YourWishes/jerryscript
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Refactor RegExp builtin (#3136)

Browse Source 
JerryScript-DCO-1.0-Signed-off-by: Dániel Bátyai dbatyai@inf.u-szeged.hu
This commit is contained in:
Dániel Bátyai
2019-10-02 16:55:16 +02:00
committed by Robert Fancsik
parent c3510fc03d
commit c3bb516e4a
20 changed files with 1415 additions and 1337 deletions
Download Patch File Download Diff File Expand all files Collapse all files
jerry-core/ecma/operations/ecma-regexp-object.c
+585 -694
View File
File diff suppressed because it is too large Load Diff
jerry-core/ecma/operations/ecma-regexp-object.h
+46 -14
View File
@@ -35,34 +35,66 @@
*/
typedef enum
{
RE_FLAG_EMPTY = 0u, /* Empty RegExp flags */
RE_FLAG_GLOBAL = (1u << 1), /**< ECMA-262 v5, 15.10.7.2 */
RE_FLAG_IGNORE_CASE = (1u << 2), /**< ECMA-262 v5, 15.10.7.3 */
RE_FLAG_MULTILINE = (1u << 3) /**< ECMA-262 v5, 15.10.7.4 */
} re_flags_t;
} ecma_regexp_flags_t;
/**
* Structure for storing capturing group results
*/
typedef struct
{
const lit_utf8_byte_t *begin_p; /**< substring start pointer */
const lit_utf8_byte_t *end_p; /**< substring end pointer */
} ecma_regexp_capture_t;
/**
* Structure for storing non-capturing group results
*/
typedef struct
{
const lit_utf8_byte_t *str_p; /**< string pointer */
} ecma_regexp_non_capture_t;
#if (JERRY_STACK_LIMIT != 0)
/**
* Value used ase result when stack limit is reached
*/
#define ECMA_RE_OUT_OF_STACK ((const lit_utf8_byte_t *) UINTPTR_MAX)
/**
* Checks if the stack limit has been reached during regexp matching
*/
#define ECMA_RE_STACK_LIMIT_REACHED(p) (JERRY_UNLIKELY (p == ECMA_RE_OUT_OF_STACK))
#else /* JERRY_STACK_LIMIT == 0 */
#define ECMA_RE_STACK_LIMIT_REACHED(p) (false)
#endif /* JERRY_STACK_LIMIT != 0 */
/**
* RegExp executor context
*/
typedef struct
{
const lit_utf8_byte_t **saved_p; /**< saved result string pointers, ECMA 262 v5, 15.10.2.1, State */
const lit_utf8_byte_t *input_start_p; /**< start of input pattern string */
const lit_utf8_byte_t *input_end_p; /**< end of input pattern string */
uint32_t num_of_captures; /**< number of capture groups */
uint32_t num_of_non_captures; /**< number of non-capture groups */
uint32_t *num_of_iterations_p; /**< number of iterations */
uint16_t flags; /**< RegExp flags */
} re_matcher_ctx_t;
const lit_utf8_byte_t *input_end_p; /**< end of input string */
const lit_utf8_byte_t *input_start_p; /**< start of input string */
uint32_t captures_count; /**< number of capture groups */
ecma_regexp_capture_t *captures_p; /**< capturing groups */
uint32_t non_captures_count; /**< number of non-capture groups */
ecma_regexp_non_capture_t *non_captures_p; /**< non-capturing groups */
uint32_t *iterations_p; /**< number of iterations */
uint16_t flags; /**< RegExp flags */
} ecma_regexp_ctx_t;
ecma_value_t ecma_op_create_regexp_object_from_bytecode (re_compiled_code_t *bytecode_p);
ecma_value_t ecma_op_create_regexp_object (ecma_string_t *pattern_p, uint16_t flags);
ecma_value_t ecma_regexp_exec_helper (ecma_value_t regexp_value, ecma_value_t input_string, bool ignore_global);
ecma_value_t ecma_regexp_read_pattern_str_helper (ecma_value_t pattern_arg, ecma_string_t **pattern_string_p);
ecma_char_t re_canonicalize (ecma_char_t ch, bool is_ignorecase);
void re_set_result_array_properties (ecma_object_t *array_obj_p, ecma_string_t *input_str_p, uint32_t num_of_elements,
int32_t index);
ecma_value_t re_parse_regexp_flags (ecma_string_t *flags_str_p, uint16_t *flags_p);
void re_initialize_props (ecma_object_t *re_obj_p, ecma_string_t *source_p, uint16_t flags);
ecma_char_t ecma_regexp_canonicalize (ecma_char_t ch, bool is_ignorecase);
ecma_char_t ecma_regexp_canonicalize_char (ecma_char_t ch);
ecma_value_t ecma_regexp_parse_flags (ecma_string_t *flags_str_p, uint16_t *flags_p);
void ecma_regexp_initialize_props (ecma_object_t *re_obj_p, ecma_string_t *source_p, uint16_t flags);
/**
* @}