datadog-ruby_core_source 3.4.1 → 3.5.0

data/lib/datadog/ruby_core_source/ruby-4.0.0-preview2/prism/util/pm_char.h

@@ -0,0 +1,204 @@
+/**
+ * @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/lib/datadog/ruby_core_source/ruby-4.0.0-preview2/prism/util/pm_constant_pool.h

@@ -0,0 +1,218 @@
+/**
+ * @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);
+
+/**
+ * Initialize a list of constant ids with a given capacity.
+ *
+ * @param list The list to initialize.
+ * @param capacity The initial capacity of the list.
+ */
+void pm_constant_id_list_init_capacity(pm_constant_id_list_t *list, size_t capacity);
+
+/**
+ * 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);
+
+/**
+ * Insert a constant id into a list of constant ids at the specified index.
+ *
+ * @param list The list to insert into.
+ * @param index The index at which to insert.
+ * @param id The id to insert.
+ */
+void pm_constant_id_list_insert(pm_constant_id_list_t *list, size_t index, 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);
+
+/**
+ * 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, 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

data/lib/datadog/ruby_core_source/ruby-4.0.0-preview2/prism/util/pm_integer.h

@@ -0,0 +1,130 @@
+/**
+ * @file pm_integer.h
+ *
+ * This module provides functions for working with arbitrary-sized integers.
+ */
+#ifndef PRISM_NUMBER_H
+#define PRISM_NUMBER_H
+
+#include "prism/defines.h"
+#include "prism/util/pm_buffer.h"
+
+#include <assert.h>
+#include <stdbool.h>
+#include <stdint.h>
+#include <stdlib.h>
+
+/**
+ * A structure represents an arbitrary-sized integer.
+ */
+typedef struct {
+    /**
+     * The number of allocated values. length is set to 0 if the integer fits
+     * into uint32_t.
+     */
+    size_t length;
+
+    /**
+     * List of 32-bit integers. Set to NULL if the integer fits into uint32_t.
+     */
+    uint32_t *values;
+
+    /**
+     * Embedded value for small integer. This value is set to 0 if the value
+     * does not fit into uint32_t.
+     */
+    uint32_t value;
+
+    /**
+     * Whether or not the integer is negative. It is stored this way so that a
+     * zeroed pm_integer_t is always positive zero.
+     */
+    bool negative;
+} pm_integer_t;
+
+/**
+ * An enum controlling the base of an integer. It is expected that the base is
+ * already known before parsing the integer, even though it could be derived
+ * from the string itself.
+ */
+typedef enum {
+    /** The default decimal base, with no prefix. Leading 0s will be ignored. */
+    PM_INTEGER_BASE_DEFAULT,
+
+    /** The binary base, indicated by a 0b or 0B prefix. */
+    PM_INTEGER_BASE_BINARY,
+
+    /** The octal base, indicated by a 0, 0o, or 0O prefix. */
+    PM_INTEGER_BASE_OCTAL,
+
+    /** The decimal base, indicated by a 0d, 0D, or empty prefix. */
+    PM_INTEGER_BASE_DECIMAL,
+
+    /** The hexadecimal base, indicated by a 0x or 0X prefix. */
+    PM_INTEGER_BASE_HEXADECIMAL,
+
+    /**
+     * An unknown base, in which case pm_integer_parse will derive it based on
+     * the content of the string. This is less efficient and does more
+     * comparisons, so if callers know the base ahead of time, they should use
+     * that instead.
+     */
+    PM_INTEGER_BASE_UNKNOWN
+} pm_integer_base_t;
+
+/**
+ * Parse an integer from a string. This assumes that the format of the integer
+ * has already been validated, as internal validation checks are not performed
+ * here.
+ *
+ * @param integer The integer to parse into.
+ * @param base The base of the integer.
+ * @param start The start of the string.
+ * @param end The end of the string.
+ */
+void pm_integer_parse(pm_integer_t *integer, pm_integer_base_t base, const uint8_t *start, const uint8_t *end);
+
+/**
+ * Compare two integers. This function returns -1 if the left integer is less
+ * than the right integer, 0 if they are equal, and 1 if the left integer is
+ * greater than the right integer.
+ *
+ * @param left The left integer to compare.
+ * @param right The right integer to compare.
+ * @return The result of the comparison.
+ */
+int pm_integer_compare(const pm_integer_t *left, const pm_integer_t *right);
+
+/**
+ * Reduce a ratio of integers to its simplest form.
+ *
+ * If either the numerator or denominator do not fit into a 32-bit integer, then
+ * this function is a no-op. In the future, we may consider reducing even the
+ * larger numbers, but for now we're going to keep it simple.
+ *
+ * @param numerator The numerator of the ratio.
+ * @param denominator The denominator of the ratio.
+ */
+void pm_integers_reduce(pm_integer_t *numerator, pm_integer_t *denominator);
+
+/**
+ * Convert an integer to a decimal string.
+ *
+ * @param buffer The buffer to append the string to.
+ * @param integer The integer to convert to a string.
+ *
+ * \public \memberof pm_integer_t
+ */
+PRISM_EXPORTED_FUNCTION void pm_integer_string(pm_buffer_t *buffer, const pm_integer_t *integer);
+
+/**
+ * Free the internal memory of an integer. This memory will only be allocated if
+ * the integer exceeds the size of a single node in the linked list.
+ *
+ * @param integer The integer to free.
+ *
+ * \public \memberof pm_integer_t
+ */
+PRISM_EXPORTED_FUNCTION void pm_integer_free(pm_integer_t *integer);
+
+#endif

data/lib/datadog/ruby_core_source/ruby-4.0.0-preview2/prism/util/pm_list.h

@@ -0,0 +1,103 @@
+/**
+ * @file pm_list.h
+ *
+ * An abstract linked list.
+ */
+#ifndef PRISM_LIST_H
+#define PRISM_LIST_H
+
+#include "prism/defines.h"
+
+#include <stdbool.h>
+#include <stddef.h>
+#include <stdint.h>
+#include <stdlib.h>
+
+/**
+ * This struct represents an abstract linked list that provides common
+ * functionality. It is meant to be used any time a linked list is necessary to
+ * store data.
+ *
+ * The linked list itself operates off a set of pointers. Because the pointers
+ * are not necessarily sequential, they can be of any size. We use this fact to
+ * allow the consumer of this linked list to extend the node struct to include
+ * any data they want. This is done by using the pm_list_node_t as the first
+ * member of the struct.
+ *
+ * For example, if we want to store a list of integers, we can do the following:
+ *
+ * ```c
+ * typedef struct {
+ *     pm_list_node_t node;
+ *     int value;
+ * } pm_int_node_t;
+ *
+ * pm_list_t list = { 0 };
+ * pm_int_node_t *node = xmalloc(sizeof(pm_int_node_t));
+ * node->value = 5;
+ *
+ * pm_list_append(&list, &node->node);
+ * ```
+ *
+ * The pm_list_t struct is used to represent the overall linked list. It
+ * contains a pointer to the head and tail of the list. This allows for easy
+ * iteration and appending of new nodes.
+ */
+typedef struct pm_list_node {
+    /** A pointer to the next node in the list. */
+    struct pm_list_node *next;
+} pm_list_node_t;
+
+/**
+ * This represents the overall linked list. It keeps a pointer to the head and
+ * tail so that iteration is easy and pushing new nodes is easy.
+ */
+typedef struct {
+    /** The size of the list. */
+    size_t size;
+
+    /** A pointer to the head of the list. */
+    pm_list_node_t *head;
+
+    /** A pointer to the tail of the list. */
+    pm_list_node_t *tail;
+} pm_list_t;
+
+/**
+ * Returns true if the given list is empty.
+ *
+ * @param list The list to check.
+ * @return True if the given list is empty, otherwise false.
+ *
+ * \public \memberof pm_list_t
+ */
+PRISM_EXPORTED_FUNCTION bool pm_list_empty_p(pm_list_t *list);
+
+/**
+ * Returns the size of the list.
+ *
+ * @param list The list to check.
+ * @return The size of the list.
+ *
+ * \public \memberof pm_list_t
+ */
+PRISM_EXPORTED_FUNCTION size_t pm_list_size(pm_list_t *list);
+
+/**
+ * Append a node to the given list.
+ *
+ * @param list The list to append to.
+ * @param node The node to append.
+ */
+void pm_list_append(pm_list_t *list, pm_list_node_t *node);
+
+/**
+ * Deallocate the internal state of the given list.
+ *
+ * @param list The list to free.
+ *
+ * \public \memberof pm_list_t
+ */
+PRISM_EXPORTED_FUNCTION void pm_list_free(pm_list_t *list);
+
+#endif

data/lib/datadog/ruby_core_source/ruby-4.0.0-preview2/prism/util/pm_memchr.h

@@ -0,0 +1,29 @@
+/**
+ * @file pm_memchr.h
+ *
+ * A custom memchr implementation.
+ */
+#ifndef PRISM_MEMCHR_H
+#define PRISM_MEMCHR_H
+
+#include "prism/defines.h"
+#include "prism/encoding.h"
+
+#include <stddef.h>
+
+/**
+ * We need to roll our own memchr to handle cases where the encoding changes and
+ * we need to search for a character in a buffer that could be the trailing byte
+ * of a multibyte character.
+ *
+ * @param source The source string.
+ * @param character The character to search for.
+ * @param number The maximum number of bytes to search.
+ * @param encoding_changed Whether the encoding changed.
+ * @param encoding A pointer to the encoding.
+ * @return A pointer to the first occurrence of the character in the source
+ *     string, or NULL if no such character exists.
+ */
+void * pm_memchr(const void *source, int character, size_t number, bool encoding_changed, const pm_encoding_t *encoding);
+
+#endif