stack-graphs/stack-graphs/include/stack-graphs.h at stack-orientation · github/stack-graphs

This repository was archived by the owner on Sep 9, 2025. It is now read-only.
862 lines (760 loc) · 44 KB
#ifndef STACK_GRAPHS_H_
#define STACK_GRAPHS_H_
/* Warning, this file is autogenerated by cbindgen. Don't modify this manually. */
#include <stdarg.h>
#include <stdbool.h>
#include <stddef.h>
#include <stdint.h>
#include <stdlib.h>
// The null value for all of our handles.
#define SG_NULL_HANDLE 0
// The handle of an empty list.
#define SG_LIST_EMPTY_HANDLE 4294967295
// The local_id of the singleton root node.
#define SG_ROOT_NODE_ID 1
// The local_id of the singleton "jump to scope" node.
#define SG_JUMP_TO_NODE_ID 2
// Describes in which direction the content of a deque is stored in memory.
enum sg_deque_direction {
    SG_DEQUE_FORWARDS,
    SG_DEQUE_BACKWARDS,
// The different kinds of node that can appear in a stack graph.
enum sg_node_kind {
    // Removes everything from the current scope stack.
    SG_NODE_KIND_DROP_SCOPES,
    // The singleton "jump to" node, which allows a name binding path to jump back to another part
    // of the graph.
    SG_NODE_KIND_JUMP_TO,
    // Pops a scoped symbol from the symbol stack.  If the top of the symbol stack doesn't match
    // the requested symbol, or if the top of the symbol stack doesn't have an attached scope
    // list, then the path is not allowed to enter this node.
    SG_NODE_KIND_POP_SCOPED_SYMBOL,
    // Pops a symbol from the symbol stack.  If the top of the symbol stack doesn't match the
    // requested symbol, then the path is not allowed to enter this node.
    SG_NODE_KIND_POP_SYMBOL,
    // Pushes a scoped symbol onto the symbol stack.
    SG_NODE_KIND_PUSH_SCOPED_SYMBOL,
    // Pushes a symbol onto the symbol stack.
    SG_NODE_KIND_PUSH_SYMBOL,
    // The singleton root node, which allows a name binding path to cross between files.
    SG_NODE_KIND_ROOT,
    // A node that adds structure to the graph. If the node is exported, it can be
    // referred to on the scope stack, which allows "jump to" nodes in any other
    // part of the graph can jump back here.
    SG_NODE_KIND_SCOPE,
// Describes the result of a computation
enum sg_result {
    SG_RESULT_SUCCESS,
    SG_RESULT_CANCELLED,
// Manages the state of a collection of partial paths to be used in the path-stitching algorithm.
struct sg_partial_path_arena;
// Contains a "database" of partial paths.
// This type is meant to be a lazily loaded "view" into a proper storage layer.  During the
// path-stitching algorithm, we repeatedly try to extend a currently incomplete path with any
// partial paths that are compatible with it.  For large codebases, or projects with a large
// number of dependencies, it can be prohibitive to load in _all_ of the partial paths up-front.
// We've written the path-stitching algorithm so that you have a chance to only load in the
// partial paths that are actually needed, placing them into a sg_partial_path_database instance
// as they're needed.
struct sg_partial_path_database;
// A list of paths found by the path-finding algorithm.
struct sg_partial_path_list;
// Contains all of the nodes and edges that make up a stack graph.
struct sg_stack_graph;
// A name that we are trying to resolve using stack graphs.
// This typically represents a portion of an identifier as it appears in the source language.  It
// can also represent some other "operation" that can occur in source code, and which needs to be
// modeled in a stack graph — for instance, many languages will use a "fake" symbol named `.` to
// represent member access.
struct sg_symbol {
    const char *symbol;
    size_t symbol_len;
// An array of all of the symbols in a stack graph.  Symbol handles are indices into this array.
// There will never be a valid symbol at index 0; a handle with the value 0 represents a missing
struct sg_symbols {
    const struct sg_symbol *symbols;
    size_t count;
// A handle to a symbol in a stack graph.  A zero handle represents a missing symbol.
// We deduplicate symbols in a stack graph — that is, we ensure that there are never multiple
// `struct sg_symbol` instances with the same content.  That means that you can compare symbol
// handles using simple equality, without having to dereference them.
typedef uint32_t sg_symbol_handle;
// Arbitrary string content associated with some part of a stack graph.
struct sg_string {
    const char *content;
    size_t length;
// An array of all of the interned strings in a stack graph.  String handles are indices into this
// array. There will never be a valid string at index 0; a handle with the value 0 represents a
// missing string.
struct sg_strings {
    const struct sg_string *strings;
    size_t count;
// A handle to an interned string in a stack graph.  A zero handle represents a missing string.
// We deduplicate strings in a stack graph — that is, we ensure that there are never multiple
// `struct sg_string` instances with the same content.  That means that you can compare string
// handles using simple equality, without having to dereference them.
typedef uint32_t sg_string_handle;
// A source file that we have extracted stack graph data from.
// It's up to you to choose what names to use for your files, but they must be unique within a
// stack graph.  If you are analyzing files from the local filesystem, the file's path is a good
// choice.  If your files belong to packages or repositories, they should include the package or
// repository IDs to make sure that files in different packages or repositories don't clash with
// each other.
struct sg_file {
    const char *name;
    size_t name_len;
// An array of all of the files in a stack graph.  File handles are indices into this array.
// There will never be a valid file at index 0; a handle with the value 0 represents a missing
struct sg_files {
    const struct sg_file *files;
    size_t count;
// A handle to a file in a stack graph.  A zero handle represents a missing file.
// We deduplicate files in a stack graph — that is, we ensure that there are never multiple
// `struct sg_file` instances with the same filename.  That means that you can compare file
// handles using simple equality, without having to dereference them.
typedef uint32_t sg_file_handle;
// Uniquely identifies a node in a stack graph.
// Each node (except for the _root node_ and _jump to scope_ node) lives in a file, and has a
// _local ID_ that must be unique within its file.
struct sg_node_id {
    sg_file_handle file;
    uint32_t local_id;
// A node in a stack graph.
struct sg_node {
    enum sg_node_kind kind;
    struct sg_node_id id;
    // The symbol associated with this node.  For push nodes, this is the symbol that will be
    // pushed onto the symbol stack.  For pop nodes, this is the symbol that we expect to pop off
    // the symbol stack.  For all other node types, this will be null.
    sg_symbol_handle symbol;
    // The scope associated with this node.  For push scope nodes, this is the scope that will be
    // attached to the symbol before it's pushed onto the symbol stack.  For all other node types,
    // this will be null.
    struct sg_node_id scope;
    // Whether this node is an endpoint.  For push nodes, this indicates that the node represents
    // a reference in the source.  For pop nodes, this indicates that the node represents a
    // definition in the source.  For scopes, this indicates that the scope is exported. For all
    // other node types, this field will be unused.
    bool is_endpoint;
// An array of all of the nodes in a stack graph.  Node handles are indices into this array.
// There will never be a valid node at index 0; a handle with the value 0 represents a missing
struct sg_nodes {
    const struct sg_node *nodes;
    size_t count;
// A handle to a node in a stack graph.  A zero handle represents a missing node.
typedef uint32_t sg_node_handle;
// Connects two nodes in a stack graph.
// These edges provide the basic graph connectivity that allow us to search for name binding paths
// in a stack graph.  (Though not all sequence of edges is a well-formed name binding: the nodes
// that you encounter along the path must also satisfy all of the rules for maintaining correct
// symbol and scope stacks.)
struct sg_edge {
    sg_node_handle source;
    sg_node_handle sink;
    int32_t precedence;
// The offset of a character within a string (typically a line of source code), using several
// different units
// All offsets are 0-indexed.
struct sg_offset {
    // The number of UTF-8-encoded bytes appearing before this character in the string
    size_t utf8_offset;
    // The number of UTF-16 code units appearing before this character in the string
    size_t utf16_offset;
    // The number of graphemes appearing before this character in the string
    size_t grapheme_offset;
// A half-open range identifying a range of characters in a string.
struct sg_utf8_bounds {
    // The UTF-8 byte index of the first character in the range.
    size_t start;
    // The UTF-8 byte index of the first character _after_ the range.
    size_t end;
// All of the position information that we have about a character in a source file
struct sg_position {
    // The 0-indexed line number containing the character
    size_t line;
    // The offset of the character within its containing line, expressed as both a UTF-8 byte
    // index and a UTF-16 code point index
    struct sg_offset column;
    // The UTF-8 byte indexes (within the file) of the start and end of the line containing the
    // character
    struct sg_utf8_bounds containing_line;
    // The UTF-8 byte indexes (within the file) of the start and end of the line containing the
    // character, with any leading and trailing whitespace removed
    struct sg_utf8_bounds trimmed_line;
// All of the position information that we have about a range of content in a source file
struct sg_span {
    struct sg_position start;
    struct sg_position end;
// Contains information about a range of code in a source code file.
struct sg_source_info {
    // The location in its containing file of the source code that this node represents.
    struct sg_span span;
    // The kind of syntax entity this node represents (e.g. `function`, `class`, `method`, etc.).
    sg_string_handle syntax_type;
    // The full content of the line containing this node in its source file.
    sg_string_handle containing_line;
    // The location in its containing file of the source code that this node's definiens represents.
    // This is used for things like the bodies of functions, rather than the RHSes of equations.
    // If you need one of these to make the type checker happy, but you don't have one, just use
    // sg_span::default(), as this will correspond to the all-0s spans which mean "no definiens".
    struct sg_span definiens_span;
// An array of all of the source information in a stack graph.  Source information is associated
// with nodes, so node handles are indices into this array.  It is _not_ guaranteed that there
// will an entry in this array for every node handle; if you have a node handle whose value is
// larger than `count`, then use a 0-valued `sg_source_info` if you need source information for
// that node.
// There will never be a valid entry at index 0; a handle with the value 0 represents a missing
struct sg_source_infos {
    const struct sg_source_info *infos;
    size_t count;
// A tuple of a node handle and source information for that node.  Used with the
// `sg_add_source_info` function to add source information to a stack graph.
struct sg_node_source_info {
    sg_node_handle node;
    struct sg_source_info source_info;
// A handle to an element of a partial scope stack.  A zero handle represents a missing partial
// scope stack.  A UINT32_MAX handle represents an empty partial scope stack.
typedef uint32_t sg_partial_scope_stack_cell_handle;
// Represents an unknown list of exported scopes.
typedef uint32_t sg_scope_stack_variable;
// A pattern that might match against a scope stack.  Consists of a (possibly empty) list of
// exported scopes, along with an optional scope stack variable.
struct sg_partial_scope_stack {
    // The handle of the first element in the partial scope stack, or SG_LIST_EMPTY_HANDLE if the
    // list is empty, or 0 if the list is null.
    sg_partial_scope_stack_cell_handle cells;
    uint32_t length;
    // The scope stack variable representing the unknown content of a partial scope stack, or 0 if
    // the variable is missing.  (If so, this partial scope stack can only match a scope stack
    // with exactly the list of scopes in `cells`, instead of any scope stack with those scopes as
    // a prefix.)
    sg_scope_stack_variable variable;
// A symbol with an unknown, but possibly empty, list of exported scopes attached to it.
struct sg_partial_scoped_symbol {
    sg_symbol_handle symbol;
    struct sg_partial_scope_stack scopes;
// A handle to an element of a partial symbol stack.  A zero handle represents a missing partial
// symbol stack.  A UINT32_MAX handle represents an empty partial symbol stack.
typedef uint32_t sg_partial_symbol_stack_cell_handle;
// An element of a partial symbol stack.
struct sg_partial_symbol_stack_cell {
    // The partial scoped symbol at this position in the partial symbol stack.
    struct sg_partial_scoped_symbol head;
    // The handle of the next element in the partial symbol stack, or SG_LIST_EMPTY_HANDLE if this
    // is the last element.
    sg_partial_symbol_stack_cell_handle tail;
    // The handle of the reversal of this partial scope stack.
    sg_partial_symbol_stack_cell_handle reversed;
// The array of all of the partial symbol stack content in a partial path arena.
struct sg_partial_symbol_stack_cells {
    const struct sg_partial_symbol_stack_cell *cells;
    size_t count;
// Represents an unknown list of scoped symbols.
typedef uint32_t sg_symbol_stack_variable;
// A pattern that might match against a symbol stack.  Consists of a (possibly empty) list of
// partial scoped symbols.
// (Note that unlike partial scope stacks, we don't store any "symbol stack variable" here.  We
// could!  But with our current path-finding rules, every partial path will always have exactly
// one symbol stack variable, which will appear at the end of its precondition and postcondition.
// So for simplicity we just leave it out.  At some point in the future we might add it in so that
// the symbol and scope stack formalisms and implementations are more obviously symmetric.)
struct sg_partial_symbol_stack {
    // The handle of the first element in the partial symbol stack, or SG_LIST_EMPTY_HANDLE if the
    // list is empty, or 0 if the list is null.
    sg_partial_symbol_stack_cell_handle cells;
    uint32_t length;
    // The symbol stack variable representing the unknown content of a partial symbol stack, or 0
    // if the variable is missing.  (If so, this partial symbol stack can only match a symbol
    // stack with exactly the list of symbols in `cells`, instead of any symbol stack with those
    // symbols as a prefix.)
    sg_symbol_stack_variable variable;
// An element of a partial scope stack.
struct sg_partial_scope_stack_cell {
    // The exported scope at this position in the partial scope stack.
    sg_node_handle head;
    // The handle of the next element in the partial scope stack, or SG_LIST_EMPTY_HANDLE if this
    // is the last element.
    sg_partial_scope_stack_cell_handle tail;
    // The handle of the reversal of this partial scope stack.
    sg_partial_scope_stack_cell_handle reversed;
// The array of all of the partial scope stack content in a partial path arena.
struct sg_partial_scope_stack_cells {
    const struct sg_partial_scope_stack_cell *cells;
    size_t count;
// Details about one of the edges in a partial path
struct sg_partial_path_edge {
    struct sg_node_id source_node_id;
    int32_t precedence;
// A handle to an element of a partial path edge list.  A zero handle represents a missing partial
// path edge list.  A UINT32_MAX handle represents an empty partial path edge list.
typedef uint32_t sg_partial_path_edge_list_cell_handle;
// An element of a partial path edge list.
struct sg_partial_path_edge_list_cell {
    // The partial path edge at this position in the partial path edge list.
    struct sg_partial_path_edge head;
    // The handle of the next element in the partial path edge list, or SG_LIST_EMPTY_HANDLE if
    // this is the last element.
    sg_partial_path_edge_list_cell_handle tail;
    // The handle of the reversal of this list.
    sg_partial_path_edge_list_cell_handle reversed;
// The array of all of the partial path edge list content in a partial path arena.
struct sg_partial_path_edge_list_cells {
    const struct sg_partial_path_edge_list_cell *cells;
    size_t count;
// The edges in a path keep track of precedence information so that we can correctly handle
// shadowed definitions.
struct sg_partial_path_edge_list {
    // The handle of the first element in the edge list, or SG_LIST_EMPTY_HANDLE if the list is
    // empty, or 0 if the list is null.
    sg_partial_path_edge_list_cell_handle cells;
    enum sg_deque_direction direction;
    uint32_t length;
// A portion of a name-binding path.
// Partial paths can be computed _incrementally_, in which case all of the edges in the partial
// path belong to a single file.  At query time, we can efficiently concatenate partial paths to
// yield a name-binding path.
// Paths describe the contents of the symbol stack and scope stack at the end of the path.
// Partial paths, on the other hand, have _preconditions_ and _postconditions_ for each stack.
// The precondition describes what the stack must look like for us to be able to concatenate this
// partial path onto the end of a path.  The postcondition describes what the resulting stack
// looks like after doing so.
// The preconditions can contain _scope stack variables_, which describe parts of the scope stack
// (or parts of a scope symbol's attached scope list) whose contents we don't care about.  The
// postconditions can _also_ refer to those variables, and describe how those variable parts of
// the input scope stacks are carried through unmodified into the resulting scope stack.
struct sg_partial_path {
    sg_node_handle start_node;
    sg_node_handle end_node;
    struct sg_partial_symbol_stack symbol_stack_precondition;
    struct sg_partial_symbol_stack symbol_stack_postcondition;
    struct sg_partial_scope_stack scope_stack_precondition;
    struct sg_partial_scope_stack scope_stack_postcondition;
    struct sg_partial_path_edge_list edges;
// An array of all of the partial paths in a partial path database.  Partial path handles are
// indices into this array.  There will never be a valid partial path at index 0; a handle with
// the value 0 represents a missing partial path.
struct sg_partial_paths {
    const struct sg_partial_path *paths;
    size_t count;
// A handle to a partial path in a partial path database.  A zero handle represents a missing
// partial path.
typedef uint32_t sg_partial_path_handle;
// Encodes a set of node handles.
// The elements are encoded in a bit set.  Use the traditional mask and shift pattern to determine
// if a particular handle is in the set:
// size_t element_index = handle / 32;
// size_t bit_index = handle % 32;
// size_t bit_mask = 1 << bit_index;
// bool bit_is_set =
//     element_index < set.length &&
//     (set.elements[element_index] & bit_mask) != 0;
struct sg_node_handle_set {
    const uint32_t *elements;
    // Note that this is the number of uint32_t's in `elements`, NOT the number of bits in the set.
    size_t length;
// Implements a phased forward partial path stitching algorithm.
// Our overall goal is to start with a set of _seed_ partial paths, and to repeatedly extend each
// partial path by concatenating another, compatible partial path onto the end of it.  (If there
// are multiple compatible partial paths, we concatenate each of them separately, resulting in
// more than one extension for the current path.)
// We perform this processing in _phases_.  At the start of each phase, we have a _current set_ of
// partial paths that need to be processed.  As we extend those partial paths, we add the
// extensions to the set of partial paths to process in the _next_ phase.  Phases are processed
// one at a time, each time you invoke `sg_forward_partial_path_stitcher_process_next_phase`.
// After each phase has completed, the `previous_phase_paths` and `previous_phase_paths_length`
// fields give you all of the partial paths that were discovered during that phase.  That gives
// you a chance to add to the `sg_partial_path_database` all of the other partial paths that we
// might need to extend those partial paths with before invoking the next phase.
struct sg_forward_partial_path_stitcher {
    // The new candidate partial paths that were discovered in the most recent phase.
    const struct sg_partial_path *previous_phase_partial_paths;
    // The number of new candidate partial paths that were discovered in the most recent phase.
    // If this is 0, then the partial path stitching algorithm is complete.
    size_t previous_phase_partial_paths_length;
    // Whether the stitching algorithm is complete.  You should keep calling
    // `sg_forward_partial_path_stitcher_process_next_phase` until this field is true.
    bool is_complete;
// The handle of the singleton root node.
#define SG_ROOT_NODE_HANDLE 1
// The handle of the singleton "jump to scope" node.
#define SG_JUMP_TO_NODE_HANDLE 2
#ifdef __cplusplus
extern "C" {
#endif // __cplusplus
// Creates a new, initially empty stack graph.
struct sg_stack_graph *sg_stack_graph_new(void);
// Frees a stack graph, and all of its contents.
void sg_stack_graph_free(struct sg_stack_graph *graph);
// Creates a new, initially empty partial path arena.
struct sg_partial_path_arena *sg_partial_path_arena_new(void);
// Frees a path arena, and all of its contents.
void sg_partial_path_arena_free(struct sg_partial_path_arena *partials);
// Creates a new, initially empty partial path database.
struct sg_partial_path_database *sg_partial_path_database_new(void);
// Frees a partial path database, and all of its contents.
void sg_partial_path_database_free(struct sg_partial_path_database *db);
// Ensures all partial paths in the database are availabe in both forwards and backwards orientation.
void sg_partial_path_database_ensure_both_directions(struct sg_partial_path_database *db,
                                                     struct sg_partial_path_arena *partials);
// Ensures all partial paths in the database are in forwards orientation.
void sg_partial_path_database_ensure_forwards(struct sg_partial_path_database *db,
                                              struct sg_partial_path_arena *partials);
// Returns a reference to the array of symbol data in this stack graph.  The resulting array
// pointer is only valid until the next call to any function that mutates the stack graph.
struct sg_symbols sg_stack_graph_symbols(const struct sg_stack_graph *graph);
// Adds new symbols to the stack graph.  You provide all of the symbol content concatenated
// together into a single string, and an array of the lengths of each symbol.  You also provide an
// output array, which must have the same size as `lengths`.  We will place each symbol's handle
// in the output array.
// We ensure that there is only ever one copy of a particular symbol stored in the graph — we
// guarantee that identical symbols will have the same handles, meaning that you can compare the
// handles using simple integer equality.
// We copy the symbol data into the stack graph.  The symbol content you pass in does not need to
// outlive the call to this function.
// Each symbol must be a valid UTF-8 string.  If any symbol isn't valid UTF-8, it won't be added
// to the stack graph, and the corresponding entry in the output array will be the null handle.
void sg_stack_graph_add_symbols(struct sg_stack_graph *graph,
                                size_t count,
                                const char *symbols,
                                const size_t *lengths,
                                sg_symbol_handle *handles_out);
// Returns a reference to the array of string data in this stack graph.  The resulting array
// pointer is only valid until the next call to any function that mutates the stack graph.
struct sg_strings sg_stack_graph_strings(const struct sg_stack_graph *graph);
// Adds new strings to the stack graph.  You provide all of the string content concatenated
// together into a single string, and an array of the lengths of each string.  You also provide an
// output array, which must have the same size as `lengths`.  We will place each string's handle
// in the output array.
// We ensure that there is only ever one copy of a particular string stored in the graph — we
// guarantee that identical strings will have the same handles, meaning that you can compare the
// handles using simple integer equality.
// We copy the string data into the stack graph.  The string content you pass in does not need to
// outlive the call to this function.
// Each string must be a valid UTF-8 string.  If any string isn't valid UTF-8, it won't be added
// to the stack graph, and the corresponding entry in the output array will be the null handle.
void sg_stack_graph_add_strings(struct sg_stack_graph *graph,
                                size_t count,
                                const char *strings,
                                const size_t *lengths,
                                sg_string_handle *handles_out);
// Returns a reference to the array of file data in this stack graph.  The resulting array pointer
// is only valid until the next call to any function that mutates the stack graph.
struct sg_files sg_stack_graph_files(const struct sg_stack_graph *graph);
// Adds new files to the stack graph.  You provide all of the file content concatenated together
// into a single string, and an array of the lengths of each file.  You also provide an output
// array, which must have the same size as `lengths`.  We will place each file's handle in the
// output array.
// There can only ever be one file with a particular name in the graph.  If you try to add a file
// with a name that already exists, you'll get the same handle as a result.
// We copy the filenames into the stack graph.  The filenames you pass in do not need to outlive
// the call to this function.
// Each filename must be a valid UTF-8 string.  If any filename isn't valid UTF-8, it won't be
// added to the stack graph, and the corresponding entry in the output array will be the null
void sg_stack_graph_add_files(struct sg_stack_graph *graph,
                              size_t count,
                              const char *files,
                              const size_t *lengths,
                              sg_file_handle *handles_out);
// Returns a reference to the array of nodes in this stack graph.  The resulting array pointer is
// only valid until the next call to any function that mutates the stack graph.
struct sg_nodes sg_stack_graph_nodes(const struct sg_stack_graph *graph);
// Adds new nodes to the stack graph.  You provide an array of `struct sg_node` instances.  You
// also provide an output array, which must have the same length as `nodes`, in which we will
// place each node's handle in the stack graph.
// We copy the node content into the stack graph.  The array you pass in does not need to outlive
// the call to this function.
// You cannot add new instances of the root node or "jump to scope" node, since those are
// singletons and already exist in the stack graph.
// If you try to add a new node that has the same ID as an existing node in the stack graph, the
// new node will be ignored, and the corresponding entry in the `handles_out` array will contain
// the handle of the _existing_ node with that ID.
// If any node that you pass in is invalid, it will not be added to the graph, and the
// corresponding entry in the `handles_out` array will be null.
void sg_stack_graph_get_or_create_nodes(struct sg_stack_graph *graph,
                                        size_t count,
                                        const struct sg_node *nodes,
                                        sg_node_handle *handles_out);
// Adds new edges to the stack graph.  You provide an array of `struct sg_edges` instances.  A
// stack graph can contain at most one edge between any two nodes.  It is not an error if you try
// to add an edge that already exists, but it won't have any effect on the graph.
void sg_stack_graph_add_edges(struct sg_stack_graph *graph,
                              size_t count,
                              const struct sg_edge *edges);
// Returns a reference to the array of source information in this stack graph.  The resulting
// array pointer is only valid until the next call to any function that mutates the stack graph.
struct sg_source_infos sg_stack_graph_source_infos(const struct sg_stack_graph *graph);
// Adds new source information to the stack graph.  You provide an array of `sg_node_source_info`
// instances.  Any existing source information for any node mentioned in the array is overwritten.
void sg_stack_graph_add_source_infos(struct sg_stack_graph *graph,
                                     size_t count,
                                     const struct sg_node_source_info *infos);
// Returns a reference to the array of partial symbol stack content in a partial path arena.  The
// resulting array pointer is only valid until the next call to any function that mutates the path
struct sg_partial_symbol_stack_cells sg_partial_path_arena_partial_symbol_stack_cells(const struct sg_partial_path_arena *partials);
// Adds new partial symbol stacks to the partial path arena.  `count` is the number of partial
// symbol stacks you want to create.  The content of each partial symbol stack comes from two
// arrays.  The `lengths` array must have `count` elements, and provides the number of symbols in
// each partial symbol stack.  The `symbols` array contains the contents of each of these partial
// symbol stacks in one contiguous array.  Its length must be the sum of all of the counts in the
// `lengths` array.  The `variables` array must have `count` elements, and provides the optional
// symbol stack variable for each partial symbol stack.
// You must also provide an `out` array, which must also have room for `count` elements.  We will
// fill this array in with the `sg_partial_symbol_stack` instances for each partial symbol stack
// that is created.
void sg_partial_path_arena_add_partial_symbol_stacks(struct sg_partial_path_arena *partials,
                                                     const struct sg_partial_scoped_symbol *symbols,
                                                     const size_t *lengths,
                                                     const sg_symbol_stack_variable *variables,
                                                     struct sg_partial_symbol_stack *out);
// Returns a reference to the array of partial scope stack content in a partial path arena.  The
// resulting array pointer is only valid until the next call to any function that mutates the
// partial path arena.
struct sg_partial_scope_stack_cells sg_partial_path_arena_partial_scope_stack_cells(const struct sg_partial_path_arena *partials);
// Adds new partial scope stacks to the partial path arena.  `count` is the number of partial
// scope stacks you want to create.  The content of each partial scope stack comes from three
// arrays.  The `lengths` array must have `count` elements, and provides the number of scopes in
// each scope stack.  The `scopes` array contains the contents of each of these scope stacks in
// one contiguous array.  Its length must be the sum of all of the counts in the `lengths` array.
// The `variables` array must have `count` elements, and provides the optional scope stack
// variable for each partial scope stack.
// You must also provide an `out` array, which must also have room for `count` elements.  We will
// fill this array in with the `sg_partial_scope_stack` instances for each partial scope stack
// that is created.
void sg_partial_path_arena_add_partial_scope_stacks(struct sg_partial_path_arena *partials,
                                                    const sg_node_handle *scopes,
                                                    const size_t *lengths,
                                                    const sg_scope_stack_variable *variables,
                                                    struct sg_partial_scope_stack *out);
// Returns a reference to the array of partial path edge list content in a partial path arena.
// The resulting array pointer is only valid until the next call to any function that mutates the
// partial path arena.
struct sg_partial_path_edge_list_cells sg_partial_path_arena_partial_path_edge_list_cells(const struct sg_partial_path_arena *partials);
// Adds new partial path edge lists to the partial path arena.  `count` is the number of partial
// path edge lists you want to create.  The content of each partial path edge list comes from two
// arrays.  The `lengths` array must have `count` elements, and provides the number of edges in
// each partial path edge list.  The `edges` array contains the contents of each of these partial
// path edge lists in one contiguous array.  Its length must be the sum of all of the counts in
// the `lengths` array.
// You must also provide an `out` array, which must also have room for `count` elements.  We will
// fill this array in with the `sg_partial_path_edge_list` instances for each partial path edge
// list that is created.
void sg_partial_path_arena_add_partial_path_edge_lists(struct sg_partial_path_arena *partials,
                                                       const struct sg_partial_path_edge *edges,
                                                       const size_t *lengths,
                                                       struct sg_partial_path_edge_list *out);
// Creates a new, empty sg_partial_path_list.
struct sg_partial_path_list *sg_partial_path_list_new(void);
void sg_partial_path_list_free(struct sg_partial_path_list *partial_path_list);
size_t sg_partial_path_list_count(const struct sg_partial_path_list *partial_path_list);
const struct sg_partial_path *sg_partial_path_list_paths(const struct sg_partial_path_list *partial_path_list);
// Finds all partial paths in a file that are _productive_ and _as complete as possible_, placing
// the result into the `partial_path_list` output parameter.  You must free the path list when you
// are done with it by calling `sg_partial_path_list_done`.
// This function will not return until all reachable paths have been processed, so `graph` must
// already contain a complete stack graph.  If you have a very large stack graph stored in some
// other storage system, and want more control over lazily loading only the necessary pieces, then
// you should use sg_forward_path_stitcher.
enum sg_result sg_partial_path_arena_find_partial_paths_in_file(const struct sg_stack_graph *graph,
                                                                struct sg_partial_path_arena *partials,
                                                                sg_file_handle file,
                                                                struct sg_partial_path_list *partial_path_list,
                                                                const size_t *cancellation_flag);
// Finds all complete paths reachable from a set of starting nodes, placing the result into the
// `path_list` output parameter.  You must free the path list when you are done with it by calling
// `sg_path_list_done`.
// This function will not return until all reachable paths have been processed, so `graph` must
// already contain a complete stack graph.  If you have a very large stack graph stored in some
// other storage system, and want more control over lazily loading only the necessary pieces, then
// you should use sg_forward_path_stitcher.
enum sg_result sg_partial_path_arena_find_all_complete_paths(const struct sg_stack_graph *graph,
                                                             struct sg_partial_path_arena *partials,
                                                             size_t starting_node_count,
                                                             const sg_node_handle *starting_nodes,
                                                             struct sg_partial_path_list *path_list,
                                                             const size_t *cancellation_flag);
// Returns a reference to the array of partial path data in this partial path database.  The
// resulting array pointer is only valid until the next call to any function that mutates the
// partial path database.
struct sg_partial_paths sg_partial_path_database_partial_paths(const struct sg_partial_path_database *db);
// Adds new partial paths to the partial path database.  `paths` is the array of partial paths
// that you want to add; `count` is the number of them.
// We copy the partial path content into the partial path database.  The array you pass in does
// not need to outlive the call to this function.
// You should take care not to add a partial path to the database multiple times.  This won't
// cause an _error_, in that nothing will break, but it will probably cause you to get duplicate
// paths from the path-stitching algorithm.
// You must also provide an `out` array, which must also have room for `count` elements.  We will
// fill this array in with the `sg_partial_path_edge_list` instances for each partial path edge
// list that is created.
void sg_partial_path_database_add_partial_paths(const struct sg_stack_graph *graph,
                                                struct sg_partial_path_arena *partials,
                                                struct sg_partial_path_database *db,
                                                const struct sg_partial_path *paths,
                                                sg_partial_path_handle *out);
// Determines which nodes in the stack graph are “local”, taking into account the partial paths in
// this database.  The result is valid until the next call to this function, or until the database
// is freed.
// A local node has no partial path that connects it to the root node in either direction. That
// means that it cannot participate in any paths that leave the file.
// This method is meant to be used at index time, to calculate the set of nodes that are local
// after having just calculated the set of partial paths for the file.
void sg_partial_path_database_find_local_nodes(struct sg_partial_path_database *db);
// Marks that a list of stack graph nodes are local.
// This method is meant to be used at query time.  You will have precalculated the set of local
// nodes for a file at index time; at query time, you will load this information from your storage
// layer and use this method to update our internal view of which nodes are local.
void sg_partial_path_database_mark_local_nodes(struct sg_partial_path_database *db,
                                               const sg_node_handle *nodes);
// Returns a reference to the set of stack graph nodes that are local, according to this database
// of partial paths.  The resulting set is only valid until the next call to any function that
// mutates the partial path database.
struct sg_node_handle_set sg_partial_path_database_local_nodes(const struct sg_partial_path_database *db);
// Creates a new forward partial path stitcher that is "seeded" with a set of starting stack graph
// Before calling this method, you must ensure that `db` contains all of the possible partial
// paths that start with any of your requested starting nodes.
// Before calling `sg_forward_partial_path_stitcher_process_next_phase` for the first time, you
// must ensure that `db` contains all possible extensions of any of those initial partial paths.
// You can retrieve a list of those extensions via the `previous_phase_partial_paths` and
// `previous_phase_partial_paths_length` fields.
struct sg_forward_partial_path_stitcher *sg_forward_partial_path_stitcher_from_nodes(const struct sg_stack_graph *graph,
                                                                                     struct sg_partial_path_arena *partials,
                                                                                     struct sg_partial_path_database *db,
                                                                                     const sg_node_handle *starting_nodes);
// Creates a new forward partial path stitcher that is "seeded" with a set of initial partial
// Before calling `sg_forward_partial_path_stitcher_process_next_phase` for the first time, you
// must ensure that `db` contains all possible extensions of any of those initial partial paths.
// You can retrieve a list of those extensions via the `previous_phase_partial_paths` and
// `previous_phase_partial_paths_length` fields.
struct sg_forward_partial_path_stitcher *sg_forward_partial_path_stitcher_from_partial_paths(const struct sg_stack_graph *graph,
                                                                                             struct sg_partial_path_arena *partials,
                                                                                             struct sg_partial_path_database *db,
                                                                                             const struct sg_partial_path *initial_partial_paths);
// Sets whether similar path detection should be enabled during path stitching. Paths are similar
// if start and end node, and pre- and postconditions are the same. The presence of similar paths
// can lead to exponential blow up during path stitching. Similar path detection is disabled by
// default because of the accociated preformance cost.
void sg_forward_partial_path_stitcher_set_similar_path_detection(struct sg_forward_partial_path_stitcher *stitcher,
                                                                 bool detect_similar_paths);
// Sets the maximum amount of work that can be performed during each phase of the algorithm. By
// bounding our work this way, you can ensure that it's not possible for our CPU-bound algorithm
// to starve any worker threads or processes that you might be using.  If you don't call this
// method, then we allow ourselves to process all of the extensions of all of the paths found in
// the previous phase, with no additional bound.
void sg_forward_partial_path_stitcher_set_max_work_per_phase(struct sg_forward_partial_path_stitcher *stitcher,
                                                             size_t max_work);
// Runs the next phase of the algorithm.  We will have built up a set of incomplete partial paths
// during the _previous_ phase.  Before calling this function, you must ensure that `db` contains
// all of the possible partial paths that we might want to extend any of those candidate partial
// paths with.
// After this method returns, you can retrieve a list of the (possibly incomplete) partial paths
// that were encountered during this phase via the `previous_phase_partial_paths` and
// `previous_phase_partial_paths_length` fields.
void sg_forward_partial_path_stitcher_process_next_phase(const struct sg_stack_graph *graph,
                                                         struct sg_partial_path_arena *partials,
                                                         struct sg_partial_path_database *db,
                                                         struct sg_forward_partial_path_stitcher *stitcher);
// Frees a forward path stitcher.
void sg_forward_partial_path_stitcher_free(struct sg_forward_partial_path_stitcher *stitcher);
#ifdef __cplusplus
} // extern "C"
#endif // __cplusplus
#endif /* STACK_GRAPHS_H_ */
Provide feedback

Saved searches

Use saved searches to filter your results more quickly

FilesExpand file tree

stack-graphs.h

Latest commit

History

stack-graphs.h

File metadata and controls
