string.c: answer an ASCII-case build from the header's fallback - #7222

include/mruby/internal.h already answers -1 for a build that compiles no
case tables:

/* include/mruby/internal.h */
#ifdef MRB_UTF8_STRING
int mrb_str_case_convert_unicode(mrb_state *mrb, mrb_value str, enum mrb_case_mode mode);
#else
#define mrb_str_case_convert_unicode(mrb, str, mode) (-1)
#endif

The guard asks only about MRB_UTF8_STRING, but that is not the condition the
tables are compiled under. MRB_USE_ASCII_CASE keeps MRB_UTF8_STRING and
drops the tables, so such a build takes the declared side of the guard, and
src/string.c has to supply a body for the promise:

/* src/string.c */
#elif defined(MRB_UTF8_STRING)
int
mrb_str_case_convert_unicode(mrb_state *mrb, mrb_value str, enum mrb_case_mode mode)
{
  (void)mrb;
  (void)str;
  (void)mode;
  return -1;
}
#endif  /* MRB_UTF8_STRING */

Two spellings of one answer, and the one costing a call and a return is the one
that build takes.

The change

Widen the guard to the condition the tables are actually compiled under, and
the #elif arm goes away whole:

#if defined(MRB_UTF8_STRING) && !defined(MRB_USE_ASCII_CASE)
int mrb_str_case_convert_unicode(mrb_state *mrb, mrb_value str, enum mrb_case_mode mode);
#else
#define mrb_str_case_convert_unicode(mrb, str, mode) (-1)
#endif

That condition is the one src/string.c opens the walk with and the one
include/mruby/internal.h already uses a few lines below for the codepoint
accessors, so the header now names the tables the same way in both places.

All seven call sites pass plain locals, so a macro evaluating none of its
arguments answers each of them as before:

Each already reads the result as int uc = ...; if (uc >= 0) or as ... < 0,
which (-1) satisfies. A build without MRB_UTF8_STRING has taken the macro
for as long as it has existed, so this is the path those five methods are
already compiled from where the tables are absent; what changes is which builds
count as absent.

The stale #endif /* MRB_UTF8_STRING */ comment, which closed a chain whose
first arm is MRB_UTF8_STRING && !MRB_USE_ASCII_CASE, is carried off by the
same change.

Size

.text summed over every .o, each side built from an empty build directory:

ascii-case is the only build whose guard moves, and every other one is
unchanged to the byte, which is what a change that only redraws a #if should
show.

Testing

build_config/ci/gcc-clang.rb and build_config/asan.rb, run per build so the
counts are attributable:

Identical to master, build for build. bintest (the binary tests) passes 117
of 117 under ci/gcc-clang and 79 of 79 under asan.

No new tests come with this. ascii-case already asks all five methods for the
-1 answer: mrbgems/mruby-string-ext/test/string.rb skips only the Unicode
assertions of String#swapcase and String#casecmp? there and keeps the ASCII
ones, and test/t/string.rb asserts downcase, upcase and capitalize on
strings the walk would have refused either way.

Environment

# ci/gcc-clang, full-debug
gcc -std=gnu99 -g -O3 -Wall -Wundef -Werror-implicit-function-declaration -Wwrite-strings -g3 -O0 -DMRB_GC_STRESS -DMRB_USE_DEBUG_HOOK -DMRB_DEBUG -DMRB_USE_BIGINT -DMRB_USE_COMPLEX -DHAVE_MRUBY_ENCODING_GEM -DMRB_UTF8_STRING -DHAVE_MRUBY_IO_GEM -DMRB_USE_RATIONAL -DHAVE_MRUBY_REGEXP_GEM -DMRB_USE_SET -DMRB_USE_TASK_SCHEDULER src/string.c

# ci/gcc-clang, bintest
gcc -std=gnu99 -g -O3 -Wall -Wundef -Werror-implicit-function-declaration -Wwrite-strings -DMRB_GC_FIXED_ARENA -DMRB_USE_BIGINT -DMRB_USE_COMPLEX -DHAVE_MRUBY_ENCODING_GEM -DMRB_UTF8_STRING -DHAVE_MRUBY_IO_GEM -DMRB_USE_RATIONAL -DHAVE_MRUBY_REGEXP_GEM -DMRB_USE_SET -DMRB_USE_TASK_SCHEDULER -DMRB_USE_DEBUG_HOOK src/string.c

# ci/gcc-clang, cxx_abi
gcc -g -O3 -Wall -Wundef -Wwrite-strings -x c++ -std=gnu++03 -DMRB_GC_FIXED_ARENA -DMRB_USE_CXX_EXCEPTION -DMRB_USE_CXX_ABI -DMRB_USE_BIGINT -DMRB_USE_COMPLEX -DHAVE_MRUBY_ENCODING_GEM -DMRB_UTF8_STRING -DHAVE_MRUBY_IO_GEM -DMRB_USE_RATIONAL -DHAVE_MRUBY_REGEXP_GEM -DMRB_USE_SET -DMRB_USE_TASK_SCHEDULER src/string.c

# ci/gcc-clang, byte-string
gcc -std=gnu99 -g -O3 -Wall -Wundef -Werror-implicit-function-declaration -Wwrite-strings -DMRB_USE_BIGINT -DMRB_USE_COMPLEX -DHAVE_MRUBY_IO_GEM -DMRB_USE_RATIONAL -DHAVE_MRUBY_REGEXP_GEM -DMRB_USE_SET -DMRB_USE_TASK_SCHEDULER src/string.c

# ci/gcc-clang, ascii-case
gcc -std=gnu99 -g -O3 -Wall -Wundef -Werror-implicit-function-declaration -Wwrite-strings -DMRB_USE_ASCII_CASE -DMRB_USE_BIGINT -DMRB_USE_COMPLEX -DHAVE_MRUBY_ENCODING_GEM -DMRB_UTF8_STRING -DHAVE_MRUBY_IO_GEM -DMRB_USE_RATIONAL -DHAVE_MRUBY_REGEXP_GEM -DMRB_USE_SET -DMRB_USE_TASK_SCHEDULER src/string.c

# build_config/asan.rb
gcc -std=gnu99 -g -O3 -Wall -Wundef -Werror-implicit-function-declaration -Wwrite-strings -fsanitize=address,undefined -g3 -O0 -DMRB_DEBUG -DMRB_USE_BIGINT -DMRB_USE_COMPLEX -DHAVE_MRUBY_ENCODING_GEM -DMRB_UTF8_STRING -DHAVE_MRUBY_IO_GEM -DMRB_USE_RATIONAL -DHAVE_MRUBY_REGEXP_GEM -DMRB_USE_SET -DMRB_USE_TASK_SCHEDULER src/string.c

Summary by CodeRabbit

• Bug Fixes
  • Corrected Unicode string case conversion behavior for configurations using ASCII case conversion.
  • Prevented unavailable Unicode case-conversion functionality from being exposed in those configurations.