get_numeric_arg returns a number out of range and we're going to print an error message, make sure to skip over an argument of `--' so we can print the right argument builtins/break.def - break_builtin,continue_builtin: ditto with get_numeric_arg and `--' 9/28 ---- builtins/history.def - history_builtin: changed error message if the numeric argument to -d is invalid 9/29 ---- subst.c - array_length_reference: include the open bracket in the error message passed to err_badarraysub; it looks cleaner 10/2 ---- jobs.c - delete_all_jobs: if we are clearing the jobs list, reset the stats about running and dead jobs and child processes Report from Koichi Murase 10/4 ---- lib/readline/complete.c - rl_completion_rewrite_hook: rewrite hook applied to the completion word before comparing against the possible completions from the file system (modified by rl_filename_rewrite_hook) - rl_filename_completion_function: call rl_completion_rewrite_hook, if set by the calling application, on the filename portion of the word to be completed lib/readline/readline.h - rl_completion_rewrite_hook: extern declaration bashline.c - bash_set_filename_hooks: set rl_completion_rewrite_hook to bash_filename_rewrite_hook now that we've separated them From a report and patch from Grisha Levit ; original report from Stefan H. Holek lib/readline/doc/rltech.texi - rl_completion_rewrite_hook: document 10/5 ---- lib/readline/complete.c - rl_complete_internal: use `%' as an op character to indicate that we should just display the completions, like '?' in previous versions (nothing uses this yet) - rl_possible_completions: set last_completion_failed to 0 before we pass '?' to rl_complete_internal so we don't worry about trying to insert a single completion 10/6 ---- subst.c - string_transform: use ansicstr directly for the `E' transformation so we don't accidentally remove too many escapes. From a report and patch from Grisha Levit execute_cmd.c - execute_arith_for_command: make sure a failed init expression resets loop_level so break and continue work correctly From a report and patch from Grisha Levit - execute_for_command: set line_number for invalid identifier error messages; restore before returning error From a report and patch from Grisha Levit lib/readline/misc.c - _rl_set_insert_mode: manage RL_STATE_OVERWRITE when modifiying insert mode doc/bash.1,doc/bashref.texi - some style and formatting changes from Bjarni Ingi Gislason subst.c - getifs: now returns allocated memory to insulate it from changes to $IFS builtins/read.def - read_builtin: since getifs now returns allocated memory, note when we need to free ifs_chars and free it before returning. Reported by Robert Elz in https://www.austingroupbugs.net/view.php?id=1778#c6513 10/7 ---- pathexp.c - glob_char_p: add more of the extglob pattern characters; it doesn't hurt to quote them Report and patch from Grisha Levit - unquoted_glob_pattern_p: don't treat a `(' immediately following a `/' as a potential globbing character - unquoted_glob_pattern_p: only treat the extended glob characters followed by a `(' as a glob pattern if extended_glob is enabled - unquoted_glob_pattern_p: an unquoted backslash isn't treated specially any more, but if it's followed by a CTLESC, you still don't treat the next character as an unquoted globbing char Report and patch from Grisha Levit 10/9 ---- builtins/common.c - get_numeric_arg: if fatal > 0, set the exit status to EX_USAGE and call jump_to_top_level with EXITPROG or DISCARD. We don't need to do everything that throw_to_top_level() does here (it's really meant for signals and other exceptional failure conditions). builtins/cd.def - cd_builtin: exit with EX_USAGE on too many arguments builtins/shift.def - shift_builtin: return EX_USAGE if get_numeric_arg returns 0, indicating that the argument wasn't a valid number builtins/history.def - shift_builtin: return EX_USAGE if get_numeric_arg returns 0 builtins/exit.def - exit_or_logout: if get_exitstat (which calls get_numeric_arg) returns a value > EX_SHERRBASE, indicating an error, just return that to the caller and let the caller deal with it. This means that `exit xyz' is no longer a fatal error, but can potentially cause a non-interactive posix-mode shell to exit because exit is a special builtin builtins/return.def - return_builtin: if a non-interactive shell in posix mode gets an invalid numeric arg from get_exitstat, return immediately and let the caller deal with exiting All prompted by a report by Martin Schulte 10/13 ----- pathexp.c - unquoted_glob_pattern_p: restore some of the special treatment of backslash followed by CTLESC removed on 10/7 Report and patch from Grisha Levit parse.y - parse_matched_pair: don't add an extra CTLESC after reading \CTLESC, like in other parts of the parser subst.c - dequote_string: don't drop trailing CTLESC in a string with more than a single character Report and patch from Grisha Levit lib/sh/strtrans.c - ansicstr: handle $'\c^A' and $'\c^?' correctly when being expanded by the parser (flags&2). The parser passes these as \c^A^A and \c^A^?, respectively, so we should strip the quoting CTLESC. Report from Grisha Levit subst.[ch] - extract_dollar_brace_string: now global so brace expansion can use it braces.c - brace_gobbler: use extract_dollar_brace_string if we see ${ with the appropriate value of QUOTING, so we don't have to teach brace expansion more shell syntax. Report from Emanuele Torre - brace_gobbler: call the word extraction functions with SX_NOALLOC so we don't have to allocate memory we're just going to free 10/16 ----- builtins/read.def - read_builtin: return EX_MISCERROR (2) if there is an error trying to assign to one of the variables. This is what the newest POSIX draft specifies. variables.c - dispose_variable_value: do the right thing for att_nofree vars - makunbound: call dispose_variable_value instead of using inline code braces.c - brace_gobbler: make sure to set no_longjmp_on_fatal_error around calls to extract_dollar_brace_string 10/17 ----- braces.c - brace_gobbler: set SX_NOLONGJMP|SX_NOERROR in the flags passed to extract_command_subst; make sure no_longjmp_on_fatal_error is set before that call - brace_gobbler: revert 10/13 change to use extract_dollar_brace_string 10/20 ----- parse.y - xparse_dolparen,parse_string_to_command: if SX_NOERROR is set in FLAGS, add PST_NOERROR to parser_state. Not much effect yet. - parse_matched_pair: if PST_NOERROR is set in parser_state, don't print an error message if we hit EOF, just return an error. A start at using PST_NOERROR to suppress error messages, not just duplicate ones. We'll see how this goes before adding more builtins/printf.def - decodeprec: decode the precision into an intmax_t; clamp the return value at INT_MAX - printf_builtin: update to posix interp 1647 (even though it's about fprintf(3)) and output a NUL byte if %lc is supplied a null argument. - printf_builtin: fix case where %Q is supplied with a precision in the format string - printf_builtin: fix case where %Q is supplied with a precision greater than INT_MAX - getwidestr,getwidechar: handle case where there is no argument supplied; return NULL or NUL - convwidestr: allow a precedence of 0 for %ls - getint: don't call getintmax any more, just use the same code style inline; getintmax will consume an extra argument on an error Report and patches from Grisha Levit - printf_builtin: handle field width and precision overflow from getint() by ignoring the argument (fieldwidth = 0, precision = -1) 10/26 ----- jobs.c - wait_for: rearrange code that sets the SIGINT handler to wait_sigint_handler and saves the old handler to old_sigint_handler to avoid delay before assigning the handler Report from Wenlin Kang - wait_sigint_handler: if cur_sigint_handler (what restore_sigint_handler) just restored or ignored) is INVALID_SIGNAL_HANDLER, set the appropriate SIGINT handler with set_sigint_handler before sending ourselves SIGINT 10/30 ----- general.c - legal_number: use the same test (isspace(3)) to skip trailing whitespace that strtoimax uses to skip leading whitespace. Report and patch from Paul Eggert general.c,general.h - legal_number: renamed to valid_number, changed all callers - legal_identifier: renamed to valid_identifier, changed all callers - legal_alias_name: renamed to valid_alias_name, changed all callers lib/sh/compat.c - legal_number: definition of legal_number here for backwards compat with old loadable builtins - legal_identifier: definition of legal_identifier here for backwards compat with old loadable builtins - legal_alias_name: same subst.c - brace_expand_word_list: if brace_expand returns NULL, dummy up a single-element vector with a copy of the original word so we can add it to the result list unchanged. Inspired by https://savannah.gnu.org/support/?110948 11/3 ---- lib/readline/text.c - readstr: set of functions to read a string from the keyboard, using rl_line_buffer for temporary storage, with minimal editing and an optional caller-supplied completion function. Doesn't use the callback framework yet since none of the functions are public - rl_execute_named_command: new bindable function to read a bindable command name (from the funmap_names array) and execute it, with command name completion on SPACE and TAB lib/readline/readline.h - rl_execute_named_command: new extern declaration lib/readline/rlprivate.h - new extern declarations for the readstr function framework lib/readline/funmap.c - execute-named-command: new bindable function name lib/readline/emacs_keymap.c - rl_execute_named_function: bound to M-x by default lib/readline/doc/rluser.texi - execute-named-command: document as bindable function name with its default binding to M-x in emacs mode 11/6 ---- lib/readline/search.c - _rl_nsearch_dispatch: use ^V/^Q for rl_quoted_insert into the search string lib/readline/isearch.c - _rl_isearch_dispatch: use ^V or anything bound to rl_quoted_insert for rl_quoted_insert into the search string lib/readline/terminal.c - add support for new BE/BD and PS/PE bracketed-paste capabilities. Nothing uses them yet. 11/10 ----- builtins/enable.def - dyn_load_builtin: if we don't find a pathname argument without a slash in BASH_LOADABLES_PATH, convert it to a pathname with a slash before calling dlopen, to force the loader to look in the current directory (Linux, for example, will not). 11/14 ----- examples/loadables/fdflags.c - fdflags_builtin: only parse the setspec once, since parsing uses strtok. Report and patch from Emanuele Torre 11/15 ----- builtins/enable.def - dyn_load_builtin: if BASH_LOADABLES_PATH is set, use only it: don't fall back to looking in the current directory. This changes the historical behavior and brings the path behavior more in line with PATH, but not CDPATH. 11/20 ----- lib/readline/complete.c - rl_complete_internal: add `|' as a character for rl_complete_internal that also just displays the completions, since `%' is overloaded by rl_menu_complete 11/23 ----- examples/loadables/strptime.c - strptime: new loadable builtin, interface to strptime(3). Takes a date-time string as its arguments and tries to parse it according to a number of built-in formats. If successful, it outputs the result as a number of seconds since the epoch. Understands some handy shorthands like "now" and "tomorrow". 11/24 ----- tests/unicode1.sub - update the driver function using namerefs and ${!x[@]} - zh_TW.BIG5: fixed some problems with the test encodings. Report and patches from Grisha Levit 11/25 ----- tests/glob2.sub - zh_HK.big5hkscs -> zh_TW.big5 to work around macOS 14 issues with BSD-based locale definition bashline.c - bash_execute_unix_command: check count != 1 instead of count > 1 to handle negative arguments lib/readline/bind.c - rl_trim_arg_from_keyseq: change the parsing_digits logic to handle multiple instances of `-' the same way rl_digit_loop does (collapse to 1) - rl_trim_arg_from_keyseq: fix typo to correctly recognize a `-' after a key sequence bound to universal-argument - rl_trim_arg_from_keyseq: trim negative argument for a key sequence that is shadowed by a longer key sequence (ANYOTHERKEY) Report and patches from Grisha Levit 11/27 ----- support/shobj-conf - solaris2: change PIC option for non-gcc Solaris 2 builds lib/malloc/malloc.c - binsize: make sure indexing into binsizes casts to size_t (unsigned) examples/loadables/getconf.c - protect more of the _SC defines with #ifdefs for different systems examples/loadables/fdflags.c - ALLFLAGS: add some more flag values from Solaris/AIX/HP-UX 11/28 ----- subst.c - function_substitute: unwind return_catch_flag unconditionally Report and patch from Grisha Levit - function_substitute: save getopts state and restore it if the funsub declares a local copy of OPTIND Report and patch from Grisha Levit lib/readline/display.c - rl_redisplay: make sure lpos is initialized to 0 if local_prompt is NULL shell.c - run_startup_files: make sure ssh_reading_startup_files is set to 0 in more places; still dependent on SSH_SOURCE_BASHRC builtins/evalstring.c - should_suppress_fork: make sure ssh_reading_startup_files is 0 in order to suppress a fork. It will never be non-zero unless someone enables SSH_SOURCE_BASHRC. Fixes https://issues.redhat.com/browse/RHEL-5653 variables.c - posix_unbind_tempvar: new function to implement POSIX interp 1009, which says that `x=4 unset x' not only unsets the variable in the temporary environment but the most recent scope as well. variables.h - posix_unbind_tempvar: extern declaration builtins/set.def - unset_builtin: if we are unsetting a variable that is in some temporary environment, and the shell is in posix mode, call posix_unbind_tempvar to implement the right posix behavior doc/bashref.texi - update posix mode section with change to unset for interp 1009 12/1 ---- lib/readline/readline.c - _rl_subseq_result: add check before _rl_do_lowercase_version that _rl_to_lower actually results in a different key sequence Report and patch from Grisha Levit lib/readline/isearch.c - _rl_isearch_dispatch: if the function bound to a key and the function bound to the result of _rl_to_lower for that key are both rl_do_lowercase_version, just insert the key into the search string 12/5 ---- parse.y,parser.h - heredoc_string: new variable to indicate whether or not we're reading a here-document from an alias (a pushed string). Set in gather_here_documents individually for each call to make_here_document. parse.y - read_a_line: if heredoc_string is non-zero, use shell_getc instead of yy_getc to get the right alias processing - shell_getc: if heredoc_string is non-zero, don't add a space to the end of the alias -- it can mess up the here-document delimiter if the next character is a newline Report and sample patch from gldrk parse.y,make_cmd.c - read_a_line,make_here_document: if we're using shell_getc to read the body of a here-document, let it manage line_number 12/8 ---- bashline.c - bash_dequote_filename: fix an off-by-one error that didn't remove a closing single quote in a filename if it was backslash-escaped. Report by Ole Tange , fix from Grisha Levit lib/readline/terminal.c - _rl_check_ansi_terminal: check whether or not rl_terminal_name appears to be an ANSI/ECMA-48 terminal. Check some common terminal types and then check whether some common terminal capabilities (ce, nd, ho, up) begin with CSI (ESC-[) and then have a correct subsequent character - _rl_term_isansi: new variable, holds readline's idea about whether the value of rl_terminal_name is an ANSI/ECMA-48 terminal. Initialized to RL_ANSI_TERM_DEFAULT; set to the result of _rl_check_ansi_terminal if initialized to 0 and on a `non-dumb' terminal From an idea by John Tsiombikas in 11/2023 lib/readline/rlconf.h - RL_ANSI_TERM_DEFAULT: define to 1 (yes) or 0 (no) to tell readline whether or not to assume it's running on an ANSI/ECMA-48 terminal. The default is 1 (yes) 12/11 ----- print_cmd.c - make_command_string_internal: when printing a coproc, print the coproc name only if the coproc command is not a simple command. Report from Albert Akchurin 12/14 ----- doc/bash.1,doc/bashref.texi - bind: clarify meaning of `re-read' as being individual arguments to `bind'. Report from Dan Jacobson - history expansion: add that the quick substitution character is only active when it's the first character on the line. Suggested by Dale R. Worley lib/readline/histexpand.c - history_expand: if history_quoting_state is "'", don't try to interpret the quick substitution syntax at the start of the string. Report from Sundeep Agarwal execute_cmd.c,execute_cmd.h - save_lastarg: new function, returns saved value of $_ in allocated memory parse.y - execute_variable_command: instead of inline code, call save_lastarg mailcheck.c - check_mail: instead of inline code, call save_lastarg - check_mail: instead of calling bind_variable, use bind_lastarg parse.y - decode_prompt_string: use save_lastarg() and bind_lastarg() to preserve the value of $_ across prompt string decoding. Patch from Grisha Levit 12/21 ----- lib/readline/doc/readline.3 - updates to some vi-mode bindings Patch from Josh Brobst builtins/setattr.def - set_or_show_attributes: in posix mode, supplying an argument that is not an identifier when not acting on functions is a utility error. This will cause a non-interactive shell to exit 12/27 ----- print_cmd.c - command_print_word_list,print_select_command_head,print_case_command_head, print_function_def: if we're in pretty-print mode, dequote words before printing them Report from Seth Sabar shell.c - execute_profile_file,execute_bashrc_file: separate the (duplicated) code that sources the various profile and bashrc files out into separate functions - find_bashrc_file: new function, returns the name of the interactive shell startup file that should be executed if --rcfile (--init-file) is not supplied; currently returns DEFAULT_BASHRC - bashrc_file: default to NULL, making it just the filename supplied as an argument to --rcfile/--init-file - execute_bashrc_file: execute whatever find_bashrc_file() returns if bashrc_file is NULL From a patch from Allison Karlitskaya 1/3/2024 -------- print_cmd.c - print_case_clauses: if we're printing a comsub for subsequent parsing, don't add a newline before the first case clause, since it adds a token after the `in' that will allow reserved words to be parsed. Report from Oguz 1/5 --- lib/readline/doc/rltech.texi,lib/readline/doc/readline.3 - Note that since quoted characters are possible, the line readline() returns may contain embedded newlines. From a report by Martin Buck 1/8 --- parse.y - parse_compound_assignment,parse_string_to_word_list: if we call reset_parser, directly or indirectly via yyerror, make sure to set the pushed_strings member of the saved parser state to NULL, since reset_parser already freed it and we don't want to try and restore it in restore_parser_state. From a report by Nathan Mills 1/10 ---- builtins/hash.def, builtins/ulimit.def - add some calls to sh_chkwrite where there is builtin output lib/sh/eaccess.c - sh_stat: use strcpy/strcpy when constructing pbuf instead of strcpy/strcat lib/sh/tmpfile.c - sh_mktmpname,sh_mktmpfd,sh_mktmpdir: use snprintf (filename, PATH_MAX, ...) instead of sprintf (filename, ...) 1/11 ---- configure.ac - unconditionally AC_DEFINE(PGRP_PIPE), to prevent the problem with a pipeline and a DEBUG trap containing an external command described in https://lists.gnu.org/archive/html/bug-bash/2024-01/msg00037.html 1/12 ---- jobs.c - wait_for_any_job: if we're executing a funsub/varsub, do the wait even if the jobs list is frozen, but don't remove the job from the table or change its notification status Report from Oguz subst.c - uw_unbind_variable: unset the first instance of the named variable, don't follow namerefs. This is for REPLY in a varsub if it's made a nameref. Report from Oguz shell.c - main: call compat_init() so the linker drags in the old compatibility functions from lib/sh/compat.c. Primarily for use by existing loadable builtins parse.y - parse_compound_assignment: handle error case (wl == &parse_string_error) before restoring the parser state from ps Report from Grisha Levit 1/13 ---- lib/readline/complete.c - compute_lcd_of_matches: move a strlen out of a call to MBRTOWC by computing the length of the entire string once and then using an offset from it while going through the loop 1/15 ---- parse.y - parse_compound_assignment: treat parsing a string following (( (read by parse_dparen but not an arithmetic command) the same as expanding an alias when fixing up the pushed string list Report from Grisha Levit 1/18 ---- doc/bash.1 - typos and style fixes to satisfy groff warnings; remove uses of the `CW' constant-width font; use .EX/.EE for examples; use consistent style for double-quoted strings; fix inconsistent inter-paragraph spacing Fixes from G. Branden Robinson 1/19 ---- test.c - binary_test: make sure all calls in posix mode use TEST_LOCALE for locale-specific string comparisons shell.c - find_bashrc_file: remove 1/20 ---- shell.c - exit_shell: don't try to call rl_deprep_terminal, regardless of the readline state we inherited, if bash_readline_initialized is 0 Fixes bug reported by Oguz - shell_reinitialize: reset startup_state and reading_shell_script to 0; reset debugging_mode to 0 1/22 ---- builtins/shopt.def - reset_shopt_options: reset debugging_mode to 0 builtins/set.def - reset_shell_options: reset interactive_comments to 1 braces.c - brace_expand: if the first call to brace_gobbler consumes the entire string, don't try to call it again - brace_expand: check that i < tlen before checking to see if expand_seqterm left more of the string unconsumed - brace_gobbler: if extract_command_subst hits the end of the string without closing the command substitution, make sure we return 0 as well as set *i = tlen From a fuzzing report by Nathan Mills 1/23 ---- subst.c - function_substitute: unwind-protect current_builtin and this_shell_builtin like we do this_shell_function From a fuzzing report by Nathan Mills 1/24 ---- braces.c - brace_gobbler: handle nofork command substitutions; skip over any enclosed command and any braces it contains bashline.c - check_redir: return 0 if we're in a nofork comsub, so we will attempt command word completion if appropriate 1/26 ---- lib/readline/readline.h - RL_STATE_READSTR: new state flag saying we are reading an arbitrary string from the keyboard using rl_readstr lib/readline/rlprivate.h - READSTR_FREEPMT: new context state flag for rl_readstr; indicates that we have allocated a new prompt with _rl_make_prompt_for_search and we should free it with rl_restore_prompt because rl_clear_message will not do it lib/readline/text.c - _rl_readstr_init,_rl_readstr_cleanup: set and unset RL_STATE_READSTR - _rl_unsave_saved_readstr_line: free any rl_undo_list, which we may have accumulated while reading the string, before restoring the line Fixes leaks reported by sparrowhawk996@gmail.com. - _rl_readstr_init: set the READSTR_FREEPMT context flag after we call _rl_make_prompt_for_search, which calls rl_save_prompt - _rl_readstr_sigcleanup: new function to call from signal cleanup; restores prompt and calls _rl_readstr_cleanup to free the readstr context - _rl_readstr_restore: if the RL_READSTR_FREEPMT flag is set in the context, call rl_restore_prompt to deallocate the prompt we created lib/readline/callback.c - rl_callback_sigcleanup: call _rl_readstr_sigcleanup to deallocate readstr state on a signal lib/readline/text.c - rl_execute_named_command: free COMMAND before returning 1/27 ---- lib/readline/rlprivate.h - SF_FREEPMT: new flag for non-incremental search contexts: means we saved the prompt data with _rl_make_prompt_for_search and need to restore it with rl_restore_prompt, since rl_clear_message will not lib/readline/search.c - _rl_nsearch_sigcleanup: new function that if calls rl_restore_prompt to clean up the saved prompt data if necessary before calling the context cleanup function, avoids memory leak - _rl_nsearch_init: set the SF_FREEPMPT flag after calling _rl_make_prompt_from_search so we can clean it up properly on errors - _rl_nsearch_abort,_rl_nsearch_dosearch: call rl_restore_prompt only if cxt->sflags includes SF_FREEPMT, manage SF_FREEPMT state lib/readline/signals.c - _rl_state_sigcleanup: call _rl_nsearch_sigcleanup lib/readline/callback.c - _rl_callback_sigcleanup: call _rl_nsearch_sigcleanup 1/29 ---- builtins/cd.def - change_to_directory: don't try to canonicalize a NULL path that's NULL after make_absolute. From a report by Kerin Millar jobs.c - retrieve_proc_status,delete_proc_status: external interfaces to bgp_search and bgp_delete, respectively; these take an argument that says whether or not to block SIGCHLD jobs.h - retrieve_proc_status,delete_proc_status: extern declarations builtins/wait.def - check_bgpids: new function to check whether a requested PID is in the bgpids table (retrieve_proc_status) and optionally delete it if it is (posixly_correct delete_proc_status) while returning its status. If the PID isn't in the bgpids table, return -1 - wait_builtin: if -n is supplied with pid/job arguments, use check_bgpids to check the bgpids table for any of the requested pids. Tagged for bash-5.3, might need another option From a report by Steven Pelley 1/30 ---- redir.c - redir_open: assume the AFS bug with O_CREAT and existing files in protected directories has been fixed over the years, so take out the workaround. From a report by Etienne Champetier jobs.c - wait_for_any_job: if the jobs list is frozen and we're running a funsub, mark the job as notified so it gets cleaned up later - wait_for_any_job: if we're in posix mode, we should remove the job from the job list and not add it to the bgpids list, as posix requires builtins/set.def - set_shellopts: use ASS_FORCE in the call to bind_variable so we don't have to mess with temporarily turning off readonly builtins/shopt.def - set_bashopts: same Report by Emanuele Torre 2/1 --- builtins/printf.def - vblen: make it a size_t to avoid going negative on underflow - clearerr, ferror, fflush: don't test these if vflag is set and we're writing to a string (multiple places) - vbprintf: pass through failure returns (< 0) from vsnprintf to the caller(s) so we can return on errors - printf_builtin: remove redundant test for empty or missing format string - PF: call builtin_error if vflag is set, sh_wrerror otherwise (via PRETURN) - PF: use PRETURN so we can get partial output to the variable on error (if vflag is set) - PRETURN: free conv_buf before attempting the write (it would get cleaned up on the next call, but why not) - PRETURN: clean up vbuf only if vflag is set, and clean it up on error (it would get cleaned up on the next call, but...) Fixes from Grisha Levit 2/2 --- doc/bash.1,doc/bashref.texi - word expansions: make it clearer that quote removal is one of the shell word expansions execute_cmd.c,execute_cmd.h - currently_executing_command: no longer static, so other parts of the shell can save and restore it if necessary subst.c - function_substitute: unwind-protect currently_executing_command, since parse_and_execute can overwrite it in the current shell execution context. Report from Grisha Levit lib/readline/readline.c - readline_common_teardown: new function from the guts of readline_internal_teardown, manages and deallocates rl_undo_list. lib/readline/callback.c - rl_callback_handler_remove: if we're removing the line handler while we still have an undo list, we didn't call readline_internal_teardown. Call readline_common_teardown to manage the undo list in case we are calling this from a signal handler but not exiting the program. Fixes leaks reported by sparrowhawk996@gmail.com. variables.c - initialize_shell_variables: use ASS_FORCE when binding SHELLOPTS or BASHOPTS if we get them from the environment, in case they've already been created as shell variables and set to readonly (like changes from 1/30) - set_ppid: use ASS_FORCE in the call to bind_variable instead of temporarily turning off att_readonly subst.c - extract_delimited_string: pass FLAGS down to skip_single_quoted and skip_double_quoted so we propagate SX_COMMAND and SX_COMPLETE properly. Fixes bug reported by A4-Tacks 2/3 --- bashline.c - initialize_readline: add `bash-vi-complete' as a bindable command name so users can bind other key sequences to it - vi_advance_point: function to advance point by one character even in the presence of multibyte characters - bash_vi_complete: call vi_advance_point instead of just incrementing rl_point 2/5 --- doc/bash.1,lib/readline/doc/readline.3 - minor updates to handle old versions of troff and groff warnings Fixes from G. Branden Robinson 2/7 --- jobs.c - notify_of_job_status: in non-interactive shells, if we're not going to print information about a terminated background job (SIGINT/SIGTERM/SIGPIPE), don't mark it as notified, in the same way that we don't mark background jobs that exit cleanly as notified Inspired by a discussion with Robert Elz and https://lists.gnu.org/archive/html/bug-bash/2024-01/msg00189.html 2/9 --- lib/readline/history.c - DEFAULT_HISTORY_GROW_SIZE: bump up to 256 - real_history, real_history_size: the true history array and its allocated size. the_history begins a user-visible window into this list, extending for history_size entries - history_list_grow_size: compute how much to extend the history list; increase by roughly sqrt(history_size), don't extend by less than (new) DEFAULT_HISTORY_GROW_SIZE - history_list_resize: possibly resize/realloc real_history and reset the_history to real_history and history_size to real_history_size - advance_history: make the `window' that the_history provides into real_history advance to simulate dropping an initial entry; don't need to memmove all the entries any more - add_history: call advance_history as necessary to move the_history window - add_history: call history_list_resize as necessary Report and initial patch from Casey Johnson 2/12 ---- bashline.c,lib/readline/bind.c,lib/readline/display.c - replace some sprintf calls with snprintf to appease clang execute_cmd.c - execute_simple_command: unwind-protect this_command_name, since we assign it to something we will free if unwind-protects are run - execute_function: unwind-protect currently_executing_command, since execute_command_internal will change it to something will will free if unwind-protects are run Report from Grisha Levit - execute_function: if we didn't compile with debugger support, restore currently_executing_command after run_debug_trap, like in other cases 2/13 ---- builtins/declare.def - declare_invalid_opts: move the code that checks for invalid option combinations to its own function; it returns 0 for no errors and a return code for declare_internal to return otherwise Inspired by report from Grisha Levit - declare_invalid_opts: make -A -a and -A +A/-a +a option combinations invalid Report from Grisha Levit 2/14 ---- bashline.c,general.c,subst.c,lib/sh/pathphys.c,builtins/read.def, support/bashversion.c - changes for cygwin and msys (no new features) bashline.c,builtins/fc.def,general.c,parse.y,stringlib.c,subst.c,variables.c - MSYS-specific changes to support \r\n line endings and DOS-style paths configure.ac,cross-build/msys32.cache,support/config.rpath,support/shobj-conf - MSYS-specific build options configure.ac,config.h.in,m4/unlocked-io.m4,general.h,include/unlocked-io.h MANIFEST - adapt Gnulib's unlocked-io module for faster stdio functions. From Grisha Levit lib/malloc/malloc.c,lib/malloc/table.h,lib/sh/strlcpy.c - fixes for ISO C 23 Patch from Collin Funk 2/16 ---- jobs.c - waitchld: note whether or not we are executing in a signal handler context and break out of the loop if we receive a terminating signal instead of calling termsig_handler, which can execute unsafe functions From https://issues.redhat.com/browse/RHEL-22295 variables.h - valid_readonly_attrs,invalid_readonly_attrs: new defines for declare to use to validate attempts to modify attributes of readonly variables builtins/declare.def - declare_internal: reject attempts to modify (set or unset) attributes in invalid_readonly_attrs for readonly variables. From a discussion with Grisha Levit 2/17 ---- variables.h - add att_readonly to valid_readonly_attrs; it's valid to use declare -r var ; declare -r var Makefile.in,builtins/Makefile.in - add unlocked-io.h to dependencies and variable contents where appropriate - add additional include files to INSTALLED_INCFILES builtins/pushd.def,examples/loadables/necho.c - fix order of includes to get unlocked-io.h From Grisha Levit subst.c,subst.h - rename quote_rhs -> quote_nosplit 2/19 ---- subst.c - quote_var_value: break the code that quotes a variable value ($x, ${x}, ${x[n]}, etc.) into a separate inline function and call it from param_expand and parameter_brace_expand_word so they do it consistently - param_expand: quote $* appropriately using quote_nosplit if we are in a place where word splitting will not occur (PF_NOSPLIT2), expand_no_split_dollar_star == 1 but quoted is 0 - quote_var_value: use quote_nosplit if we are in a place where word splitting will not occur (PF_NOSPLIT2) but quoted == 0. Fixes bug with IFS=$'\1' reported by alex xmb sw ratchev arrayfunc.c - array_value: make sure to quote ${A[*]} appropriately if not quoted (quoted == 0) and in a place where word splitting does not occur (flags & AV_ASSIGNRHS); the caller expects array_value to quote in this case subst.c - parameter_brace_expand_word: make sure to call array_value with AV_ASSIGNRHS if we are expanding unquoted ${A[*]} in a place where word splitting does not occur with a non-null $IFS; array_value will quote appropriately here - parameter_brace_expand_word,param_expand: use quote_var_value when expanding $N and ${N} doc/bash.1,doc/bashref.texi - HISTIGNORE: clarify the description a little to emphasize that lines matching one of the patterns are not saved in the history list From https://savannah.gnu.org/support/index.php?111020 2/20 ---- builtins/evalfile.c - FEVAL_RETRY: if set in FLAGS, _evalfile will retry an interrupted open - _evalfile: if open() returns -1, FEVAL_RETRY is set in FLAGS, and errno == EINTR, retry the open after checking for interrupts or terminating signals - maybe_execute_file,force_execute_file: pass FEVAL_RETRY in flags bashhist.c - load_history: retry read_history if it returns EINTR after checking for interrupts or terminating signals lib/readline/bind.c - _rl_read_init_file: retry the open once if it's interrupted due to a signal. If we are at a point where readline has installed its signal handlers, check for signals readline handles From a patch from Grisha Levit 2/21 ---- subst.c - make_internal_declare: now takes two strings of option arguments: those to turn on (`-') and those to turn off (`+') - expand_declaration_argument: simplify the code; handle `-' and `+' optionsns; always call make_internal_declare From a patch from Grisha Levit 2/22 ---- aclocal.m4 - add additional character to test for broken wcwidth (0x200b) lib/readline/display.c - update_line: use IS_COMBINING_CHAR instead of UNICODE_COMBINING_CHAR plus WCWIDTH; it doesn't make sense on systems where wcwidth isn't broken 2/27 ---- print_cmd.c - print_case_clauses: if one of the case command pattern lists begins with the word `esac' (unquoted), precede the pattern list with `(', since it had to be there originally to get through the parser. Report by Emanuele Torre 2/29 ---- general.c,general.h - string_to_rlimtype: takes a second ENDP argument, like strtol, so the caller doesn't have to check that the string is all digits, but can optionally check for and disallow a 0x prefix 3/2 --- subst.c - dequote_list: unset the W_QUOTED flag in the word after dequoting it - parameter_brace_expand_rhs: if the word in the list returned by expand_string_for_rhs has W_QUOTED set, but the string being expanded was not quoted, turn on the W_QUOTED in the returned word so we can potentially avoid word splitting - expand_word_internal: if CTLNUL is a IFS character, don't add quoted null strings to istring if we're going to be word splitting, since they will be treated as word delimiters 3/6 --- examples/loadables/getconf.c - getconf_builtin,getconf_all: changes for POSIX interp 1808 proposal to allow an optional pathname argument with -a that forces its use for pathconf variables https://www.austingroupbugs.net/view.php?id=1808 3/11 ---- builtins/printf.def - printf_builtin: %lc/%ls (%C/%S) should only work on wide characters when in a locale where MB_CUR_MAX > 1. Report by Emanuele Torre 3/12 ---- jobs.c - notify_of_job_status: don't mark terminated background jobs or stopped jobs as notified in shells started to run -c command From a report by Greg Wooledge - notify_of_job_status: make the final marking the dead job as notified a true catch-all so we can put a debugging message in there to see if there are other missing cases - notify_of_job_status: don't mark terminated jobs that died due to a signal the shell trapped (on the assumption that the shell received the signal, too) as notified, since we don't report on it in the JDEAD case if the signal is trapped 3/13 ---- builtins/printf.def - PF: print an error message if printf returns an error or leaves ferror(stdout) true From a report by Paul Eggert 3/14 ---- builtins/printf.def - getint: now takes an int argument and returns it if the conversion overflows an int; changed callers in printf_builtin - getint: consolidate checks for overflow. This changes the behavior if the argument overflows an intmax_t: it returns a field width of 0 (the overflow result) instead of -1, which changes how printf behaves. This is consistent with the behavior when the argument overflows an int - printf_builtin: change check for overflow of [LONG_MIN..LONG_MAX] for %d/%i to be explicit instead of relying on integer overflow of p - printstr,printwidestr: print message if field width or precision overflow, since we don't call printf - printstr,printwidestr: set field width and precision on integer overflow the way we do in printf_builtin From a report and patch by Paul Eggert bashtypes.h - PRIdMAX: move redefinition here after including inttypes.h builtins/printf.def,examples/loadables/seq.c,examples/loadables/getconf.h - remove PRIdMAX redefinitions, since it's now fixed in bashtypes.h From a report and patch by Paul Eggert 3/15 ---- include/stdckdint.in.h,include/intprops-internal.h - new files, from gnulib configure.ac - stdckdint.h: create in the build directory if the system doesn't provide one by copying ${srcdir}/include/stdckdint.in.h; make sure we don't create a new one every time, changing the timestamp Makefile.in - CREATED_HEADERS: add stdckdint.h stringlib.c,externs.h - substring: now takes size_t arguments for START and END lib/sh/stringvec.c - strvec_create,strvec_mcreate,strvec_resize,strvec_mresize: use ckd_mul to avoid size_t overflow - strvec_len,strvec_flush: use size_t variable as array index From a report by Paul Eggert bashansi.h - stdbool.h: include if we have it, typedef bool as unsigned char if we don't and HAVE_C_BOOL isn't defined m4/c-bool.m4 - autoconf check for `bool' from gnulib configure.ac - include m4/c-bool.m4, call gl_C_BOOL; don't explicitly check for stdbool.h using AC_CHECK_HEADERS any more config.h.in - HAVE_C_BOOL: define 3/16 ---- builtins/printf.def - decodeprec: now decodeint, takes new arguments: a char ** that is the string to parse and update, an argument saying whether or not to print an error message on overflow, and an argument that should be the return value on overflow; use ckd_* macros to check overflow - printstr, printwidestr: use a different mechanism to check overflow; call decodeint to get precision and field width braces.c - include stdckdint.h for the ckd_* overflow checking macros - mkseq: the width argument is now size_t - mkseq: perform overflow detection using the ckd_* macros - mkseq: perform zero-padding directly instead of using asprintf; the sprintf family has trouble when width > INT_MAX - expand_seqterm: use size_t instead of int for length and width variables From a report and patch by Paul Eggert 3/18 ---- builtins/printf.def - printstr,bexpand: changed size/length arguments to be size_t; changed callers subst.c - string_var_assignment: if the variable is a local variable at the current scope, output a declare command even if there are no attributes or value. Still tagged as tentative 3/23 ---- test.c - posixtest: takes an argument, the number of test arguments to parse - two_arguments,three_arguments: use advance() instead of accessing and manipulating pos directly - posixtest,three_arguments,two_arguments: don't set pos = argc directly, rely on functions to advance pos appropriately Picked up from coreutils test - term: if we see a left paren, scan forward for a closing right paren and use posixtest() if there are 1-4 arguments between the parens to avoid ambiguous behavior for expressions like true -a \( ! -a \), which, though it's horribly ambiguous, a user would expect to return false instead of an error message Picked up from coreutils-9.2 test 3/25 ---- bashansi.c - HAVE_C_BOOL: prefer if defined, fall back to stdbool.h if bool isn't a compiler builtin braces.c - mkseq: simplify break condition builtins/printf.def - decodeint: let the caller decide what to do with an overflow return, don't return -1 after calling report_erange(). This will honor a precision even if the field width overflows - printf_builtin: check for overflow when adjusting precision for string length with %Q - printstr,printwidestr: now that the code uses decodeint, we don't need to check the precision against INT_MAX bashansi.h - include stddef.h if we have it, otherwise define replacements for NULL and offsetof() builtins/help.def,lib/sh/fmtulong.c,lib/sh/getcwd.c,lib/sh/makepath.c, lib/sh/snprintf.c,lib/sh/strtod.c,lib/sh/strtol.c,unwind_prot.c, lib/readline/shell.c - include bashansi.h instead of stddef.h, or remove a stddef.h include if bashansi.h is already included - remove extra definitions that appear in stddef.h or bashansi.h lib/sh/reallocarray.c,externs.h - implementation of reallocarray(), originally from OpenBSD, from Paul Eggert configure.ac - use AC_REPLACE_FUNCS for reallocarray xmalloc.c,xmalloc.h - xreallocarray: implementation of reallocarray that uses xrealloc - sh_xreallocarray: implementation of reallocarray that uses sh_xrealloc and has file and line information for error messages lib/sh/stringvec.c - strvec_create, strvec_resize: rewrite in terms of xreallocarray - strvec_mcreate, strvec_mresize: rewrite in terms of reallocarray - strvec_remove, strvec_search, strvec_copy, strvec_from_word_list, strvec_to_word_list: use size_t instead of int where appropriate lib/sh/stringvec.c,externs.h - strvec_search: return a value of type ptrdiff_t, since size_t is unsigned and we want to return -1 if not found All from a report and patch by Paul Eggert test.c - term: change paren scan algorithm to count open and close parens so we handle nested subexpressions. Still not compatible in the case of something like test true -a \( ! -a \) \) 3/27 ---- parse.y - flush_parser_state: new function, deallocates the members of a parser state struct that are dynamically allocated - parse_comsub: call flush_parser_state in places where we will return an error or jump to the top level Fixes memory leaks on error reported by bug_reports00@protonmail.com Makefile.in - CREATED_MACOS: list of dSYM directories created by building and testing - distclean,maintainer-clean: remove $(CREATED_HEADERS) and $(CREATED_MACOS) - maintainer-clean: remove ctags/etags files 3/28 ---- bashline.c,bashline.h - uw_restore_parser_state: moved to parse.y and declaration to shell.h shell.c,shell.h - parsing_command: new flag, set to 1 when calling yyparse; saved and restored by save_parser_state/restore_parser_state parse.y,shell.h - parsing_command: new element of sh_parser_state_t sig.c - throw_to_top_level: reset executing and parsing_command to 0 parse.y - parse_comsub: set parsing_command to 1 before calling yyparse() eval.c - parse_command: set parsing_command to 1 before calling yyparse(), restore old value when yyparse returns subst.c - function_substitute: if we are parsing a command, save the parser state with save_parser_state and add an unwind-protect to restore it From a report from Grisha Levit - function_substitute: only save and restore the pipestatus array if we are not parsing a command, since save_parser_state saves it parse.y,externs.h - decode_prompt_string: now takes an additional argument to determine whether this is expanding PS[0124] or the @P transformation - decode_prompt_string: keep track of the (last) real prompt string being decoded so a @P expansion embedded in the prompt returns the same result as the prompt string itself for the \! and \# expansions From a patch from Grisha Levit parse.y,print_cmd.c,eval.c,subst.c - decode_prompt_string: changed all callers 3/29 ---- doc/bash.1,lib/readline/doc/readline.3,lib/readline/doc/history.3 - fix some font-setting directives - fix some word breaks for font-changing macros lib/readline/doc/history.3 - fix the macros for function signatures for AT&T troff limitations - fix some paragraph spacing and other layout/spacing/hyphenation issues Fixes from G. Branden Robinson 4/1 --- execute_cmd.c - shell_execve: minor changes to code that checks the #! interpreter name execute_cmd.c - execute_coproc: if MULTIPLE_COPROCS is not defined, close any old file descriptors corresponding to sh_coproc. The user can still copy them from the coproc array if desired; this prevents the shell from hanging if a second coproc inherits file descriptors from a previous one From a report by Carl Edquist jobs.c,nojobs.c,jobs.h - process_exit_status: no longer static, public function with an extern declaration 4/2 --- execute_cmd.c,execute_cmd.h - coproc_pidchk: now takes a full WAIT status word as a second argument instead of just the exit status; arranged the prototype to only be included if the source file includes posixwait.h for WAIT jobs.c - waitchld: call coproc_pidchk with a WAIT second argument (like nojobs.c does) command.h - COPROC_STOPPED,COPROC_FOREGROUND: new flags values for coproc execute_cmd.c - coproc_setstatus: now takes a full WAIT status word as the second argument, passed from coproc_pidchk - coproc_setstatus: use process_exit_status to translate a WAIT into an int status, doing proper termsig+128 handling - coproc_getstate: take a WAIT and return the appropriate COPROC_ status flag (DEAD, STOPPED, etc.) - coproc_setstatus: use coproc_getstate to set cp->c_flags 4/4 --- configure.ac - changed release status to bash-5.3-alpha - AC_PROG_GCC_TRADITIONAL: removed, since it no longer is supported in autoconf-2.72 (which MacPorts upgraded me to) test.c - term: made change of 3/25 dependent on the compatibility level support/config.guess,support/config.sub,support/config.rpath - new versions, imported from gnulib 4/5 --- doc/bashref.texi - update compatibility, posix mode, and bourne shell sections for bash-5.3-alpha release version.c - update copyright date to 2024 [bash-5.3-alpha frozen] 4/6 --- builtins/cd.def - cd_builtin: a null pathname argument is now an error; POSIX interp 1047 variables.c - makunbound: revert change from 7/10/2023 about preserving the export attribute when unsetting a local variable in light of POSIX interp 1806 4/8 --- bashline.c - command_word_completion_function: we don't need to perform an extra comparison against what rl_filename_completion_function returns if we are searching $PATH for executable completions - command_word_completion_function: if a directory name from $PATH contains characters that need to be quoted, quote them and set rl_completion_found_quote to force rl_filename_completion_function to dequote the entire pathname - command_word_completion_function: since the directory name from $PATH is not quoted, use the dequoted hint to construct the full pathname to pass to rl_filename_completion_function (possibly after quoting it) From https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=306043 4/10 ---- redir.h - RX_EXPANDED: flag set when translating redirects to their final form; indicates the redirection should not be expanded again redir.c - do_redirection_internal: when translating a redirect into r_err_and_out, set the RX_EXPANDED flag so the filename doesn't get expanded again. Fixes bug reported by squeaky 4/12 ---- parse.y - expecting_in_command: new variable, set to FOR or CASE depending on which command is expecting the IN token to follow it - {save,restore}_parser_state: save and restore expecting_in_command in `incmd' member - special_case_tokens: for POSIX rule 6, check expecting_in_token instead of last_read_token/token_before_that for case/for command and whether we should return IN and maybe set PST_CASEPAT - special_case_tokens: reset expecting_in_command before returning DO - read_token_word: set expecting_in_command if last_read_token was CASE or FOR and we are returning a WORD - reset_parser: reset expecting_in_command - parse_comsub,parse_compound_assignment: reset expecting_in_command Fixes bug reported by nbowler@draconx.ca parser.h - incmd: new member of parser_state, saves and restores the value of expecting_in_command - PST_FORCMD: new parser state, used to say if we're parsing a `for' command and waiting for an IN token (not used yet) 4/22 ---- config-top.h - MULTIPLE_COPROCS: now enabled by default 4/23 ---- doc/bash.1,doc/bashref.texi - add text clarifying the difference between PIPESTATUS and $? expr.c,redir.c,jobs.c,parse.y,arrayfunc.c,test.c,subst.c builtins/declare.def,builtins/wait.def,builtins/printf.def - went through and removed or modified TAGs for bash-5.3 lib/readline/bind.c - rl_print_keybinding: handle case where there are no key sequences bound to the command name; don't just print out every unbound key builtins/bind.def - bind_builtin: if the shell compatibility level is > 52, interpret additional non-option arguments supplied with -P and -p as bindable command names and print out any bindings for those names 4/24 ---- Makefile.in,doc/Makefile.in,builtins/Makefile.in lib/sh/Makefile.in,lib/glob/Makefile.in,lib/tilde/Makefile.in lib/readline/Makefile.in - MFLAGS -> BASH_MAKEFLAGS, rely on make passing MAKEFLAGS in the environment From a report by Cedric Blancher buildconf.h - new build include file, construct from buildconf.h.in; contains a subset of config.h for use by the build tools mksyntax.c - now assumes a C90 compilation environment; includes buildconf.h for any build-time configuration variables Makefile.in - buildconf.h: add to CREATED_CONFIGURE - mksyntax: now depends on buildconf.h instead of config.h 4/25 ---- support/bashversion.c,version.c - now assumes a C90 compilation environment; includes buildconf.h for any build-time configuration variables support/printenv.c,support/recho.c,support/xcase.c,support/zecho.c - now assumes a C90 compilation environment and POSIX.1-1990 execution environment 4/26 ---- redir.h - RX_INTERNAL, RX_USER (unused), RX_SAVCLEXEC, RX_SAVEFD, RX_EXPANDED: new values starting at 0x80, now intended for rflags member of a struct redirect; don't want them colliding with O_XX flags for open and fcntl redir.c - rflags: save redirect->rflags and pass it to make_redirection if we're translating a redirection into a new one - new_redirect: pass rflags to all calls to make_redirection now that it holds more than REDIR_VARASSIGN - RX_INTERNAL, RX_USER (unused), RX_SAVCLEXEC, RX_SAVEFD, RX_EXPANDED: check and assign these in the rflags member; they're only used here - do_redirection_internal: assign new_redirect->rflags back to redirect->rflags since it can possibly be modified with RX_EXPANDED - do_redirection_internal: take care not to leave REDIR_VARASSIGN in new_redirect->rflags, since it will cause redirector to be freed, leading to accessing freed memory From a report by Clark Wang and a hint from and Grisha Levit 4/28 ---- lib/sh/anonfile.c - anonopen: call memfd_create with MFD_NOEXEC_SEAL From a report by Kerin Millar 4/29 ---- shell.c,shell.h - su_shell: now global variables.c - get_bash_name: use the user's login shell for $BASH if the shell is a login shell and the name is "-su". From https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1069978 subst.c - uw_unbind_localvar: unwind-protect to unbind a local variable at the current function context - function_substitute: make sure we unbind the local REPLY we created at the current (fake) context From a report by Koichi Murase 4/30 ---- subst.c - expand_arrayref: fix crash from freeing memory that's only allocated and initialized if ARRAY_VARS is defined Report and patch from Henrik Lindström jobs.h,jobs.c - JLIST_POSIX: new format argument for pretty_print_job; implements POSIX requirement that jobs display the status of all background jobs and all jobs the user hasn't been notified about (unused yet) - JLIST_BGONLY: new format argument for pretty_print_job; restricts output to background jobs only (unused yet) jobs.c,jobs.h,execute_cmd.c,subst.c - freeze_jobs_list: now takes an int argument with the new value of jobs_list_frozen; prep for having different values with different meanings; changed callers jobs.c - wait_for_any_job: return right away if jobs_list_frozen > 0; allow job status changes (e.g., J_NOTIFIED) if jobs_list_frozen < 0; use this instead of testing executing_funsub directly - notify_and_cleanup: allow notification and status change if jobs_list_frozen < 0; don't delete any dead jobs - should_notify: takes a job index and returns 1 if the shell would notify the user about it, given the current job state - pretty_print_job: if the jobs list is frozen, only print status about jobs for which the shell would notify users (by calling should_notify()) subst.c - function_substitute: freeze the jobs list with value -1 so jobs can change status and possibly inhibit printing by `jobs' 5/2 --- jobs.c - notify_and_cleanup: allow job notifications if an interactive shell is running a trap (interactive == 0 && interactive_shell && running_trap) Fixes report by Koichi Murase on 11/14/2022 - print_pipeline: don't print an extra space before the pipeline; push that into pretty_print_job; print the space after the pid if we print one jobs.h - LONGEST_SIGNAL_DESC: update to 27 (macos SIGPROF). This changes the test output 5/3 --- builtins/common.h - SEVAL_NOTIFY: new flag for parse_and_execute; means we want job notifications even though we're not interactive at this time and would not satisfy the conditions builtins/evalstring.c - parse_prologue: if SEVAL_NOTIFY is supplied, unwind-protect want_job_notifications and set it to 1 jobs.c,jobs.h.nojobs.c - want_job_notifications: new global variable, initialized to 0 jobs.c - notify_and_cleanup: notify about dead jobs if want_job_notifications is non-zero parse.y - execute_variable_command: call parse_and_execute with the SEVAL_NOTIFY flag eval.c - reader_loop: call notify_and_cleanup before executing $PROMPT_COMMAND if the shell is interactive and prompting trap.c - run_pending_traps,_run_trap_internal,run_exit_trap: add SEVAL_NOTIFY to the flags for parse_and_execute if the shell is interactive bashline.c - bash_execute_unix_command: add SEVAL_NOTIFY to the flags for parse_and_execute if the shell is interactive Rest of fix for report by Koichi Murase on 11/14/2022 5/6 --- execute_cmd.c,shell.c,builtins/evalfile.c,unwind_prot.c - change some translated error messages to make the text more uniform and reduce the number of gettext() calls redir.c,parse.y,bashhist.c.locale.c builtins/common.c,builtins/bind.def,builtins/exec.def,builtins/fc.def builtins/help.def,builtins/mapfile.def,builtins/printf.def,builtins/read.def builtins/ulimit.def - change some error messages to avoid having the format be the return value from gettext() to avoid clang complaints; simplified translatable strings and made them more uniform by removing printf formatting directives execute_cmd.c - notfound_str: cache the translated version of "command not found" so we can call gettext early on and work around a macOS misfeature From a MacPorts ticket https://trac.macports.org/ticket/68638 forwarded by Tom - execute_disk_command: use notfound_str instead of calling gettext every time a command isn't found execute_cmd.c,execute_cmd.h - init_notfound_str: call gettext to initialize and translate notfound_str - execute_disk_command: call init_notfound_str before forking if it hasn't been initialized locale.c - set_default_locale_vars: after we set the appropriate values for the locale shell variables, call init_notfound_str to force a call to gettext() jobs.c - wait_for: don't try to set the tty state if we're running a trap, even if a process terminates due to a signal 5/7 --- builtins/enable.def - enable_builtin: don't try to turn enable f into enable -f f if the shell is restricted 5/11 ---- doc/bash.1,lib/readline/doc/readline.3,lib/readline/doc/rltech.texi - document some of the bindable commands bound to Home, End, Insert, Page Up, and Page Down, if those keys exist From a report by Xose Vazquez Perez in 3/2024 configure - make sure to note in CHANGES that configure now supports --enable-year2038 for large time_t 5/14 ---- pathexp.c - glob_quote_char: new function, identify characters that should have backslashes preserved when quoting a glob pattern. Currently has all the glob_char_p characters plus `%' and `#' for pattern substitution and removal. Fixes issue reported by Andreas Schwab 5/16 ---- pcomplete.c - programmable_completions: change to allow a newly-loaded compspec (after a completion function returns 124) to be used in more cases From a patch from Grisha Levit mksyntax.c,version.c,support/bashversion.c - include config.h if CROSS_COMPILING is not defined, assuming that the build host and target host are the same buildconf.h.in - if cross-compiling, define _POSIX_C_SOURCE and _XOPEN_SOURCE to get the C90/1003.1-1992 compilation environment we assume From a report from Grisha Levit 5/17 ---- pathexp.c - globsort_sizecmp,globsort_blockscmp: replace subtraction with a generic transitive comparison function to avoid problems with integer underflow From a report from Grisha Levit 5/20 ---- lib/readline/kill.c - _rl_bracketed_text: make sure buf is null-terminated even if rl_read_key() returns an error From a report from Grisha Levit builtins/read.def - read_builtin: if -u and -e are both supplied, dup the file descriptor supplied as an argument to -u and use it in the new FILE * to pass to readline as rl_instream. Works around an android problem with stdio and application-managed file descriptors Report and patch from Grisha Levit 5/21 ---- lib/sh/spell.c - mindist: don't check best unless we set it Report and patch from Grisha Levit jobs.c - make_child: if FORK_NOJOB is in the flags argument, don't call setpgid to set the child's process group in either the parent or child - alloc_process,dispose_process: allocate and deallocate a PROCESS; changed callers subst.c - command_substitute: call cleanup_the_pipeline after waiting for the command substitution process, since we allocated it 5/22 ---- lib/readline/text.c - rl_execute_named_command: fix a leak if the command name is null or the bound function doesn't return (rl_abort) Report and patch from Grisha Levit bashline.c - command_word_completion_function: free directory_part if it's left over from a previous completion - command_subst_completion_function: free contents of match list left over from previous completion - bash_spell_correct_shellword: free text if it's "" before returning - build_history_completion_array: only call qsort if there's actually something in the array to sort Report and patches from Grisha Levit - bash_spell_correct_shellword: fix bug where we would correct the previous word if we start on the first character of a word - bash_spell_correct_shellword: make negative argument counts work backwards, correcting words before point Report from Grisha Levit 5/23 ---- lib/readline/text.c - rl_change_case: if mbrtowc returns -1 or -2, jump to changing case for a single character, since _rl_find_next_mbchar_internal() will treat an invalid multibyte character as a sequence of bytes Report from Grisha Levit 5/25 ---- execute_cmd.c - shell_execve: fix typo in code that chops \r off the end of the #! interpreter Report and fix from Collin Funk 5/28 ---- execute_cmd.c - time_command: only restore command->flags if we haven't longjmped Report from Michael Maurer and Grisha Levit subst.c - skip_to_delim: if we're not skipping over command substitutions lexically, call extract_command_subst instead of using the old extract_delimited_string Fixes bug reported by Oguz execute_cmd.c - execute_arith_for_command: handle the extremely unlikely case that the step or test expressions execute `break' or `continue' in a nofork command substitution Report from Oguz 5/29 ---- lib/readline/complete.c - compute_lcd_of_matches: if we have multiple matches that compare equally when using case-insensitive completion, but are different lengths due to the presence of multibyte characters, use the shorter match as the common prefix to avoid overflow Report from Grisha Levit 5/30 ---- lib/readline/text.c - _rl_readstr_init: don't call rl_maybe_replace_line since we're not actually moving off this history line Report from Grisha Levit builtins/read.def - read_builtin: free ifs_chars in more return code paths Report and patch from Grisha Levit 5/31 ---- quit.h - ZRESET: new macro, calls zreset() if interrupt_state is non-zero lib/sh/zread.c - zread: if read returns -1/EINTR, and we're executing a builtin, call zreset just in case builtins/read.def - read_builtin: if read returns > 0 (partial read) but interrupt_state is non-zero, we're going to call throw_to_top_level, so call ZRESET before that happens Report from Oguz - read_builtin: if zread/zreadc/zreadintr/zreadcintr return -1/EINTR, make sure we call ZRESET in case zread did not variables.h - ASSIGN_DISALLOWED: macro that encapsulates when an assignment to SHELL_VAR *v with flags f will be disallowed and fail arrayfunc.c,execute_cmd.c,expr.c,redir.c,subst.c,variables.c builtins/delare.def,builtins/getopts.def,builtins/printf.def,builtins/read.def - use ASSIGN_DISALLOWED where appropriate arrayfunc.c - assign_array_element_internal: if the assignment failed, as tested by ASSIGN_DISALLOWED, free the assoc array key we allocated Report from Grisha Levit 6/3 --- variables.c - bind_function_def: don't copy the command tree if we're overwriting a function definition; after function binding, the function_def struct never uses the saved command anyway From a report and patch by Koichi Murase parse.y - make sure to free WORD in coproc WORD COMMAND after creating the coproc struct Report and patch from Grisha Levit builtins/exec.def - exec_builtin: we don't need to set args to NULL; if shell_execve reallocates args to add more items at the front, it has committed to longjmp back to the top level Report and patch from Grisha Levit subst.c - expand_word_internal: dispose of the list used for $@ if we're not splitting and save the space-separated list back to tword->word Report and patch from Grisha Levit parse.y - read_token: if we try to parse a conditional command and get a syntax error, make sure to dispose of the partial command we created Report and patch from Grisha Levit - read_token: if we try to parse a conditional command and get a syntax error because we read an unexpected WORD, dispose of the WORD_DESC before returning -1 Report and patch from Grisha Levit - cond_term: if we read a WORD where we expect something else, dispose of the WORD_DESC before returning COND_ERROR Report and patch from Grisha Levit 6/10 ---- parse.y - cond_term: if we read a WORD when expecting a close paren, dispose of the WORD_DESC before returning COND_ERROR - error_token_from_token: use the TOK argument instead of current_token if we can't find the token as a reserved word or symbol 6/12 ---- builtins/source.def - source_builtin: add -p PATH option, searches PATH argument instead of $PATH; overrides sourcepath; does not search $PWD if path search fails doc/bash.1,doc/bashref.texi - source: document -p pathexp.c - if the size, blocks, or any of the time values compare equal, use the name for a secondary sort key lib/sh/winsize.c - tcgetwinsize: provide definition for missing newly-standard function - get_new_window_size: use tcgetwinsize subst.c - do_compound_assignment: if ASS_CHKLOCAL is in the flags, check for variables at previous local scopes instead of just the current local scope and the global scope Part of fix for inconsistency between arrays and scalar variables reported by Will Allan builtins/declare.def - declare_find_variable: if `declare -G' is used, check for variables at previous local scopes instead of just the current local scope and the global scope Rest of fix for inconsistency between arrays and scalar variables reported by Will Allan pathexp.h - SORT_NUMERIC: new "numeric" sort specifier pathexp.c - globsort_numericcmp: new sort function for the "numeric" sort specifier: all-digit names are sorted as numbers; names containing any non-digits sort after all-digit names and are sorted lexicographically. Inspired by a discussion with Robert Elz 6/21 ---- lib/readline/text.c - rl_change_case: make mlen size_t to assign the return value from wcrtomb; change the tests of < 0 to use MB_INVALIDCH. Report from Siteshwar Vashisht 6/26 ---- trap.c - run_debug_trap: give the terminal back to pipeline_pgrp only if job_control is enabled; that's the only way the trap command would have given it to a different process group Fixes bug reported by Mark March 6/27 ---- configure.ac - remove support for cached files for cross-compiling; it's too hard to keep them up-to-date without access to the platforms 6/28 ---- lib/readline/util.c - _rl_audit_tty: make sure to close fd before returning on error lib/sh/casemod.c - sh_modcase: use size_t variable as return value for wcrtomb; check for invalid wide char and copy all characters in the multibyte character as-is if it is lib/sh/zwrite.c - zwrite: use size_t and ssize_t variables builtins/evalfile.c - _evalfile: use size_t variable for return value from strlen() lib/readline/funmap.c - rl_funmap_names: don't bother calling qsort if result doesn't have any entries All from a report by Siteshwar Vashisht 6/29 ---- lib/readline/examples/histexamp.c,lib/readline/examples/manexamp.c, lib/readline/examples/rl-callbacktest.c,lib/readline/examples/rl-callbacktest2.c, lib/readline/examples/excallback.c,lib/readline/examples/rlptytest.c - small changes to fix warnings From a report by Hiroo Hayashi 7/1 --- execute_cmd.c: - ADJUST_LINE_NUMBER: encapsulate code that adjusts line_number in shell functions executed in interactive shells into a macro 7/2 --- doc/bash.1,doc/bashref.texi - changes to the --rcfile description to avoid the use of `force' and stay away from the word `replace' lib/readline/misc.c - _rl_maybe_replace_line: old rl_maybe_replace_line, now takes an argument saying whether or not to clear rl_undo_list - rl_maybe_replace_line: now calls _rl_maybe_replace_line with a 0 argument for backwards compatibility lib/readline/rlprivate.h - _rl_maybe_replace_line: extern declaration lib/readline/search.c - _rl_nsearch_init,rl_history_search_internal: call _rl_maybe_replace_line with a 1 argument so these searches can manage rl_undo_list and restore it appropriately - rl_history_search_internal: don't set the history offset back to oldpos if the history search is successful; we want the history offset to be set to the position of the last matching line, like isearch 7/5 --- jobs.c - procsub_reap: new function, blocks and unblocks SIGCHLD around call to procsub_prune - cleanup_dead_jobs: call procsub_reap even if there are no jobs in the jobs list or the jobs list is frozen Inspired by report from Zachary Santer 7/6 --- subst.c,subst.h,execute_cmd.c,jobs.c - reap_procsubs: rename to delete_procsubs 7/7 --- aclocal.m4 - BASH_CHECK_LIB_TERMCAP: check for ncursesw before ncurses, add -lncursesw if found, to avoid pkgconfig issues for readline (since this is a shared file) Report and fix from Rosen Penev 7/8 --- builtins/evalstring.c - should_optimize_fork: don't optimize the fork if we're executing in a shell function and a return trap is set - optimize_shell_function: mark a simple command at the end of a shell function as a candidate for optimization, rather than setting CMD_NO_FORK before the function executes -- it could set a return trap Report from Jens Schmidt - parse_and_execute: don't let process substitutions or other calls to parse_and_execute do - parse_string: if flags & SEVAL_ONECMD, stop after parsing a single command, even if it doesn't consume the entire string parse.y - parse_string_to_command: call parse_string with SEVAL_ONECMD, since we only want to parse a single command and make sure it consumes the entire string, not consume the string and return the last command parsed 7/9 --- bashline.c - _ignore_completion_names: restrict force_fignore to act only on completions that should be affected by FIGNORE; in other cases act as if it were set to 1 From a report and patch by Koichi Murase 7/15 ---- builtins/ulimit.def - add help text saying that some of the units change in posix mode 7/17 ---- subst.c - remove some unused functions jobs.c,jobs.h - procsub_search,procsub_delete: now take an extra parameter saying whether or not to block SIGCHLD while they run; avoids overhead when called when SIGCHLD is already blocked; changed callers and extern declaration - struct process now has a new FLAGS member; possible values defined in jobs.h; initialized in alloc_process - procsub_setflag,procsub_setflag: set one of the PROC_XXX flags on one or all of the processes in the procsub list - new value for the flags argument to make_child: FORK_PROCSUB - add_process: now returns the PROCESS * it creates so the caller can set flags or otherwise modify it - get_job_by_pid: return NO_JOB for negative pids immediately subst.c - process_substitute,command_substitute: call make_child with the appropriate FORK_ flag jobs.c - make_child: set PROC_ flags in the PROCESS * we create based on the FORK_ flags the caller passes - wait_for: only call set_procsub_status if the process has terminated - wait_for_any_job: check for any terminated procsubs as well as any terminated background jobs builtins/wait.def - check_bgpids: rename to check_nonjobs - wait_builtin: if the -n option is supplied and pid arguments are supplied, check the bgpids list and procsubs for terminated processes - check_nonjobs: if one of the pids in LIST corresponds to a terminated procsub, return its status and move the procsub to the bgpids list - set_waitlist: if we have a non-existent or invalid job, and a pid argument that's >= 0, look for a procsub with that pid and set PROC_WAITING if we have one - unset_waitlist: unset PROC_WAITING in all procsubs 7/18 ---- jobs.c - procsub_prune: set last_procsub_child to NULL if we are cleaning up that process because it's terminated, since that will invalidate the pointer 7/22 ---- lib/malloc/malloc.c - RIGHT_BUCKET: don't check binsizes[nu-1] unless nu >= 1; clamp at 0 otherwise - internal_realloc: don't check bucket at nunits-1 unless nunits >= 1 Report and fix from Collin Funk 7/25 ---- variables.c - push_source: new function, pushes a filename to BASH_SOURCE; changed callers (shell.c, execute_cmd.c, builtins/evalfile.c) - bash_source_fullpath: new variable, if non-zero, push_source runs the filename through sh_realpath before pushing it to BASH_SOURCE Feature requested by several, including konsolebox builtins/shopt.def - bash_source_fullpath: new option, controls bash_source_fullpath and whether or not BASH_SOURCE contains full pathnames doc/bash.1,doc/bashref.texi - bash_source_fullpath: document new shell option 7/26 ---- jobs.c,jobs.h,subst.c - last_procsub_pid: new variable, set to the pid of last_procsub_child, so it will survive the PROCESS * being deleted if we need it. Not used yet 7/29 ---- doc/bash.1,doc/bashref.texi - wait: update description to clarify wait -n and unify language Inspired by report from Zachary Santer - set: note that the -e behavior for pipelines is affected by the state of the pipefail option From a report by Martin D Kealey variables.c,hashcmd.c,builtins/hash.def,builtins/history.def,builtins/source.def - use ABSPATH/RELPATH/absolute_program instead of testing for slash explicitly 7/30 ---- lib/readline/histfile.c - history_do_write: return 0 immediately if nelements < 0 Report from ZeroIntensity Dev 8/3 --- lib/sh/zread.c,externs.h - zungetc: new function, pushes a character back into a single char of storage so it can be used with the zread* functions that call read(2) directly and don't use lbuf. If we try to push back more than one char, we use lbuf, assuming that the caller will accommodate it builtins/read.def - read_mbchar: if we read a partial multibyte character (-2) and then a delimiter that makes it an invalid multibyte character (-1), use zungetc to push the delimiter back, adjust mbchar, and return what we read so far to be added as single bytes From a report by Kerin Millar 8/7 --- builtins/evalstring.c - can_optimize_connection: only check bash_input.location.string if bash_input.type == st_string, in case we want to use this in other contexts. If we do, the caller *must* make sure it's the last command in the calling context execute_cmd.c - execute_subshell_builtin_or_function: if this is an async function call, try to optimize away the fork before the last simple command in the function body by calling optimize_shell_function - execute_command_internal: if we have a simple command for which CMD_TRY_OPTIMIZING is set in a SUBSHELL_ASYNC subshell environment, try to set CMD_NO_FORK by checking via should_optimize_fork 8/9 --- parse.y - yy_readline_get: before calling readline, call rl_clear_signals to reset the signal handling disposition to a known state. Most of the time it's a no-op lib/readline/signals.c - _rl_handle_signal: make the call to sigprocmask(SIG_UNBLOCK,... more symmetric with the corresponding call to SIG_BLOCK trap.c - run_pending_traps: before calling run_interrupt_trap, restore running_trap and trap_return_context, since _run_trap_internal will save and restore them, too builtins/read.def - read_builtin: if -e is supplied, make sure to install unwind-protects to reset the attempted completion function and the event hook even if the timeout is not supplied. Report from Oguz 8/12 ---- lib/readline/histsearch.c,lib/readline/histlib.h - history_search_internal: takes a new argument, LINEDIR, saying which direction to perform the substring/pattern search in the line, decoupling it from the direction through the history list; changed all callers (history_search, history_search_prefix, _hs_history_search, _hs_history_patsearch) - _hs_history_search,_hs_history_patsearch: takes the same new LINEDIR argument, just passes it to history_search_internal; changed all callers (noninc_search_from_pos) lib/readline/search.c - noninc_search_from_pos: changed calls to _hs_history_search and _hs_history_patsearch lib/readline/histexpand.c - get_history_event: instead of using function pointers to a history search function, set the right flags and just call _hs_history_search with those flags using -1 for list direction and line direction 8/13 ---- builtins/printf.def - printf_builtin: fix params to call to mklong for long double formats if STRTOLD_BROKEN is defined lib/sh/winsize.c - rearrange includes to make sure is included if HAVE_TCGETWINSIZE is defined Report from Grisha Levit 8/14 ---- lib/readline/rlmbutil.h - _rl_is_basic: adapted from gnulib is_basic; tells whether a character is a single byte because it's guaranteed by ISO C and POSIX lib/readline/mbutil.c - _rl_mbstrlen: counts the number of possibly multibyte characters in a string depending on the current locale; uses _rl_utf8_mbstrlen or _rl_gen_mbstrlen as appropriate (or strlen for a single-byte locale) lib/readline/readline.c - _rl_internal_char_cleanup: if we are using multibyte characters, call _rl_mbstrlen to determine whether or not we've read as many characters than rl_num_chars_to_read says we should Report from Grisha Levit builtins/read.def - delim,delim_char: once again an int so we can more easily distinguish -1 and other unsigned char values >= 0x7f; make sure delim gets an unsigned char if passed with -d - edit_line: now takes an int parameter for the delim char; only add the delim to the end of the string if it's non-negative - set_eol_delim,unset_eol_delim: now explicitly check whether or not the delimiter character is non-negativesince it's no longer unsigned - read_mbchar: make sure to cast the char read to unsigned before comparing it against the delim Report and patch from Grisha Levit 8/15 ---- include/shmbchar.h,lib/sh/shmbchar.c - is_basic: update test for IS_BASIC_ASCII to include $, @, and `, which ISO C guarantees to be single-byte - is_basic_table: update bitmap to include new characters 8/21 ---- shell.c - main: make sure to set read_but_dont_execute (-n) to 0 in an interactive shell in case it was set at invocation, but not in cases where it was forced with -i Report from Milana <94888u@riseup.net> 8/22 ---- lib/readline/bind.c - force-meta-prefix: new bindable variable, used to determine whether to convert a meta character (>= 0x80) to an escape-prefixed key sequence (using ESC as the meta prefix). If set, do the conversion unconditionally; if unset, use the value of convert-meta as in previous bash/readline versions Suggestion from Reuben Thomas - _rl_function_of_keyseq_internal: if we encounter a character/byte with the eighth bit set, use force-meta-prefix to decide whether to translate it to an escape-prefixed key sequence - rl_bind_key: if we have a meta character, assume it was generated by Meta- and honor the setting of force-meta-prefix - rl_translate_keyseq: if we have \M-, use force-meta-prefix to decide whether or not to translate it to an escape-prefixed key sequence lib/readline/doc/rluser.texi,lib/readline/doc/readline.3,doc/bash.1 - force-meta-prefix: document new bindable variable and defaults - convert-meta, enable-meta-key, input-meta, output-meta: update description with better explanation of Meta key and meta characters, and locale character encoding - Key Bindings: expand description of \M- and Meta- modifiers - Readline Introduction: expand description of the Meta key and the various things it can do, reference force-meta-prefix and enable-meta-key 8/26 ---- doc/bashref.texi - case: add process substitution to the list of expansions for the word. Reported by shynur execute_cmd.c - execute_null_command: if an empty simple command (no words, just redirections) requires a fork, reset top_level and exit if we longjmp there. This can happen with an expansion error while we evaluate the redirection. Report from youheng.lue@gmail.com - execute_connection: in default mode, bash performs jobs notifications in an interactive shell between commands separated by ';' or '\n'. It shouldn't do this in posix mode, since posix now specifies when notifications can take place examples/loadables/{realpath,csv,dsv,cut,stat,kv} - changes to make them build when arrays are not available Report from Dennis Clarke jobs.c - notify_and_cleanup: make interactive shells notifying during sourced scripts dependent on the shell compatibility level and inactive in versions beyond bash-5.2 Inspired by report from Zachary Santer doc/bashref.texi - Bash posix mode: note the change with job notifications in command lists - Shell compatibility mode: note job notification changes for compat level > 52 parse.y - CHECK_FOR_RESERVED_WORD: call set_word_top here on the reserved word token about to be returned; don't worry about trying to set it after the fact Report from Collin Funk 8/28 ---- parse.y - CHECK_FOR_RESERVED_WORD: change word_top to keep track of the line number where a compound command begins; call set_word_top on the current token about to be returned - read_token_word: call set_word_top if special_case_tokens returns something; currently used for `{' - set_word_top: add `{' to the list of compound commands we track - grammar: every time we parse a group command or equivalent for the `for' and `select' compound commands, decrement word_top after we see the `}' - parse_dparen: set arith_lineno from word_lineno[word_top] since we now set it for FOR when we read it 8/29 ---- parse.y - read_token: don't turn on PST_SUBSHELL after reading a left paren if the last token was a WORD, since that's a function definition, not a subshell - read_token,parse_dparen: don't turn on PST_SUBSHELL if we run into a left paren while parsing a conditional command (PST_CONDCMD is set in parser_state) - set_word_top: add `(' to the tokens saved in word_lineno - grammar: decrement word_top in the subshell_command production - read_token,parse_dparen: call set_word_top where we add PST_SUBSHELL to parser_state 8/30 ---- parse.y - set_word_top: add COND_START to the list of tokens that set word_lineno[word_top] - grammar: decrement word_top after COND_END - rename: word_lineno -> compoundcmd_lineno, word_top -> compoundcmd_top, set_word_top -> set_compoundcmd_top lib/readline/signals.c - _rl_release_sigint: don't check for signals if we're already handling a received signal Report from Tycho Kirchner and Eduardo A. Bustamante López 8/31 ---- builtins/printf.def - printf_builtin: when parsing a time specification format, don't increment the format pointer to check for 'T' unless it points to a closing right paren, rejecting the format if it's not a `)' when the loop breaks Report from Andrey Kovalev 9/6 --- jobs.c,jobs.h - notify_and_cleanup: now takes an argument saying which job to notify about; if arg is -1 it notifies about all jobs as previously - maybe_print_job_notifications: convenience function to encapsulate policy about when to call notify_of_job_status in one place; called by notify_and_cleanup - notify_of_job_status,maybe_print_job_notifications: now take an int argument with the same semantics as notify_and_cleanup - wait_for_any_job: now call notify_of_job_status only on the job we just retrieved and will return - wait_for: don't call notify_and_cleanup if posixly_correct unless the shell is not currently interactive, since posix says we can notify in a non-interactive shell when a foreground job changes state jobs.c,parse.y,eval.c,execute_cmd.c,builtins/jobs.def,trap.c - notify_and_cleanup: changed all callers appropriately builtins/evalfile.c - _evalfile: renamed to evalfile_internal - evalfile_internal: increment and decrement (or unwind-protect) the value of want_job_notifications; increment it to a non-zero value if the shell is interactive and the compatibility level is <= 52 (see change to notify_and_cleanup of 8/26) jobs.c - maybe_print_job_notifications: remove clause testing sourcelevel; use want_job_notifications for this 9/11 ---- [prayers for the victims of 9/11] 9/12 ---- support/mkdep - new version from Martin Kealey; modified for inclusion Makefile.in - MKDEP_ variables to be able to control various aspects of `make depend' 9/13 ---- configure.ac - relstatus: update to `beta' 9/16 ---- builtins/jobs.def - jobs_builtin: when listing individial jobs, don't call notify_and_cleanup (which notifies about outstanding terminated jobs) if the shell is interactive and in posix mode 9/20 ---- jobs.c - wait_for_single_pid: if wait finds a status for the pid argument in the bgp list, remove it from the bgp list if the shell is in posix mode 9/23 ---- execute_cmd.c - execute_command_internal: if the COMMAND * passed in to execute has the CMD_INVERT_RETURN flag set, make sure we propagate that to simple commands, for commands, arith for commands, select commands, case commands, while/until commands, if commands, group commands, shell functions, pipelines, subshells, and connections so if set -e is enabled in the body (or test for if/while/until) its effect is ignored - execute_in_subshell: if the subshell's return status is being inverted, make sure to ignore set -e in the commands the subshell executes - execute_pipeline: if the pipeline's return status is being inverted, make sure to ignore the effects of set -e in all pipeline elements - execute_connection: if the list's return status is being inverted (tough to do because of precedence), make sure the list commands ignore the effect of set -e 9/24 ---- jobs.c - start_job: don't fetch the tty settings if we're being called from a readline key binding (bind -x) or programmable completion From a report by David Moberg 9/25 ---- jobs.c - wait_for_any_job: if the jobs table is empty and there are no eligible procsubs, and the shell is in posix mode, take a random pid from the bgpids table, delete it, and return its status (since we would be deleting that pid from bgpids anyway) 9/26 ---- lib/readline/histexpand.c - history_tokenize_word: fix off-by-one error (incrementing i one too many positions after the increment in the loop) preventing words like `$((expr))' from being tokenized correctly Report from Milana <94888u@riseup.net> config.h.in,configure.ac - new option: bash-source-fullpath-default: set it to make the bash_source_fullpath option enabled by default; sets BASH_SOURCE_FULLPATH_DEFAULT Based on a patch from konsolebox variables.c - bash_source_fullpath: initialize to BASH_SOURCE_FULLPATH_DEFAULT, set by configure doc/bashref.texi - --enable-bash-source-default: document 9/27 ---- doc/bash.1,doc/bashref.texi - update builtins documenation: make language and paragraph breaks more consistent between the document formats; update active language 10/1 ---- doc/bash.1,doc/bashref.texi,lib/readline/doc/rluser.texi,lib/readline/doc/hsuser.texi - update more documentation sections to make language consistent between formats, reduce passive voice 10/7 ---- lib/readline/doc/history.3 - update history expansion section to be consistent with hsuser.texi and bash.1 doc/bash.1,doc/bashref.texi,lib/readline/doc/rluser.texi,lib/readline/doc/readline.3 lib/readline/doc/hsuser.texi,lib/readline/doc/history.3 - next set of updates for consistency between formats, update language, fix formatting conventions 10/11 ----- lib/readline/doc/rltech.texi,lib/readline/doc/hstech.texi - next set of updates for consistency between formats, update language, fix formatting conventions 10/12 ----- Makefile.in,builtins/Makefile.in - update dependencies on readline, history include files builtins/shopt.def - update bash_source_fullpath reset to use new default configure.ac - update to use the right option name for bash-source-fullpath-default Report and patch from Grisha Levit braces.c - mkseq: fix loop condition to avoid integer overflow Report and patch from Grisha Levit builtins/help.def - show_desc: don't assume that long_doc[0] ends with a newline, just add one manually when we hit a newline or NULL Report and patch from Grisha Levit builtins/read.def - read_builtin: don't check the timeout after breaking out of the character read loop unless we didn't read anything (i == 0) and the last read return EOF or error - read_builtin: move check_read_timeout call to the end of the read loop, after we check for the delimiter and whether or not we've read the right number of characters, so we always return any valid data Report from Thomas Oettli 10/14 ----- jobs.c - wait_for_any_job: in posix mode, take any terminated pid from bgpids before waiting for a running job to terminate Inspired by report from Zachary Santer 10/14 ----- doc/bash.1,doc/bashref.texi - next set of updates for consistency between formats, update language, fix formatting conventions (job control section) doc/bash.1,doc/bashref.texi,lib/readline/doc/readline.3,lib/readline/doc/rluser.texi - minor wording updates and typo fixes from G. Branden Robinson 10/15 ----- doc/bash.1,doc/bashref.texi,lib/readline/doc/readline.3,lib/readline/doc/rluser.texi - wording/formatting updates and typo fixes from G. Branden Robinson doc/bash.1 - update .FN macro, use it consistently for filenames, like @file{} doc/bashref.texi - make sure to use @dots and @file consistently doc/bash.1,lib/readline/doc/readline.3 - make sure `readline' is typeset consistently 10/18 ----- sig.c - fix minix typo lib/readline/history.c - history_do_write: for portability, make sure that the offset argument to mmap is page-aligned, since some systems (linux, macos) require it. Report and patch from Grisha Levit lib/readline/util.c - rl_tilde_expand: rearrange code to do bounds checking before data validation when finding the end of the tilde-word to avoid touching uninitialized data Report and patch from Grisha Levit 10/20 ----- doc/bash.1,doc/bashref.texi - update word splitting section to add what IFS whitespace means and how word splitting uses it. Based on a bug-bash discussion 10/21 ----- lib/readline/colors.c doc/bash.1,lib/readline/doc/readline.3,lib/readline/doc/rluser.texi - RL_COLOR_PREFIX_EXTENSION: remove the leading `.' again; this was the result of a misunderstanding about how `dircolors' works Report from Daniël Gerbrand Haasbroek lib/readline/isearch.c - _rl_isearch_dispatch: when adding the character to the search string, insert it as a single byte in the C locale (or if we're not doing multibyte characters) Report and patch from Grisha Levit 10/23 ----- doc/bash.1,doc/bashref.texi - SIGNALS: update section to make it clear how job control affects SIGINT receipt and the behavior of the shell when it's waiting for a command that receives one. From a suggestion by Simone Robinson builtins/common.c - get_job_spec: warn about deprecated notation if the job spec doesn't have a leading `%'; code to return BAD_JOBSPEC tagged for bash-5.4 - sh_invalidjob: new convenience function to print error for invalid job specifications builtins/jobs.def, builtins/ kill.def, builtins/ wait.def, builtins/fg_bg.def - handle BAD_JOBSPEC return from get_job_spec; call sh_invalidjob. Nothing returns that yet. builtins/kill.def - kill_builtin: change to use common error message via sh_badpid() if we get an argument where the first character is not a digit or `%' builtins/jobs.def - jobs_builtin: check for INVALID_JOB return from get_job_spec to avoid call to get_job_by_jid 10/24 ----- error.c, error.h - err_invalidid: common error function for invalid identifiers; changed callers in execute_cmd.c, general.c general.c - valid_function_word: separated posix check against special builtin names (flags&4) and posix check for valid identifiers (flags&1); callers need to differentiate. This means that posix mode does not require function names to be valid identifiers execute_cmd.c - execute_intern_function: don't call valid_function_word with (flags&1) in posix mode unless POSIX_RESTRICT_FUNCNAME is defined; call with flags&4 to keep the check against special builtin names print_cmd.c - print_function_def,named_function_string: don't print functions with names that are invalid identifiers with a leading `function' config-top.h - POSIX_RESTRICT_FUNCNAME: placeholder, not defined by default doc/bashref.texi - Posix mode: remove item about function names being valid shell identifiers 10/25 ----- aclocal.m4 - simplify the checks for the various members of struct dirent Report and patch from Grisha Levit m4/strtoimax.m4 - simplify and remove redundant caching and cache check Report and patch from Grisha Levit builtins/common.c - get_exitstat: make the behavior of exit with no arguments running in a trap action the same as return: check for any trap (running_trap > 0) in the current execution_context. This is POSIX interp 1602, in POSIX.1-2024 stringlib.c - find_token_in_alist: if FLAGS is non-zero, do *not* allocate new memory for the token. The existing behavior is to allocate memory and return a copy of the token, and all callers pass 0 parse.y - compoundcmd_lineno: now an array of `struct tokeninfo', which has line number and current compound command information - change compoundcmd_lineno[compoundcmd_top] to compoundcmd_lineno[compoundcmd_top].lineno in all calls - set_compoundcmd_top: in addition to assigning line number info, assign T to compoundcmd_lineno[compoundcmd_top].token - report_syntax_error: if we get a syntax error while looking for shell_eof_token (comsubs), include the token we were looking for in the error message - report_syntax_error: if we are reporting an unexpected EOF error, and compoundcmd_top >= 0, look up the token for the compound command we're parsing, including `(', and change the error message to include it 10/26 ----- Makefile.in - PURIFY,GLOBC,GLOBO,ALLOC_FILES,LIBRARY_SOURCE: remove - ADDON_CFLAGS,ADDON_LDDFLAGS: initialize to empty - GRAM_H: remove from HSOURCES - profiling-tests: use Program instead of PROGRAM builtins/Makefile.in - DEBUG: initialize from configure - SYSTEM_FLAGS: remove, since it's not passed from topdir Makefile - ADDON_CFLAGS: initialize to empty lib/glob/Makefile.in - ADDON_CFLAGS: initialize to empty lib/readline/Makefile.in - ADDON_CFLAGS: initialize to empty - APP_CFLAGS: remove lib/sh/Makefile.in - DEBUG: initialize from configure - ADDON_CFLAGS: initialize to empty - LIBOBJDIR: initialize to empty; configure uses it in LIBOBJS lib/tilde/Makefile.in - ADDON_CFLAGS: initialize to empty - CCFLAGS: add ADDON_CFLAGS, replacing ASAN_CFLAGS support/Makefile.in - SYSTEM_FLAGS: remove, since it's not passed from topdir Makefile Makefiles report and patch from Grisha Levit version.c,shell.h - default_compatibility_level: int variable, initialized to DEFAULT_COMPAT_LEVEL like shell_compatibility_level variables.c - sv_shcompat: use default_compatibility_level instead of DEFAULT_COMPAT_LEVEL so variables.o doesn't have to depend on version.h builtins/shopt.def - set_compatibility_level: use default_compatibility_level instead of DEFAULT_COMPAT_LEVEL so shopt.o doesn't have to depend on version.h - don't include version.h any more configure.ac,Makefile.in,support/Makefile.in,doc/Makefile.in - remove some unused variables: BASE_CFLAGS_FOR_BUILD, DEBUGGER_DIR, INTLOBJS, TEXINDEX, TEX, MALLOC, SIGNAMES_SUPPORT From a report by Martin D Kealey 10/28 ----- builtins/exec.def - exec_builtin: don't try to print an error message in the cases where shell_execve prints one, in case we have execfail set Report from Emanuele Torre Makefile.in - LIBRARY_SOURCE: add back in to tags targets, add definition Suggestion from Mike Jonkmans 10/31 ----- lib/readline/search.c - make_history_line_current: change to use the same strategy to move to a history line as incremental search: one or more calls to rl_get_{next,previous}_history - dispose_saved_search_line: remove - noninc_dosearch: no longer need to call _rl_unsave_saved_search_line or dispose_saved_search_line; use new calling convention for make_history_line_current - _rl_nsearch_init: don't need to call _rl_maybe_replace_line, wait until we move to the history line - _rl_nsearch_dosearch: free the undo list unconditionally, since we are committed to using the contents of rl_line_buffer as the search string - _rl_nsearch_dosearch: if we abort because of no search string, call _rl_unsave_saved_search_line and let the caller clean up - _rl_nsearch_dosearch: if the search fails, reset rl_point to the saved point, since we will be restoring the old line - rl_history_search_internal: don't bother to save and restore the search line, since we don't use rl_line_buffer for anything - rl_history_search_internal: use new calling convention for make_history_line_current - all these changes unify the incremental and non-incremental search implementations lib/readline/readline.c - rl_initialize: reset rl_eof_found to 0 at the same time we reset the EOF state Report and fix from Andrew Burgess (GDB) 11/6 ---- lib/readline/display.c - rl_redisplay: if we changed screen lines during the redisplay and we need to return to the line with the cursor, make sure to account for differing numbers of invisible characters in the two lines - rl_redisplay: if we need to redisplay the prompt because we moved into a portion of it containing invisible characters while moving to the line containing the cursor, make sure to account for the cursor being on a line other than 0 and the prompt wrapping lines. If it wraps, we only want to redisplay the portion after the line wrap before the line the cursor is on Fix for C locale issues reported by Grisha Levit 11/7 ---- lib/termcap/termcap.c - tgetent: fix small memory leak on file open error Makefile.in - y.tab.h: move from CREATED_HEADERS to INSTALLED_HEADERS so we don't clean it - maybe-clean: slightly change the syntax of the recipe doc/Makefile.in - PSFILES: split into CREATED_PS and ps files from the distro so we just remove CREATED_PS in maintainer-clean target - PDFFILES: split into CREATED_PDF and ps files from the distro so we just remove CREATED_PDF in maintainer-clean targets - clean,mostlyclean: swap - mostlyclean: don't remove Makefile - maybe-clean: use realpath equivalent like in top-level Makefile.in Report from Mike Jonkmans 11/11 ----- Makefile.in,doc/Makefile.in,lib/glob/Makefile.in,lib/readline/Makefile.in lib/malloc/Makefile.in.,lib/tilde/Makefile.in - mostlyclean: swap with clean where appropriate - add .PHONY target for clean targets where appropriate Makefile.in,configure.ac - STDCKDINT_H: variable for possibly-created file stdckdint.h, for dependencies; add dependencies where appropriate Makefile.in,doc/Makefile.in - add various gnu-standard install targets that do nothing 11/14 ----- CHANGES,NEWS - updated for bash-5.3-beta 11/15 ----- lib/sh/snprintf.c - wchars: fix small memory leak by freeing lbuf if wcrtomb fails Report from SerGoGubarev lib/readline/complete.c - vector_len: an inline convenience function to return the length of an array of strings; callers are remove_duplicate_matches, rl_old_menu_complete, rl_menu_complete - rl_complete_internal: add `$' as a value for what_to_do that exports a (possibly null) list of possible completions to an application or another process by writing them, using a specific protocol, to rl_outstream - rl_export_completions: bindable command to perform word completion and write the results, with other information, to rl_outstream - _rl_export_completions: workhorse function to export the completions Based on a patch from Matthew Tromp , original request from Spencer Baugh back in 11/23 lib/readline/readline.h - rl_export_completions: extern declaration lib/readline/funmap.c - export-completions: new bindable command name doc/bash.1,lib/readline/doc/readline.3,lib/readline/doc/rluser.texi - export-completions: document, including output format 11/19 ----- Makefile.in - HYBRID_HEADERS: new variable containing header files that may appear in the build directory or the source directory - maybe-install-headers: rule to install header files that may appear in the build directory or a subdirectory of it; install-headers depends on it - uninstall-headers: remove $(HYBRID_HEADERS); fix typo in recipe 11/21 ----- builtins/printf.def - chk_converror: inline function to use for single place to check for numeric overflow, change callers to use it 11/25 ----- doc/bash.1,doc/bashref.texi - minor updates to the section on bracket expressions in pattern matching builtins/command.def - update -v help string to indicate that the output is a single word Report from Andrew Davis 11/26 ----- builtins/printf.def - chk_converr: add check for no characters being converted at all, print same warning message as if *ep != 0. This covers the case of an empty string argument. Report by Paul Eggert Makefile.in,doc/Makefile.in - uninstall: make sure the bash-specific directories created by `make install' are emptied and removed 11/27 ----- doc/bash.1,doc/bashref.texi lib/readline/doc/readline.3,lib/readline/doc/history.3 lib/readline/doc/rluser.texi,lib/readline/doc/rltech.texi - another set of updates for formatting consistency, language consistency between man pages and info files 12/1 ---- lib/readline/misc.c - rl_get_previous_history: call _rl_maybe_replace_line with arg of 1 so it clears rl_undo_list, since this may not immediately replace rl_undo_list with something from history Report from Grisha Levit 12/5 ---- doc/bash.1,doc/bashref.texi lib/readline/doc/rluser.texi - updates for active voice, future tense, formatting and fonts, edits to Programmable Completion Suggestions from G. Branden Robinson 12/10 ----- lib/readline/complete.c - fnprint: make sure _rl_completion_prefix_display_length and _rl_colored_completion_prefix are mutually exclusive, since the callers assume they are. bashline.c - bash_default_completion: if we're trying to complete a glob pattern, and rl_completion_type is '?', tell readline not to append a slash if the pattern ends in one. doc/bash.1,lib/readline/doc/readline.3,lib/readline/doc/rluser.texi - completion-prefix-display-length: note that readline uses `___' instead of an ellipsis if the filename begins with a period All reported by Frederick Abell doc/bash.1,doc/bashref.texi - PIPESTATUS: enumerate the commands that set PIPESTATUS From a report by Ulrich Müller lib/readline/isearch.c - _rl_isearch_init: change to call _rl_maybe_replace_line with arg of 1 so rl_undo_list gets cleared (see change from 12/1) Report from Grisha Levit 12/13 ----- CHANGES,NEWS - updated for bash-5.3-beta 12/15 ----- lib/readline/histexpand.c - get_history_event: make sure we don't change history_offset whether the search succeeds or fails, don't set it to history_length. Callers don't expect to have the history offset changed out from under them. Fixes bug reported by Grisha Levit 12/16 ----- lib/readline/util.c - _rl_abort_internal: make sure to clear _rl_command_to_execute Report and patch from Grisha Levit tests/printf7.sub,tests/cond-regexp2.sub - accommodate different error messages across different systems 1/2/2025 -------- tests/run-all, tests/run-minimal - BASHOPTS: unset or unexport as appropriate, same as SHELLOPTS From a report by Martin D Kealey 1/9 --- subst.c - string_extract_verbatim: take into account the fact that CTLESC can quote multibyte characters; use ADVANCE_CHAR instead of increment Fixes bug reported by jktseug@gmail.com 1/10 ---- expr.c - expr_bind_variable,expr_streval: suppress additional expansion if called by the `let' builtin, whether or not array_expand_once is set - expr_skipsubscript: suppress expansion while parsing subscripts whether array_expand_once is set or not Still have to change tests if this goes final These are all conditional on shell_compatibility_level > 51 From a bug-bash post by Greg Wooledge 1/14 ---- lib/readline/complete.c - _rl_export_completions: use print_filename instead of fprintf to display the possible completions so slashes and file types can be included From a patch from Matthew Tromp 1/15 ---- examples/loadables/fltexpr.c - fltexpr: new loadable builtin to do floating-point arithmetic expression evaluation and optionally print the result 1/16 ---- builtins/printf.def - getarg(), advancearg(): cosmetic changes to make it easier to implement %N$ format specifiers in the future 1/18 ---- subst.c - array_length_reference: requesting the length of an element of an unset variable using array syntax is a fatal error when `set -u' is enabled and the subscript is not `@' or `*'. Inspired by a report from Ian Kelling 1/20 ---- parse.y - decode_prompt_string: since Debian has made the spectaculary bad call to allow arbitrary characters in usernames, protect the results of the \u expansion in the same way as \[Ww] and \[Hh] (externally- controlled data) From https://savannah.gnu.org/patch/?10496 1/21 ---- eval.c - reader_loop: if we longjmp with ERREXIT, don't reset any local function contexts in posix mode, since POSIX says to exit as if executing the `exit builtin with no arguments'. Revision of change from 7/2022. From a report by Ivan Shapovalov in 11/2024 lib/readline/display.c - rl_redisplay: after calling update_line, when putting the cursor where it needs to be and reprinting part of the line with the prompt that contains invisible characters, make sure we start to output characters at the portion in the prompt buffer that begins on the current line and only offset by the number of invisible characters on the current line, and set _rl_last_c_pos accordingly in both multibyte and singlebyte locales. Fixes bug reported by Andreas Schwab 1/22 ---- execute_cmd.c - execute_builtin_or_function: if we're executing the exec builtin, and there are redirections, let exec_builtin handle disposing of redirection_undo_list if there is no program name to execute, so it's still there if exec returns and doesn't exit the shell. Then we can manage disposing of exec_redirection_undo_list and letting the redirections be undone as normal. This is POSIX interp 1896, with modifications from https://www.austingroupbugs.net/view.php?id=1896#c7037 builtins/exec.def - exec_builtin: dispose of redirection_undo_list and set it to NULL if there is no program name operand; let execute_builtin_or_function clean it up 1/23 ---- lib/readline/terminal.c - _rl_tcgetwinsize,_rl_tcsetwinsize: function wrappers for tcgetwinsize/ tcsetwinsize with fallbacks to ioctl if those functions aren't available - _rl_get_screensize: use _rl_tcgetwinsize lib/readline/rlwinsize.h - _rl_tcgetwinsize,_rl_tcsetwinsize: extern function declarations lib/readine/rltty.c - set_winsize: use _rl_tcgetwinsize/_rl_tcsetwinsize - set_winsize: if _rl_tcgetwinsize succeeds, use the values to set readline's idea of the screen size via _rl_set_screen_size Inspired by a discussion with Olav Mørkrid 1/24 ---- lib/readline/signals.c - _rl_handling_signal: new private variable, set to the signal we're currently handling in _rl_signal_handler (SIGWINCH) or _rl_handle_signal (all others). Only valid if RL_STATE_SIGHANDLER is set lib/readline/rlprivate.h - _rl_handling_signal: new extern declaration lib/readline/rltty.c - set_winsize: only set readline's idea of the screen size if we're executing in a signal handler context and handling SIGTSTP and the application has indicated that it doesn't want the environment variables LINES and COLUMNS to have priority (rl_prefer_env_winsize) 1/30 ---- jobs.c - wait_for_single_pid: if wait_for returns -1/ECHILD, make sure to return 257 just as if the child was not found, optionally printing an error message if JWAIT_PERROR is supplied From a discussion with Ian 2/8 --- redir.c - do_redirection_internal: keep track of whether the redirector fd was active with a new variable. Not used yet, should be optimized out. - add_undo_fd_redirect: new undo function using r_move_input so we don't need a dup redir and a separate close redir to undo. Not used yet - do_redirection_internal: only call check_bash_input if we're creating an undo list and not running one command.h - OUTPUT_REDIRECT: add r_output_force input.c - check_bash_input: if we dup the bash input fd to another in save_bash_input, return that fd. Continue to return -1 on error and 0 if we don't do anything but sync execute_cmd.c - execute_null_command: force a subshell if any of the redirections involve the bash input buffered fd. From a report by Martin D Kealey 2/10 ---- examples/loadables/chmod.c - chmod: new loadable builtin, inspired by zhixu.liu@gmail.com in https://savannah.gnu.org/patch/?10499 2/11 ---- builtins/kill.def - kill_builtin: in posix mode, return a failure status if any of the pid/job arguments are not found or sending the signal fails. This is consistent with https://pubs.opengroup.org/onlinepubs/9799919799/utilities/kill.html#tag_20_64_14 2/12 ---- braces.c - brace_gobbler: take a new argument in which to return the type of brace expansion (comma or sequence) we think we've found; change all callers - brace_expand: even if we find a BRACE_ARG_SEPARATOR, don't try to perform expansion if brace_gobbler doesn't think we have a BRACE_COMMA expansion - valid_seqterm: new function, performs minimal validation on a sequence expansion, used to skip over invalid sequence expressions in brace_expand - brace_expand: call valid_seqterm if brace_gobbler finds a BRACE_SEQ expansion Inspired by report from Robert Elz 2/13 ---- doc/bash.1,doc/bashref.texi - clarified the handling of double quotes in EXPRESSION in (( EXPRESSION )) and $(( EXPRESSION )) to say that only unescaped double quotes are removed expr.c - expr_bind_variable,expr_streval: revert change from 1/10 that suppresses additional expansion when called by the `let' builtin, tag for possible inclusion in a future release when I can look at parallel changes to ((...)) and $((...)) (and probably declare 'a[" "]=12' or similar) 2/14 ---- lib/sh/zmapfd.c,externs.h - zmapfd: now returns ssize_t in case of large files when size_t is larger than int builtins/evalfile.c,builtins/help.def - zmapfd: changed callers print_cmd.c - print_redirection: the default file descriptor for <> is 0 Report and fix from Emanuele Torre 2/18 ---- configure.ac - check for d_type member of struct dirent for future use 2/24 ---- doc/bash.1,doc/bashref.texi - document the word expansions performed on indexed array subscripts and assoc array keys - document the possibility and effect of empty strings on arithmetic evaluation for substring expansion, arithmetic expansion, operands to the the [[ command's arithmetic operators, the expression in the (( compound command, and indexed array subscripts 2/24 ---- execute_cmd.c - execute_command_internal: if executing a (...) subshell, make sure to update BASH_COMMAND before exiting on an error return from the subshell with -e set Report from Max Bowsher 3/3 --- redir.c - here_document_to_fd: if we are trying to use a pipe for a here document, but the F_GETPIPE_SZ fcntl returns a value shorter than the here document length, make sure to close the pipe file descriptors builtins/ulimit.def - pipesize: use pathconf and _PC_PIPE_BUF if available 3/8 --- braces.c - brace_expand: make sure to call valid_seqterm with the right length so that text[tlen] == '}' 3/10 ---- builtins/help.def - show_manpage: if the first line of long_doc doesn't contain a newline, e.g., if it's a loadable builtin, add one so there's the usual space between NAME and SYNOPSIS m4/codeset.m4 - define HAVE_NL_LANGINFO lib/sh/strftime.c - updates for locale-awareness 3/14 ---- lib/termcap/termcap.c,lib/termcap/tparam.c - updates for ISO C 3/17 ---- configure.ac - update release to rc1, update CHANGES and NEWS 3/19 ---- lib/readline/rlmbutil.h - _rl_wcwidth: make sure wchar_t constants have an `L' prefix 3/24 ---- subst.c - function_substitute: set temporary_env to NULL (we've already set up an unwind-protect for it) before pushing the context, so pop_context doesn't free it out from underneath the caller. Report from Emanuele Torre 3/27 ---- shell.c - set_bash_input: if the call to with_input_from_buffered_stream does not result in a valid buffer corresponding to the file descriptor, print an error message and set last_command_exit_value From a report in https://savannah.gnu.org/patch/?10512 4/3 --- lib/sh/gettimeofday.c - updated from version in gnulib [bash-5.3-rc1 frozen] 4/4 --- examples/loadables/fltexpr.c - add support for function calls; support almost all math function calls in math.h - add support for pi, gamma, e mathematical constants 4/7 --- doc/bash.1,doc/bashref.texi - HISTFILESIZE: note that the history file will contain complete history entries whose lines may total slightly more than $HISTFILESIZE if the entries contain multiple lines From a report from Jens Schmidt - export: note that you can use -f with -p to display exported functions From a report from Duncan Roe builtins/history.def - add history -d start-end to the long doc From a report from Duncan Roe 4/9 --- lib/readline/histfile.c - read_history_range: changes to not skip blank lines if we are reading a history file with multiline history entries. From a report by Jens Schmidt 4/11 ---- bashline.c - attempt_shell_completion: move the check for char_is_quoted from check_redir to the caller, set in_command_position to 0 if the command separator is quoted From https://savannah.gnu.org/support/?111224 david@mandelberg.org - check_extglob: break check for extended glob out into separate function, call from attempted_shell_completion, set in_command_position to -1 if it returns 1, as with the old call to check_redir 4/21 ---- builtins/read.def - read_mbchar: if we read a delimiter character that makes an incomplete multibyte character into an invalid multibyte character, only push the delimiter back if we read that character ourselves (i > 1). Report from Greg Wooledge 4/23 ---- parse.y - pop_delimiter: only decrement delimiter_depth if it's > 0, since reset_parser() may have set it to 0 after the matching call to push_delimiter Report from Grisha Levit based on a report from Александр Ушаков 4/25 ---- parse.y - read_a_line: if the shell is interactive, and not reading from a string, check whether a previous call to shell_getc has set EOF_Reached and return EOF in this case, after resetting the current token to '\n'. This makes EOFs that are not the first character on the line `sticky' instead of just token delimiters. From https://savannah.gnu.org/bugs/?67045 - history_delimiting_chars: if it looks like we just finished a subshell, and the line we're adding begins with an operator that can't follow a semicolon, return a newline From https://savannah.gnu.org/patch/?10517 4/28 ---- parse.y - parse_arith_command: if the character after the first right paren isn't a right paren, making the construct a nested subshell, push that character back and return the subshell command as the current token string, which we push onto the pushed string list. Reading one character more can cause synchronization problems with backslash newline, among other things. From https://savannah.gnu.org/patch/?10517 4/29 ---- builtins/evalstring.c - parse_string: only run the top-level unwind-protects if we're actually going to be calling jump_to_top_level(); otherwise let xparse_dolparen take care of it From a report by Александр Ушаков 5/1 --- lib/sh/zread.c - zungetc: rework to use a local 16-byte buffer so we can push back multiple characters whether we use read or zread - zread/zreadretry/zreadintr/zreadc/zreadcintr/zreadn: changed to use zpopbuf to check whether we have pushed back bytes - zreset,zsyncfd: reset the pushback buffer indices builtins/read.def - read_mbchar: handle the delimiter being part of an invalid multibyte character, not just the byte that makes the multibyte character invalid: keep track of where the delimiter char is in the buffer and use zungetc to push back that character and everything we read after it so subsequent buffered (zread) or unbuffered (read) reads will return them 5/2 --- lib/readline/display.c,lib/readline/histexpand.c,lib/readline/history.c, lib/readline/input.c,lib/readline/search.c,lib/readline/shell.c, lib/readline/util.c,lib/readline/readline.c,lib/readline/histfile.c - fixes for issues (unused variables, etc.) uncovered by static analysis All from a report by Siteshwar Vashisht 5/16 ---- lib/readline/terminal.c - NEED_EXTERN_PC: add __gnu_hurd__ to the systems that require this if not using ncurses; rework the structure of the defines Report from Collin Funk braces.c - brace_gobbler: don't mark an expression as BRACE_SEQ if we've already seen a comma and marked it as BRACE_COMMA, so we treat it as a comma brace expansion. From a report by Sam James support/install-sh - new version from coreutils-git (2024-12-03.03); renamed from install.sh aclocal.m4 - AM_PROG_INSTALL_SH, AM_PROG_INSTALL_STRIP: added from automake configure.ac - call AM_PROG_INSTALL_SH, AM_PROG_INSTALL_STRIP Makefile.in - STRIP: let configure substitute it from AC_CHECK_TOOL result - INSTALL_STRIP_PROGRAM: let configure substitute it from AM_PROG_INSTALL_STRIP - install-strip: change to set INSTALL_PROGRAM=${INSTALL_STRIP_PROGRAM} to avoid hard-coding `-s' in the recipe All from a report by NR 5/18 ---- subst.c - param_expand: if we expand a nameref that references an unset array variable reference, make sure we enforce `nounset', with the usual exceptions for `@' and `*' subscripts From a report from Emanuele Torre 5/22 ---- buildconf.h.in - add additional defines for conservative ISO C/POSIX environment - add HAVE_C_BOOL define in case someone cross-compiles with a C23 compiler From a report by Bruce Dubbs builtins/mkbuiltins.c - include buildconf.h instead of having separate defines when cross-compiling 6/11 ---- lib/readline/terminal.c - rework change from 5/16 to add __gnu_hurd__ to the list of systems that don't need extern PC/UP/BC Reported by Joel Ebel as a static link failure due to multiply-defined symbols 6/12 ---- general.c - valid_function_word: back out change from 2/4/2023 that allowed quoted characters in function names. Inspired by discussion with Robert Elz and Koichi Murase parse.y - read_token_word: set dollar_present if we see <( or >), backing out change from 2/4/2023