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

data/include/prism/prettyprint.h

@@ -0,0 +1,26 @@
+/**
+ * @file prettyprint.h
+ *
+ * An AST node pretty-printer.
+ */
+#ifndef PRISM_PRETTYPRINT_H
+#define PRISM_PRETTYPRINT_H
+
+#include "prism/defines.h"
+
+#include <stdio.h>
+
+#include "prism/ast.h"
+#include "prism/parser.h"
+#include "prism/util/pm_buffer.h"
+
+/**
+ * Pretty-prints the AST represented by the given node to the given buffer.
+ *
+ * @param output_buffer The buffer to write the pretty-printed AST to.
+ * @param parser The parser that parsed the AST.
+ * @param node The root node of the AST to pretty-print.
+ */
+PRISM_EXPORTED_FUNCTION void pm_prettyprint(pm_buffer_t *output_buffer, const pm_parser_t *parser, const pm_node_t *node);
+
+#endif

data/include/prism/regexp.h

@@ -0,0 +1,33 @@
+/**
+ * @file regexp.h
+ *
+ * A regular expression parser.
+ */
+#ifndef PRISM_REGEXP_H
+#define PRISM_REGEXP_H
+
+#include "prism/defines.h"
+#include "prism/parser.h"
+#include "prism/encoding.h"
+#include "prism/util/pm_memchr.h"
+#include "prism/util/pm_string_list.h"
+#include "prism/util/pm_string.h"
+
+#include <stdbool.h>
+#include <stddef.h>
+#include <string.h>
+
+/**
+ * Parse a regular expression and extract the names of all of the named capture
+ * groups.
+ *
+ * @param source The source code to parse.
+ * @param size The size of the source code.
+ * @param named_captures The list to add the names of the named capture groups.
+ * @param encoding_changed Whether or not the encoding changed from the default.
+ * @param encoding The encoding of the source code.
+ * @return Whether or not the parsing was successful.
+ */
+PRISM_EXPORTED_FUNCTION bool pm_regexp_named_capture_group_names(const uint8_t *source, size_t size, pm_string_list_t *named_captures, bool encoding_changed, const pm_encoding_t *encoding);
+
+#endif

data/include/prism/util/pm_buffer.h

@@ -0,0 +1,155 @@
+/**
+ * @file pm_buffer.h
+ *
+ * A wrapper around a contiguous block of allocated memory.
+ */
+#ifndef PRISM_BUFFER_H
+#define PRISM_BUFFER_H
+
+#include "prism/defines.h"
+
+#include <assert.h>
+#include <stdbool.h>
+#include <stdint.h>
+#include <stdlib.h>
+#include <string.h>
+
+/**
+ * A pm_buffer_t is a simple memory buffer that stores data in a contiguous
+ * block of memory.
+ */
+typedef struct {
+    /** The length of the buffer in bytes. */
+    size_t length;
+
+    /** The capacity of the buffer in bytes that has been allocated. */
+    size_t capacity;
+
+    /** A pointer to the start of the buffer. */
+    char *value;
+} pm_buffer_t;
+
+/**
+ * Return the size of the pm_buffer_t struct.
+ *
+ * @returns The size of the pm_buffer_t struct.
+ */
+PRISM_EXPORTED_FUNCTION size_t pm_buffer_sizeof(void);
+
+/**
+ * Initialize a pm_buffer_t with the given capacity.
+ *
+ * @param buffer The buffer to initialize.
+ * @param capacity The capacity of the buffer.
+ * @returns True if the buffer was initialized successfully, false otherwise.
+ */
+bool pm_buffer_init_capacity(pm_buffer_t *buffer, size_t capacity);
+
+/**
+ * Initialize a pm_buffer_t with its default values.
+ *
+ * @param buffer The buffer to initialize.
+ * @returns True if the buffer was initialized successfully, false otherwise.
+ */
+PRISM_EXPORTED_FUNCTION bool pm_buffer_init(pm_buffer_t *buffer);
+
+/**
+ * Return the value of the buffer.
+ *
+ * @param buffer The buffer to get the value of.
+ * @returns The value of the buffer.
+ */
+PRISM_EXPORTED_FUNCTION char * pm_buffer_value(pm_buffer_t *buffer);
+
+/**
+ * Return the length of the buffer.
+ *
+ * @param buffer The buffer to get the length of.
+ * @returns The length of the buffer.
+ */
+PRISM_EXPORTED_FUNCTION size_t pm_buffer_length(pm_buffer_t *buffer);
+
+/**
+ * Append the given amount of space as zeroes to the buffer.
+ *
+ * @param buffer The buffer to append to.
+ * @param length The amount of space to append and zero.
+ */
+void pm_buffer_append_zeroes(pm_buffer_t *buffer, size_t length);
+
+/**
+ * Append a formatted string to the buffer.
+ *
+ * @param buffer The buffer to append to.
+ * @param format The format string to append.
+ * @param ... The arguments to the format string.
+ */
+void pm_buffer_append_format(pm_buffer_t *buffer, const char *format, ...) PRISM_ATTRIBUTE_FORMAT(2, 3);
+
+/**
+ * Append a string to the buffer.
+ *
+ * @param buffer The buffer to append to.
+ * @param value The string to append.
+ * @param length The length of the string to append.
+ */
+void pm_buffer_append_string(pm_buffer_t *buffer, const char *value, size_t length);
+
+/**
+ * Append a list of bytes to the buffer.
+ *
+ * @param buffer The buffer to append to.
+ * @param value The bytes to append.
+ * @param length The length of the bytes to append.
+ */
+void pm_buffer_append_bytes(pm_buffer_t *buffer, const uint8_t *value, size_t length);
+
+/**
+ * Append a single byte to the buffer.
+ *
+ * @param buffer The buffer to append to.
+ * @param value The byte to append.
+ */
+void pm_buffer_append_byte(pm_buffer_t *buffer, uint8_t value);
+
+/**
+ * Append a 32-bit unsigned integer to the buffer as a variable-length integer.
+ *
+ * @param buffer The buffer to append to.
+ * @param value The integer to append.
+ */
+void pm_buffer_append_varuint(pm_buffer_t *buffer, uint32_t value);
+
+/**
+ * Append a 32-bit signed integer to the buffer as a variable-length integer.
+ *
+ * @param buffer The buffer to append to.
+ * @param value The integer to append.
+ */
+void pm_buffer_append_varsint(pm_buffer_t *buffer, int32_t value);
+
+/**
+ * Prepend the given string to the buffer.
+ *
+ * @param buffer The buffer to prepend to.
+ * @param value The string to prepend.
+ * @param length The length of the string to prepend.
+ */
+void pm_buffer_prepend_string(pm_buffer_t *buffer, const char *value, size_t length);
+
+/**
+ * Concatenate one buffer onto another.
+ *
+ * @param destination The buffer to concatenate onto.
+ * @param source The buffer to concatenate.
+ */
+void pm_buffer_concat(pm_buffer_t *destination, const pm_buffer_t *source);
+
+/**
+ * Free the memory associated with the buffer.
+ *
+ * @param buffer The buffer to free.
+ */
+PRISM_EXPORTED_FUNCTION void pm_buffer_free(pm_buffer_t *buffer);
+
+#endif

data/include/prism/util/pm_char.h

@@ -0,0 +1,205 @@
+/**
+ * @file pm_char.h
+ *
+ * Functions for working with characters and strings.
+ */
+#ifndef PRISM_CHAR_H
+#define PRISM_CHAR_H
+
+#include "prism/defines.h"
+#include "prism/util/pm_newline_list.h"
+
+#include <stdbool.h>
+#include <stddef.h>
+
+/**
+ * Returns the number of characters at the start of the string that are
+ * whitespace. Disallows searching past the given maximum number of characters.
+ *
+ * @param string The string to search.
+ * @param length The maximum number of characters to search.
+ * @return The number of characters at the start of the string that are
+ *     whitespace.
+ */
+size_t pm_strspn_whitespace(const uint8_t *string, ptrdiff_t length);
+
+/**
+ * Returns the number of characters at the start of the string that are
+ * whitespace while also tracking the location of each newline. Disallows
+ * searching past the given maximum number of characters.
+ *
+ * @param string The string to search.
+ * @param length The maximum number of characters to search.
+ * @param newline_list The list of newlines to populate.
+ * @return The number of characters at the start of the string that are
+ *     whitespace.
+ */
+size_t
+pm_strspn_whitespace_newlines(const uint8_t *string, ptrdiff_t length, pm_newline_list_t *newline_list);
+
+/**
+ * Returns the number of characters at the start of the string that are inline
+ * whitespace. Disallows searching past the given maximum number of characters.
+ *
+ * @param string The string to search.
+ * @param length The maximum number of characters to search.
+ * @return The number of characters at the start of the string that are inline
+ *     whitespace.
+ */
+size_t pm_strspn_inline_whitespace(const uint8_t *string, ptrdiff_t length);
+
+/**
+ * Returns the number of characters at the start of the string that are decimal
+ * digits. Disallows searching past the given maximum number of characters.
+ *
+ * @param string The string to search.
+ * @param length The maximum number of characters to search.
+ * @return The number of characters at the start of the string that are decimal
+ *     digits.
+ */
+size_t pm_strspn_decimal_digit(const uint8_t *string, ptrdiff_t length);
+
+/**
+ * Returns the number of characters at the start of the string that are
+ * hexadecimal digits. Disallows searching past the given maximum number of
+ * characters.
+ *
+ * @param string The string to search.
+ * @param length The maximum number of characters to search.
+ * @return The number of characters at the start of the string that are
+ *     hexadecimal digits.
+ */
+size_t pm_strspn_hexadecimal_digit(const uint8_t *string, ptrdiff_t length);
+
+/**
+ * Returns the number of characters at the start of the string that are octal
+ * digits or underscores. Disallows searching past the given maximum number of
+ * characters.
+ *
+ * If multiple underscores are found in a row or if an underscore is
+ * found at the end of the number, then the invalid pointer is set to the index
+ * of the first invalid underscore.
+ *
+ * @param string The string to search.
+ * @param length The maximum number of characters to search.
+ * @param invalid The pointer to set to the index of the first invalid
+ *     underscore.
+ * @return The number of characters at the start of the string that are octal
+ *     digits or underscores.
+ */
+size_t pm_strspn_octal_number(const uint8_t *string, ptrdiff_t length, const uint8_t **invalid);
+
+/**
+ * Returns the number of characters at the start of the string that are decimal
+ * digits or underscores. Disallows searching past the given maximum number of
+ * characters.
+ *
+ * If multiple underscores are found in a row or if an underscore is
+ * found at the end of the number, then the invalid pointer is set to the index
+ * of the first invalid underscore.
+ *
+ * @param string The string to search.
+ * @param length The maximum number of characters to search.
+ * @param invalid The pointer to set to the index of the first invalid
+ *     underscore.
+ * @return The number of characters at the start of the string that are decimal
+ *     digits or underscores.
+ */
+size_t pm_strspn_decimal_number(const uint8_t *string, ptrdiff_t length, const uint8_t **invalid);
+
+/**
+ * Returns the number of characters at the start of the string that are
+ * hexadecimal digits or underscores. Disallows searching past the given maximum
+ * number of characters.
+ *
+ * If multiple underscores are found in a row or if an underscore is
+ * found at the end of the number, then the invalid pointer is set to the index
+ * of the first invalid underscore.
+ *
+ * @param string The string to search.
+ * @param length The maximum number of characters to search.
+ * @param invalid The pointer to set to the index of the first invalid
+ *     underscore.
+ * @return The number of characters at the start of the string that are
+ *     hexadecimal digits or underscores.
+ */
+size_t pm_strspn_hexadecimal_number(const uint8_t *string, ptrdiff_t length, const uint8_t **invalid);
+
+/**
+ * Returns the number of characters at the start of the string that are regexp
+ * options. Disallows searching past the given maximum number of characters.
+ *
+ * @param string The string to search.
+ * @param length The maximum number of characters to search.
+ * @return The number of characters at the start of the string that are regexp
+ *     options.
+ */
+size_t pm_strspn_regexp_option(const uint8_t *string, ptrdiff_t length);
+
+/**
+ * Returns the number of characters at the start of the string that are binary
+ * digits or underscores. Disallows searching past the given maximum number of
+ * characters.
+ *
+ * If multiple underscores are found in a row or if an underscore is
+ * found at the end of the number, then the invalid pointer is set to the index
+ * of the first invalid underscore.
+ *
+ * @param string The string to search.
+ * @param length The maximum number of characters to search.
+ * @param invalid The pointer to set to the index of the first invalid
+ *     underscore.
+ * @return The number of characters at the start of the string that are binary
+ *     digits or underscores.
+ */
+size_t pm_strspn_binary_number(const uint8_t *string, ptrdiff_t length, const uint8_t **invalid);
+
+/**
+ * Returns true if the given character is a whitespace character.
+ *
+ * @param b The character to check.
+ * @return True if the given character is a whitespace character.
+ */
+bool pm_char_is_whitespace(const uint8_t b);
+
+/**
+ * Returns true if the given character is an inline whitespace character.
+ *
+ * @param b The character to check.
+ * @return True if the given character is an inline whitespace character.
+ */
+bool pm_char_is_inline_whitespace(const uint8_t b);
+
+/**
+ * Returns true if the given character is a binary digit.
+ *
+ * @param b The character to check.
+ * @return True if the given character is a binary digit.
+ */
+bool pm_char_is_binary_digit(const uint8_t b);
+
+/**
+ * Returns true if the given character is an octal digit.
+ *
+ * @param b The character to check.
+ * @return True if the given character is an octal digit.
+ */
+bool pm_char_is_octal_digit(const uint8_t b);
+
+/**
+ * Returns true if the given character is a decimal digit.
+ *
+ * @param b The character to check.
+ * @return True if the given character is a decimal digit.
+ */
+bool pm_char_is_decimal_digit(const uint8_t b);
+
+/**
+ * Returns true if the given character is a hexadecimal digit.
+ *
+ * @param b The character to check.
+ * @return True if the given character is a hexadecimal digit.
+ */
+bool pm_char_is_hexadecimal_digit(const uint8_t b);
+
+#endif

data/include/prism/util/pm_constant_pool.h

@@ -0,0 +1,209 @@
+/**
+ * @file pm_constant_pool.h
+ *
+ * A data structure that stores a set of strings.
+ *
+ * Each string is assigned a unique id, which can be used to compare strings for
+ * equality. This comparison ends up being much faster than strcmp, since it
+ * only requires a single integer comparison.
+ */
+#ifndef PRISM_CONSTANT_POOL_H
+#define PRISM_CONSTANT_POOL_H
+
+#include "prism/defines.h"
+
+#include <assert.h>
+#include <stdbool.h>
+#include <stdint.h>
+#include <stdlib.h>
+#include <string.h>
+
+/**
+ * When we allocate constants into the pool, we reserve 0 to mean that the slot
+ * is not yet filled. This constant is reused in other places to indicate the
+ * lack of a constant id.
+ */
+#define PM_CONSTANT_ID_UNSET 0
+
+/**
+ * A constant id is a unique identifier for a constant in the constant pool.
+ */
+typedef uint32_t pm_constant_id_t;
+
+/**
+ * A list of constant IDs. Usually used to represent a set of locals.
+ */
+typedef struct {
+    /** The number of constant ids in the list. */
+    size_t size;
+
+    /** The number of constant ids that have been allocated in the list. */
+    size_t capacity;
+
+    /** The constant ids in the list. */
+    pm_constant_id_t *ids;
+} pm_constant_id_list_t;
+
+/**
+ * Initialize a list of constant ids.
+ *
+ * @param list The list to initialize.
+ */
+void pm_constant_id_list_init(pm_constant_id_list_t *list);
+
+/**
+ * Append a constant id to a list of constant ids. Returns false if any
+ * potential reallocations fail.
+ *
+ * @param list The list to append to.
+ * @param id The id to append.
+ * @return Whether the append succeeded.
+ */
+bool pm_constant_id_list_append(pm_constant_id_list_t *list, pm_constant_id_t id);
+
+/**
+ * Checks if the current constant id list includes the given constant id.
+ *
+ * @param list The list to check.
+ * @param id The id to check for.
+ * @return Whether the list includes the given id.
+ */
+bool pm_constant_id_list_includes(pm_constant_id_list_t *list, pm_constant_id_t id);
+
+/**
+ * Get the memory size of a list of constant ids.
+ *
+ * @param list The list to get the memory size of.
+ * @return The memory size of the list.
+ */
+size_t pm_constant_id_list_memsize(pm_constant_id_list_t *list);
+
+/**
+ * Free the memory associated with a list of constant ids.
+ *
+ * @param list The list to free.
+ */
+void pm_constant_id_list_free(pm_constant_id_list_t *list);
+
+/**
+ * The type of bucket in the constant pool hash map. This determines how the
+ * bucket should be freed.
+ */
+typedef unsigned int pm_constant_pool_bucket_type_t;
+
+/** By default, each constant is a slice of the source. */
+static const pm_constant_pool_bucket_type_t PM_CONSTANT_POOL_BUCKET_DEFAULT = 0;
+
+/** An owned constant is one for which memory has been allocated. */
+static const pm_constant_pool_bucket_type_t PM_CONSTANT_POOL_BUCKET_OWNED = 1;
+
+/** A constant constant is known at compile time. */
+static const pm_constant_pool_bucket_type_t PM_CONSTANT_POOL_BUCKET_CONSTANT = 2;
+
+/** A bucket in the hash map. */
+typedef struct {
+    /** The incremental ID used for indexing back into the pool. */
+    unsigned int id: 30;
+
+    /** The type of the bucket, which determines how to free it. */
+    pm_constant_pool_bucket_type_t type: 2;
+
+    /** The hash of the bucket. */
+    uint32_t hash;
+} pm_constant_pool_bucket_t;
+
+/** A constant in the pool which effectively stores a string. */
+typedef struct {
+    /** A pointer to the start of the string. */
+    const uint8_t *start;
+
+    /** The length of the string. */
+    size_t length;
+} pm_constant_t;
+
+/** The overall constant pool, which stores constants found while parsing. */
+typedef struct {
+    /** The buckets in the hash map. */
+    pm_constant_pool_bucket_t *buckets;
+
+    /** The constants that are stored in the buckets. */
+    pm_constant_t *constants;
+
+    /** The number of buckets in the hash map. */
+    uint32_t size;
+
+    /** The number of buckets that have been allocated in the hash map. */
+    uint32_t capacity;
+} pm_constant_pool_t;
+
+/**
+ * Initialize a new constant pool with a given capacity.
+ *
+ * @param pool The pool to initialize.
+ * @param capacity The initial capacity of the pool.
+ * @return Whether the initialization succeeded.
+ */
+bool pm_constant_pool_init(pm_constant_pool_t *pool, uint32_t capacity);
+
+/**
+ * Return a pointer to the constant indicated by the given constant id.
+ *
+ * @param pool The pool to get the constant from.
+ * @param constant_id The id of the constant to get.
+ * @return A pointer to the constant.
+ */
+pm_constant_t * pm_constant_pool_id_to_constant(const pm_constant_pool_t *pool, pm_constant_id_t constant_id);
+
+/**
+ * Find a constant in a constant pool. Returns the id of the constant, or 0 if
+ * the constant is not found.
+ *
+ * @param pool The pool to find the constant in.
+ * @param start A pointer to the start of the constant.
+ * @param length The length of the constant.
+ * @return The id of the constant.
+ */
+pm_constant_id_t pm_constant_pool_find(const pm_constant_pool_t *pool, const uint8_t *start, size_t length);
+
+/**
+ * Insert a constant into a constant pool that is a slice of a source string.
+ * Returns the id of the constant, or 0 if any potential calls to resize fail.
+ *
+ * @param pool The pool to insert the constant into.
+ * @param start A pointer to the start of the constant.
+ * @param length The length of the constant.
+ * @return The id of the constant.
+ */
+pm_constant_id_t pm_constant_pool_insert_shared(pm_constant_pool_t *pool, const uint8_t *start, size_t length);
+
+/**
+ * Insert a constant into a constant pool from memory that is now owned by the
+ * constant pool. Returns the id of the constant, or 0 if any potential calls to
+ * resize fail.
+ *
+ * @param pool The pool to insert the constant into.
+ * @param start A pointer to the start of the constant.
+ * @param length The length of the constant.
+ * @return The id of the constant.
+ */
+pm_constant_id_t pm_constant_pool_insert_owned(pm_constant_pool_t *pool, const uint8_t *start, size_t length);
+
+/**
+ * Insert a constant into a constant pool from memory that is constant. Returns
+ * the id of the constant, or 0 if any potential calls to resize fail.
+ *
+ * @param pool The pool to insert the constant into.
+ * @param start A pointer to the start of the constant.
+ * @param length The length of the constant.
+ * @return The id of the constant.
+ */
+pm_constant_id_t pm_constant_pool_insert_constant(pm_constant_pool_t *pool, const uint8_t *start, size_t length);
+
+/**
+ * Free the memory associated with a constant pool.
+ *
+ * @param pool The pool to free.
+ */
+void pm_constant_pool_free(pm_constant_pool_t *pool);
+
+#endif