jruby-prism-parser 0.23.0.pre.SNAPSHOT-java

data/include/prism/node.h

@@ -0,0 +1,57 @@
+/**
+ * @file node.h
+ *
+ * Functions related to nodes in the AST.
+ */
+#ifndef PRISM_NODE_H
+#define PRISM_NODE_H
+
+#include "prism/defines.h"
+#include "prism/parser.h"
+
+/**
+ * Append a new node onto the end of the node list.
+ *
+ * @param list The list to append to.
+ * @param node The node to append.
+ */
+void pm_node_list_append(pm_node_list_t *list, pm_node_t *node);
+
+/**
+ * Deallocate a node and all of its children.
+ *
+ * @param parser The parser that owns the node.
+ * @param node The node to deallocate.
+ */
+PRISM_EXPORTED_FUNCTION void pm_node_destroy(pm_parser_t *parser, struct pm_node *node);
+
+/**
+ * This struct stores the information gathered by the pm_node_memsize function.
+ * It contains both the memory footprint and additionally metadata about the
+ * shape of the tree.
+ */
+typedef struct {
+    /** The total memory footprint of the node and all of its children. */
+    size_t memsize;
+
+    /** The number of children the node has. */
+    size_t node_count;
+} pm_memsize_t;
+
+/**
+ * Calculates the memory footprint of a given node.
+ *
+ * @param node The node to calculate the memory footprint of.
+ * @param memsize The memory footprint of the node and all of its children.
+ */
+PRISM_EXPORTED_FUNCTION void pm_node_memsize(pm_node_t *node, pm_memsize_t *memsize);
+
+/**
+ * Returns a string representation of the given node type.
+ *
+ * @param node_type The node type to convert to a string.
+ * @return A string representation of the given node type.
+ */
+PRISM_EXPORTED_FUNCTION const char * pm_node_type_to_str(pm_node_type_t node_type);
+
+#endif

data/include/prism/options.h

@@ -0,0 +1,230 @@
+/**
+ * @file options.h
+ *
+ * The options that can be passed to parsing.
+ */
+#ifndef PRISM_OPTIONS_H
+#define PRISM_OPTIONS_H
+
+#include "prism/defines.h"
+#include "prism/util/pm_string.h"
+
+#include <stdbool.h>
+#include <stddef.h>
+#include <stdint.h>
+
+/**
+ * A scope of locals surrounding the code that is being parsed.
+ */
+typedef struct pm_options_scope {
+    /** The number of locals in the scope. */
+    size_t locals_count;
+
+    /** The names of the locals in the scope. */
+    pm_string_t *locals;
+} pm_options_scope_t;
+
+/**
+ * The version of Ruby syntax that we should be parsing with. This is used to
+ * allow consumers to specify which behavior they want in case they need to
+ * parse in the same way as a specific version of CRuby would have.
+ */
+typedef enum {
+    /** The current version of prism. */
+    PM_OPTIONS_VERSION_LATEST = 0,
+
+    /** The vendored version of prism in CRuby 3.3.0. */
+    PM_OPTIONS_VERSION_CRUBY_3_3_0 = 1
+} pm_options_version_t;
+
+/**
+ * The options that can be passed to the parser.
+ */
+typedef struct {
+    /** The name of the file that is currently being parsed. */
+    pm_string_t filepath;
+
+    /**
+     * The line within the file that the parse starts on. This value is
+     * 1-indexed.
+     */
+    int32_t line;
+
+    /**
+     * The name of the encoding that the source file is in. Note that this must
+     * correspond to a name that can be found with Encoding.find in Ruby.
+     */
+    pm_string_t encoding;
+
+    /**
+     * The number of scopes surrounding the code that is being parsed.
+     */
+    size_t scopes_count;
+
+    /**
+     * The scopes surrounding the code that is being parsed. For most parses
+     * this will be NULL, but for evals it will be the locals that are in scope
+     * surrounding the eval. Scopes are ordered from the outermost scope to the
+     * innermost one.
+     */
+    pm_options_scope_t *scopes;
+
+    /**
+     * The version of prism that we should be parsing with. This is used to
+     * allow consumers to specify which behavior they want in case they need to
+     * parse exactly as a specific version of CRuby.
+     */
+    pm_options_version_t version;
+
+    /** Whether or not the frozen string literal option has been set. */
+    bool frozen_string_literal;
+} pm_options_t;
+
+/**
+ * Set the filepath option on the given options struct.
+ *
+ * @param options The options struct to set the filepath on.
+ * @param filepath The filepath to set.
+ */
+PRISM_EXPORTED_FUNCTION void pm_options_filepath_set(pm_options_t *options, const char *filepath);
+
+/**
+ * Set the line option on the given options struct.
+ *
+ * @param options The options struct to set the line on.
+ * @param line The line to set.
+ */
+PRISM_EXPORTED_FUNCTION void pm_options_line_set(pm_options_t *options, int32_t line);
+
+/**
+ * Set the encoding option on the given options struct.
+ *
+ * @param options The options struct to set the encoding on.
+ * @param encoding The encoding to set.
+ */
+PRISM_EXPORTED_FUNCTION void pm_options_encoding_set(pm_options_t *options, const char *encoding);
+
+/**
+ * Set the frozen string literal option on the given options struct.
+ *
+ * @param options The options struct to set the frozen string literal value on.
+ * @param frozen_string_literal The frozen string literal value to set.
+ */
+PRISM_EXPORTED_FUNCTION void pm_options_frozen_string_literal_set(pm_options_t *options, bool frozen_string_literal);
+
+/**
+ * Set the version option on the given options struct by parsing the given
+ * string. If the string contains an invalid option, this returns false.
+ * Otherwise, it returns true.
+ *
+ * @param options The options struct to set the version on.
+ * @param version The version to set.
+ * @param length The length of the version string.
+ * @return Whether or not the version was parsed successfully.
+ */
+PRISM_EXPORTED_FUNCTION bool pm_options_version_set(pm_options_t *options, const char *version, size_t length);
+
+/**
+ * Allocate and zero out the scopes array on the given options struct.
+ *
+ * @param options The options struct to initialize the scopes array on.
+ * @param scopes_count The number of scopes to allocate.
+ */
+PRISM_EXPORTED_FUNCTION void pm_options_scopes_init(pm_options_t *options, size_t scopes_count);
+
+/**
+ * Return a pointer to the scope at the given index within the given options.
+ *
+ * @param options The options struct to get the scope from.
+ * @param index The index of the scope to get.
+ * @return A pointer to the scope at the given index.
+ */
+PRISM_EXPORTED_FUNCTION const pm_options_scope_t * pm_options_scope_get(const pm_options_t *options, size_t index);
+
+/**
+ * Create a new options scope struct. This will hold a set of locals that are in
+ * scope surrounding the code that is being parsed.
+ *
+ * @param scope The scope struct to initialize.
+ * @param locals_count The number of locals to allocate.
+ */
+PRISM_EXPORTED_FUNCTION void pm_options_scope_init(pm_options_scope_t *scope, size_t locals_count);
+
+/**
+ * Return a pointer to the local at the given index within the given scope.
+ *
+ * @param scope The scope struct to get the local from.
+ * @param index The index of the local to get.
+ * @return A pointer to the local at the given index.
+ */
+PRISM_EXPORTED_FUNCTION const pm_string_t * pm_options_scope_local_get(const pm_options_scope_t *scope, size_t index);
+
+/**
+ * Free the internal memory associated with the options.
+ *
+ * @param options The options struct whose internal memory should be freed.
+ */
+PRISM_EXPORTED_FUNCTION void pm_options_free(pm_options_t *options);
+
+/**
+ * Deserialize an options struct from the given binary string. This is used to
+ * pass options to the parser from an FFI call so that consumers of the library
+ * from an FFI perspective don't have to worry about the structure of our
+ * options structs. Since the source of these calls will be from Ruby
+ * implementation internals we assume it is from a trusted source.
+ *
+ * `data` is assumed to be a valid pointer pointing to well-formed data. The
+ * layout of this data should be the same every time, and is described below:
+ *
+ * | # bytes | field                      |
+ * | ------- | -------------------------- |
+ * | `4`     | the length of the filepath |
+ * | ...     | the filepath bytes         |
+ * | `4`     | the line number            |
+ * | `4`     | the length the encoding    |
+ * | ...     | the encoding bytes         |
+ * | `1`     | frozen string literal      |
+ * | `1`     | suppress warnings          |
+ * | `1`     | the version                |
+ * | `4`     | the number of scopes       |
+ * | ...     | the scopes                 |
+ *
+ * The version field is an enum, so it should be one of the following values:
+ *
+ * | value | version                   |
+ * | ----- | ------------------------- |
+ * | `0`   | use the latest version of prism |
+ * | `1`   | use the version of prism that is vendored in CRuby 3.3.0 |
+ *
+ * Each scope is layed out as follows:
+ *
+ * | # bytes | field                      |
+ * | ------- | -------------------------- |
+ * | `4`     | the number of locals       |
+ * | ...     | the locals                 |
+ *
+ * Each local is layed out as follows:
+ *
+ * | # bytes | field                      |
+ * | ------- | -------------------------- |
+ * | `4`     | the length of the local    |
+ * | ...     | the local bytes            |
+ *
+ * Some additional things to note about this layout:
+ *
+ * * The filepath can have a length of 0, in which case we'll consider it an
+ *   empty string.
+ * * The line number should be 0-indexed.
+ * * The encoding can have a length of 0, in which case we'll use the default
+ *   encoding (UTF-8). If it's not 0, it should correspond to a name of an
+ *   encoding that can be passed to `Encoding.find` in Ruby.
+ * * The frozen string literal and suppress warnings fields are booleans, so
+ *   their values should be either 0 or 1.
+ * * The number of scopes can be 0.
+ *
+ * @param options The options struct to deserialize into.
+ * @param data The binary string to deserialize from.
+ */
+void pm_options_read(pm_options_t *options, const char *data);
+
+#endif

data/include/prism/pack.h

@@ -0,0 +1,152 @@
+/**
+ * @file pack.h
+ *
+ * A pack template string parser.
+ */
+#ifndef PRISM_PACK_H
+#define PRISM_PACK_H
+
+#include "prism/defines.h"
+
+#include <stdint.h>
+#include <stdlib.h>
+
+/** The version of the pack template language that we are parsing. */
+typedef enum pm_pack_version {
+    PM_PACK_VERSION_3_2_0
+} pm_pack_version;
+
+/** The type of pack template we are parsing. */
+typedef enum pm_pack_variant {
+    PM_PACK_VARIANT_PACK,
+    PM_PACK_VARIANT_UNPACK
+} pm_pack_variant;
+
+/** A directive within the pack template. */
+typedef enum pm_pack_type {
+    PM_PACK_SPACE,
+    PM_PACK_COMMENT,
+    PM_PACK_INTEGER,
+    PM_PACK_UTF8,
+    PM_PACK_BER,
+    PM_PACK_FLOAT,
+    PM_PACK_STRING_SPACE_PADDED,
+    PM_PACK_STRING_NULL_PADDED,
+    PM_PACK_STRING_NULL_TERMINATED,
+    PM_PACK_STRING_MSB,
+    PM_PACK_STRING_LSB,
+    PM_PACK_STRING_HEX_HIGH,
+    PM_PACK_STRING_HEX_LOW,
+    PM_PACK_STRING_UU,
+    PM_PACK_STRING_MIME,
+    PM_PACK_STRING_BASE64,
+    PM_PACK_STRING_FIXED,
+    PM_PACK_STRING_POINTER,
+    PM_PACK_MOVE,
+    PM_PACK_BACK,
+    PM_PACK_NULL,
+    PM_PACK_END
+} pm_pack_type;
+
+/** The signness of a pack directive. */
+typedef enum pm_pack_signed {
+    PM_PACK_UNSIGNED,
+    PM_PACK_SIGNED,
+    PM_PACK_SIGNED_NA
+} pm_pack_signed;
+
+/** The endianness of a pack directive. */
+typedef enum pm_pack_endian {
+    PM_PACK_AGNOSTIC_ENDIAN,
+    PM_PACK_LITTLE_ENDIAN,      // aka 'VAX', or 'V'
+    PM_PACK_BIG_ENDIAN,         // aka 'network', or 'N'
+    PM_PACK_NATIVE_ENDIAN,
+    PM_PACK_ENDIAN_NA
+} pm_pack_endian;
+
+/** The size of an integer pack directive. */
+typedef enum pm_pack_size {
+    PM_PACK_SIZE_SHORT,
+    PM_PACK_SIZE_INT,
+    PM_PACK_SIZE_LONG,
+    PM_PACK_SIZE_LONG_LONG,
+    PM_PACK_SIZE_8,
+    PM_PACK_SIZE_16,
+    PM_PACK_SIZE_32,
+    PM_PACK_SIZE_64,
+    PM_PACK_SIZE_P,
+    PM_PACK_SIZE_NA
+} pm_pack_size;
+
+/** The type of length of a pack directive. */
+typedef enum pm_pack_length_type {
+    PM_PACK_LENGTH_FIXED,
+    PM_PACK_LENGTH_MAX,
+    PM_PACK_LENGTH_RELATIVE,  // special case for unpack @*
+    PM_PACK_LENGTH_NA
+} pm_pack_length_type;
+
+/** The type of encoding for a pack template string. */
+typedef enum pm_pack_encoding {
+    PM_PACK_ENCODING_START,
+    PM_PACK_ENCODING_ASCII_8BIT,
+    PM_PACK_ENCODING_US_ASCII,
+    PM_PACK_ENCODING_UTF_8
+} pm_pack_encoding;
+
+/** The result of parsing a pack template. */
+typedef enum pm_pack_result {
+    PM_PACK_OK,
+    PM_PACK_ERROR_UNSUPPORTED_DIRECTIVE,
+    PM_PACK_ERROR_UNKNOWN_DIRECTIVE,
+    PM_PACK_ERROR_LENGTH_TOO_BIG,
+    PM_PACK_ERROR_BANG_NOT_ALLOWED,
+    PM_PACK_ERROR_DOUBLE_ENDIAN
+} pm_pack_result;
+
+/**
+ * Parse a single directive from a pack or unpack format string.
+ *
+ * @param variant (in) pack or unpack
+ * @param format (in, out) the start of the next directive to parse on calling,
+ *     and advanced beyond the parsed directive on return, or as much of it as
+ *     was consumed until an error was encountered
+ * @param format_end (in) the end of the format string
+ * @param type (out) the type of the directive
+ * @param signed_type (out) whether the value is signed
+ * @param endian (out) the endianness of the value
+ * @param size (out) the size of the value
+ * @param length_type (out) what kind of length is specified
+ * @param length (out) the length of the directive
+ * @param encoding (in, out) takes the current encoding of the string which
+ *     would result from parsing the whole format string, and returns a possibly
+ *     changed directive - the encoding should be `PM_PACK_ENCODING_START` when
+ *     pm_pack_parse is called for the first directive in a format string
+ *
+ * @return `PM_PACK_OK` on success or `PM_PACK_ERROR_*` on error
+ * @note Consult Ruby documentation for the meaning of directives.
+ */
+PRISM_EXPORTED_FUNCTION pm_pack_result
+pm_pack_parse(
+    pm_pack_variant variant,
+    const char **format,
+    const char *format_end,
+    pm_pack_type *type,
+    pm_pack_signed *signed_type,
+    pm_pack_endian *endian,
+    pm_pack_size *size,
+    pm_pack_length_type *length_type,
+    uint64_t *length,
+    pm_pack_encoding *encoding
+);
+
+/**
+ * Prism abstracts sizes away from the native system - this converts an abstract
+ * size to a native size.
+ *
+ * @param size The abstract size to convert.
+ * @return The native size.
+ */
+PRISM_EXPORTED_FUNCTION size_t pm_size_to_native(pm_pack_size size);
+
+#endif