RegExp refactoring and improvements

Move RegExp bytecode functions to a separate file.
Optimize bytecode lenght on character matching.
Implement a basic RegExp cache to optimize memory
usage on duplicated RegExp in JS files. Also fix
minor style issues and add missing comments. Improve
existing comments.

JerryScript-DCO-1.0-Signed-off-by: László Langó llango.u-szeged@partner.samsung.com
This commit is contained in:
László Langó
2016-02-18 10:23:45 +00:00
parent 3f377692d9
commit 2c72bb1139
16 changed files with 844 additions and 601 deletions
+20 -29
View File
@@ -32,48 +32,39 @@
/**
* RegExp flags
*/
#define RE_FLAG_GLOBAL (1u << 1) /* ECMA-262 v5, 15.10.7.2 */
#define RE_FLAG_IGNORE_CASE (1u << 2) /* ECMA-262 v5, 15.10.7.3 */
#define RE_FLAG_MULTILINE (1u << 3) /* ECMA-262 v5, 15.10.7.4 */
typedef enum
{
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;
/**
* RegExp executor context
*/
typedef struct
{
lit_utf8_byte_t **saved_p; /**< saved result string pointers, ECMA 262 v5, 15.10.2.1, State */
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 */
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;
extern ecma_value_t
ecma_op_create_regexp_object_from_bytecode (re_compiled_code_t *);
extern ecma_value_t
ecma_op_create_regexp_object (ecma_string_t *, ecma_string_t *);
extern ecma_value_t
ecma_regexp_exec_helper (ecma_value_t, ecma_value_t, bool);
extern ecma_char_t
re_canonicalize (ecma_char_t, bool);
extern void
re_set_result_array_properties (ecma_object_t *, ecma_string_t *, uint32_t, int32_t);
extern ecma_value_t
re_parse_regexp_flags (ecma_string_t *, uint16_t *);
extern void
re_initialize_props (ecma_object_t *, ecma_string_t *, uint16_t);
ecma_value_t ecma_op_create_regexp_object_from_bytecode (re_compiled_code_t *);
ecma_value_t ecma_op_create_regexp_object (ecma_string_t *, ecma_string_t *);
ecma_value_t ecma_regexp_exec_helper (ecma_value_t, ecma_value_t, bool);
ecma_char_t re_canonicalize (ecma_char_t, bool);
void re_set_result_array_properties (ecma_object_t *, ecma_string_t *, uint32_t, int32_t);
ecma_value_t re_parse_regexp_flags (ecma_string_t *, uint16_t *);
void re_initialize_props (ecma_object_t *, ecma_string_t *, uint16_t);
/**
* @}
* @}
*/
#endif /* CONFIG_ECMA_COMPACT_PROFILE_DISABLE_REGEXP_BUILTIN */
#endif /* !CONFIG_ECMA_COMPACT_PROFILE_DISABLE_REGEXP_BUILTIN */
#endif /* !ECMA_REGEXP_OBJECT_H */