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/parser/regexp/re-bytecode.c
+98 -63
View File
@@ -30,8 +30,18 @@
/**
* Size of block of RegExp bytecode. Used for allocation
*
* @return pointer to the RegExp compiled code header
*/
#define REGEXP_BYTECODE_BLOCK_SIZE 256UL
#define REGEXP_BYTECODE_BLOCK_SIZE 64UL
void
re_initialize_regexp_bytecode (re_bytecode_ctx_t *bc_ctx_p) /**< RegExp bytecode context */
{
bc_ctx_p->block_start_p = jmem_heap_alloc_block (REGEXP_BYTECODE_BLOCK_SIZE);
bc_ctx_p->block_end_p = bc_ctx_p->block_start_p + REGEXP_BYTECODE_BLOCK_SIZE;
bc_ctx_p->current_p = bc_ctx_p->block_start_p + sizeof (re_compiled_code_t);
} /* re_initialize_regexp_bytecode */
/**
* Realloc the bytecode container
@@ -42,26 +52,22 @@ static uint8_t *
re_realloc_regexp_bytecode_block (re_bytecode_ctx_t *bc_ctx_p) /**< RegExp bytecode context */
{
JERRY_ASSERT (bc_ctx_p->block_end_p >= bc_ctx_p->block_start_p);
size_t old_size = (size_t) (bc_ctx_p->block_end_p - bc_ctx_p->block_start_p);
const size_t old_size = (size_t) (bc_ctx_p->block_end_p - bc_ctx_p->block_start_p);
/* If one of the members of RegExp bytecode context is NULL, then all member should be NULL
* (it means first allocation), otherwise all of the members should be a non NULL pointer. */
JERRY_ASSERT ((!bc_ctx_p->current_p && !bc_ctx_p->block_end_p && !bc_ctx_p->block_start_p)
|| (bc_ctx_p->current_p && bc_ctx_p->block_end_p && bc_ctx_p->block_start_p));
size_t new_block_size = old_size + REGEXP_BYTECODE_BLOCK_SIZE;
const size_t new_size = old_size + REGEXP_BYTECODE_BLOCK_SIZE;
JERRY_ASSERT (bc_ctx_p->current_p >= bc_ctx_p->block_start_p);
size_t current_ptr_offset = (size_t) (bc_ctx_p->current_p - bc_ctx_p->block_start_p);
const size_t current_ptr_offset = (size_t) (bc_ctx_p->current_p - bc_ctx_p->block_start_p);
uint8_t *new_block_start_p = (uint8_t *) jmem_heap_alloc_block (new_block_size);
if (bc_ctx_p->current_p)
{
memcpy (new_block_start_p, bc_ctx_p->block_start_p, (size_t) (current_ptr_offset));
jmem_heap_free_block (bc_ctx_p->block_start_p, old_size);
}
bc_ctx_p->block_start_p = new_block_start_p;
bc_ctx_p->block_end_p = new_block_start_p + new_block_size;
bc_ctx_p->current_p = new_block_start_p + current_ptr_offset;
bc_ctx_p->block_start_p = jmem_heap_realloc_block (bc_ctx_p->block_start_p,
old_size,
new_size);
bc_ctx_p->block_end_p = bc_ctx_p->block_start_p + new_size;
bc_ctx_p->current_p = bc_ctx_p->block_start_p + current_ptr_offset;
return bc_ctx_p->current_p;
} /* re_realloc_regexp_bytecode_block */
@@ -69,54 +75,71 @@ re_realloc_regexp_bytecode_block (re_bytecode_ctx_t *bc_ctx_p) /**< RegExp bytec
/**
* Append a new bytecode to the and of the bytecode container
*/
static void
re_bytecode_list_append (re_bytecode_ctx_t *bc_ctx_p, /**< RegExp bytecode context */
uint8_t *bytecode_p, /**< input bytecode */
size_t length) /**< length of input */
static uint8_t *
re_bytecode_reserve (re_bytecode_ctx_t *bc_ctx_p, /**< RegExp bytecode context */
const size_t size) /**< size */
{
JERRY_ASSERT (length <= REGEXP_BYTECODE_BLOCK_SIZE);
JERRY_ASSERT (size <= REGEXP_BYTECODE_BLOCK_SIZE);
uint8_t *current_p = bc_ctx_p->current_p;
if (current_p + length > bc_ctx_p->block_end_p)
if (current_p + size > bc_ctx_p->block_end_p)
{
current_p = re_realloc_regexp_bytecode_block (bc_ctx_p);
}
memcpy (current_p, bytecode_p, length);
bc_ctx_p->current_p += length;
} /* re_bytecode_list_append */
bc_ctx_p->current_p += size;
return current_p;
} /* re_bytecode_reserve */
/**
* Insert a new bytecode to the bytecode container
*/
void
re_bytecode_list_insert (re_bytecode_ctx_t *bc_ctx_p, /**< RegExp bytecode context */
size_t offset, /**< distance from the start of the container */
uint8_t *bytecode_p, /**< input bytecode */
size_t length) /**< length of input */
static void
re_bytecode_insert (re_bytecode_ctx_t *bc_ctx_p, /**< RegExp bytecode context */
const size_t offset, /**< distance from the start of the container */
const size_t size) /**< size */
{
JERRY_ASSERT (length <= REGEXP_BYTECODE_BLOCK_SIZE);
JERRY_ASSERT (size <= REGEXP_BYTECODE_BLOCK_SIZE);
uint8_t *current_p = bc_ctx_p->current_p;
if (current_p + length > bc_ctx_p->block_end_p)
if (current_p + size > bc_ctx_p->block_end_p)
{
re_realloc_regexp_bytecode_block (bc_ctx_p);
}
uint8_t *src_p = bc_ctx_p->block_start_p + offset;
if ((re_get_bytecode_length (bc_ctx_p) - offset) > 0)
uint8_t *dest_p = bc_ctx_p->block_start_p + offset;
const size_t bytecode_length = re_get_bytecode_length (bc_ctx_p);
if (bytecode_length - offset > 0)
{
uint8_t *dest_p = src_p + length;
uint8_t *tmp_block_start_p;
tmp_block_start_p = (uint8_t *) jmem_heap_alloc_block (re_get_bytecode_length (bc_ctx_p) - offset);
memcpy (tmp_block_start_p, src_p, (size_t) (re_get_bytecode_length (bc_ctx_p) - offset));
memcpy (dest_p, tmp_block_start_p, (size_t) (re_get_bytecode_length (bc_ctx_p) - offset));
jmem_heap_free_block (tmp_block_start_p, re_get_bytecode_length (bc_ctx_p) - offset);
memmove (dest_p + size, dest_p, bytecode_length - offset);
}
memcpy (src_p, bytecode_p, length);
bc_ctx_p->current_p += length;
} /* re_bytecode_list_insert */
bc_ctx_p->current_p += size;
} /* re_bytecode_insert */
/**
* Encode ecma_char_t into bytecode
*/
static void
re_encode_char (uint8_t *dest_p, /**< destination */
const ecma_char_t c) /**< character */
{
*dest_p++ = (uint8_t) ((c >> 8) & 0xFF);
*dest_p = (uint8_t) (c & 0xFF);
} /* re_encode_char */
/**
* Encode uint32_t into bytecode
*/
static void
re_encode_u32 (uint8_t *dest_p, /**< destination */
const uint32_t u) /**< uint32 value */
{
*dest_p++ = (uint8_t) ((u >> 24) & 0xFF);
*dest_p++ = (uint8_t) ((u >> 16) & 0xFF);
*dest_p++ = (uint8_t) ((u >> 8) & 0xFF);
*dest_p = (uint8_t) (u & 0xFF);
} /* re_encode_u32 */
/**
* Get a character from the RegExp bytecode and increase the bytecode position
@@ -124,10 +147,12 @@ re_bytecode_list_insert (re_bytecode_ctx_t *bc_ctx_p, /**< RegExp bytecode conte
* @return ecma character
*/
inline ecma_char_t JERRY_ATTR_ALWAYS_INLINE
re_get_char (uint8_t **bc_p) /**< pointer to bytecode start */
re_get_char (const uint8_t **bc_p) /**< pointer to bytecode start */
{
ecma_char_t chr;
memcpy (&chr, *bc_p, sizeof (ecma_char_t));
const uint8_t *src_p = *bc_p;
ecma_char_t chr = (ecma_char_t) *src_p++;
chr = (ecma_char_t) (chr << 8);
chr = (ecma_char_t) (chr | *src_p);
(*bc_p) += sizeof (ecma_char_t);
return chr;
} /* re_get_char */
@@ -138,11 +163,9 @@ re_get_char (uint8_t **bc_p) /**< pointer to bytecode start */
* @return current RegExp opcode
*/
inline re_opcode_t JERRY_ATTR_ALWAYS_INLINE
re_get_opcode (uint8_t **bc_p) /**< pointer to bytecode start */
re_get_opcode (const uint8_t **bc_p) /**< pointer to bytecode start */
{
uint8_t bytecode = **bc_p;
(*bc_p) += sizeof (uint8_t);
return (re_opcode_t) bytecode;
return (re_opcode_t) *((*bc_p)++);
} /* re_get_opcode */
/**
@@ -151,10 +174,17 @@ re_get_opcode (uint8_t **bc_p) /**< pointer to bytecode start */
* @return opcode parameter
*/
inline uint32_t JERRY_ATTR_ALWAYS_INLINE
re_get_value (uint8_t **bc_p) /**< pointer to bytecode start */
re_get_value (const uint8_t **bc_p) /**< pointer to bytecode start */
{
uint32_t value;
memcpy (&value, *bc_p, sizeof (uint32_t));
const uint8_t *src_p = *bc_p;
uint32_t value = (uint32_t) (*src_p++);
value <<= 8;
value |= ((uint32_t) (*src_p++));
value <<= 8;
value |= ((uint32_t) (*src_p++));
value <<= 8;
value |= ((uint32_t) (*src_p++));
(*bc_p) += sizeof (uint32_t);
return value;
} /* re_get_value */
@@ -175,9 +205,10 @@ re_get_bytecode_length (re_bytecode_ctx_t *bc_ctx_p) /**< RegExp bytecode contex
*/
void
re_append_opcode (re_bytecode_ctx_t *bc_ctx_p, /**< RegExp bytecode context */
re_opcode_t opcode) /**< input opcode */
const re_opcode_t opcode) /**< input opcode */
{
re_bytecode_list_append (bc_ctx_p, (uint8_t *) &opcode, sizeof (uint8_t));
uint8_t *dest_p = re_bytecode_reserve (bc_ctx_p, sizeof (uint8_t));
*dest_p = (uint8_t) opcode;
} /* re_append_opcode */
/**
@@ -185,9 +216,10 @@ re_append_opcode (re_bytecode_ctx_t *bc_ctx_p, /**< RegExp bytecode context */
*/
void
re_append_u32 (re_bytecode_ctx_t *bc_ctx_p, /**< RegExp bytecode context */
uint32_t value) /**< input value */
const uint32_t value) /**< input value */
{
re_bytecode_list_append (bc_ctx_p, (uint8_t *) &value, sizeof (uint32_t));
uint8_t *dest_p = re_bytecode_reserve (bc_ctx_p, sizeof (uint32_t));
re_encode_u32 (dest_p, value);
} /* re_append_u32 */
/**
@@ -195,9 +227,10 @@ re_append_u32 (re_bytecode_ctx_t *bc_ctx_p, /**< RegExp bytecode context */
*/
void
re_append_char (re_bytecode_ctx_t *bc_ctx_p, /**< RegExp bytecode context */
ecma_char_t input_char) /**< input char */
const ecma_char_t input_char) /**< input char */
{
re_bytecode_list_append (bc_ctx_p, (uint8_t *) &input_char, sizeof (ecma_char_t));
uint8_t *dest_p = re_bytecode_reserve (bc_ctx_p, sizeof (ecma_char_t));
re_encode_char (dest_p, input_char);
} /* re_append_char */
/**
@@ -216,10 +249,11 @@ re_append_jump_offset (re_bytecode_ctx_t *bc_ctx_p, /**< RegExp bytecode context
*/
void
re_insert_opcode (re_bytecode_ctx_t *bc_ctx_p, /**< RegExp bytecode context */
uint32_t offset, /**< distance from the start of the container */
re_opcode_t opcode) /**< input opcode */
const uint32_t offset, /**< distance from the start of the container */
const re_opcode_t opcode) /**< input opcode */
{
re_bytecode_list_insert (bc_ctx_p, offset, (uint8_t *) &opcode, sizeof (uint8_t));
re_bytecode_insert (bc_ctx_p, offset, sizeof (uint8_t));
*(bc_ctx_p->block_start_p + offset) = (uint8_t) opcode;
} /* re_insert_opcode */
/**
@@ -230,7 +264,8 @@ re_insert_u32 (re_bytecode_ctx_t *bc_ctx_p, /**< RegExp bytecode context */
uint32_t offset, /**< distance from the start of the container */
uint32_t value) /**< input value */
{
re_bytecode_list_insert (bc_ctx_p, offset, (uint8_t *) &value, sizeof (uint32_t));
re_bytecode_insert (bc_ctx_p, offset, sizeof (uint32_t));
re_encode_u32 (bc_ctx_p->block_start_p + offset, value);
} /* re_insert_u32 */
#if ENABLED (JERRY_REGEXP_DUMP_BYTE_CODE)
@@ -242,10 +277,10 @@ re_dump_bytecode (re_bytecode_ctx_t *bc_ctx_p) /**< RegExp bytecode context */
{
re_compiled_code_t *compiled_code_p = (re_compiled_code_t *) bc_ctx_p->block_start_p;
JERRY_DEBUG_MSG ("%d ", compiled_code_p->header.status_flags);
JERRY_DEBUG_MSG ("%d ", compiled_code_p->num_of_captures);
JERRY_DEBUG_MSG ("%d | ", compiled_code_p->num_of_non_captures);
JERRY_DEBUG_MSG ("%d ", compiled_code_p->captures_count);
JERRY_DEBUG_MSG ("%d | ", compiled_code_p->non_captures_count);
uint8_t *bytecode_p = (uint8_t *) (compiled_code_p + 1);
const uint8_t *bytecode_p = (const uint8_t *) (compiled_code_p + 1);
re_opcode_t op;
while ((op = re_get_opcode (&bytecode_p)))
jerry-core/parser/regexp/re-bytecode.h
+17 -12
View File
@@ -85,9 +85,9 @@ typedef enum
typedef struct
{
ecma_compiled_code_t header; /**< compiled code header */
ecma_value_t pattern; /**< original RegExp pattern */
uint32_t num_of_captures; /**< number of capturing brackets */
uint32_t num_of_non_captures; /**< number of non capturing brackets */
ecma_value_t source; /**< original RegExp pattern */
uint32_t captures_count; /**< number of capturing brackets */
uint32_t non_captures_count; /**< number of non capturing brackets */
} re_compiled_code_t;
/**
@@ -100,19 +100,24 @@ typedef struct
uint8_t *current_p; /**< current position in bytecode */
} re_bytecode_ctx_t;
re_opcode_t re_get_opcode (uint8_t **bc_p);
ecma_char_t re_get_char (uint8_t **bc_p);
uint32_t re_get_value (uint8_t **bc_p);
re_opcode_t re_get_opcode (const uint8_t **bc_p);
ecma_char_t re_get_char (const uint8_t **bc_p);
uint32_t re_get_value (const uint8_t **bc_p);
uint32_t JERRY_ATTR_PURE re_get_bytecode_length (re_bytecode_ctx_t *bc_ctx_p);
void re_append_opcode (re_bytecode_ctx_t *bc_ctx_p, re_opcode_t opcode);
void re_append_u32 (re_bytecode_ctx_t *bc_ctx_p, uint32_t value);
void re_append_char (re_bytecode_ctx_t *bc_ctx_p, ecma_char_t input_char);
void re_initialize_regexp_bytecode (re_bytecode_ctx_t *bc_ctx_p);
void re_append_opcode (re_bytecode_ctx_t *bc_ctx_p, const re_opcode_t opcode);
void re_append_u32 (re_bytecode_ctx_t *bc_ctx_p, const uint32_t value);
void re_append_char (re_bytecode_ctx_t *bc_ctx_p, const ecma_char_t input_char);
void re_append_jump_offset (re_bytecode_ctx_t *bc_ctx_p, uint32_t value);
void re_insert_opcode (re_bytecode_ctx_t *bc_ctx_p, uint32_t offset, re_opcode_t opcode);
void re_insert_u32 (re_bytecode_ctx_t *bc_ctx_p, uint32_t offset, uint32_t value);
void re_bytecode_list_insert (re_bytecode_ctx_t *bc_ctx_p, size_t offset, uint8_t *bytecode_p, size_t length);
void re_insert_opcode (re_bytecode_ctx_t *bc_ctx_p, const uint32_t offset, const re_opcode_t opcode);
void re_insert_u32 (re_bytecode_ctx_t *bc_ctx_p, const uint32_t offset, const uint32_t value);
void re_bytecode_list_insert (re_bytecode_ctx_t *bc_ctx_p,
const size_t offset,
const uint8_t *bytecode_p,
const size_t length);
#if ENABLED (JERRY_REGEXP_DUMP_BYTE_CODE)
void re_dump_bytecode (re_bytecode_ctx_t *bc_ctx);
jerry-core/parser/regexp/re-compiler.c
+312 -78
View File
@@ -17,6 +17,7 @@
#include "ecma-helpers.h"
#include "ecma-regexp-object.h"
#include "ecma-try-catch-macro.h"
#include "lit-char-helpers.h"
#include "jcontext.h"
#include "jrt-libc-includes.h"
#include "jmem.h"
@@ -36,20 +37,6 @@
* @{
*/
/**
* Callback function of character class generation
*/
static void
re_append_char_class (void *re_ctx_p, /**< RegExp compiler context */
ecma_char_t start, /**< character class range from */
ecma_char_t end) /**< character class range to */
{
re_compiler_ctx_t *ctx_p = (re_compiler_ctx_t *) re_ctx_p;
re_append_char (ctx_p->bytecode_ctx_p, start);
re_append_char (ctx_p->bytecode_ctx_p, end);
ctx_p->parser_ctx_p->num_of_classes++;
} /* re_append_char_class */
/**
* Insert simple atom iterator
*
@@ -234,6 +221,276 @@ re_insert_into_group_with_jump (re_compiler_ctx_t *re_ctx_p, /**< RegExp compile
return re_insert_into_group (re_ctx_p, group_start_offset, idx, is_capturable);
} /* re_insert_into_group_with_jump */
/**
* Append a character class range to the bytecode
*/
static void
re_append_char_class (re_compiler_ctx_t *re_ctx_p, /**< RegExp compiler context */
ecma_char_t start, /**< character class range from */
ecma_char_t end) /**< character class range to */
{
re_append_char (re_ctx_p->bytecode_ctx_p, ecma_regexp_canonicalize (start, re_ctx_p->flags & RE_FLAG_IGNORE_CASE));
re_append_char (re_ctx_p->bytecode_ctx_p, ecma_regexp_canonicalize (end, re_ctx_p->flags & RE_FLAG_IGNORE_CASE));
re_ctx_p->parser_ctx_p->classes_count++;
} /* re_append_char_class */
/**
* Read the input pattern and parse the range of character class
*
* @return empty ecma value - if parsed successfully
* error ecma value - otherwise
*
* Returned value must be freed with ecma_free_value
*/
static ecma_value_t
re_parse_char_class (re_compiler_ctx_t *re_ctx_p, /**< number of classes */
re_token_t *out_token_p) /**< [out] output token */
{
re_parser_ctx_t *const parser_ctx_p = re_ctx_p->parser_ctx_p;
out_token_p->qmax = out_token_p->qmin = 1;
parser_ctx_p->classes_count = 0;
ecma_char_t start = LIT_CHAR_UNDEF;
bool is_range = false;
const bool is_char_class = (re_ctx_p->current_token.type == RE_TOK_START_CHAR_CLASS
|| re_ctx_p->current_token.type == RE_TOK_START_INV_CHAR_CLASS);
const ecma_char_t prev_char = lit_utf8_peek_prev (parser_ctx_p->input_curr_p);
if (prev_char != LIT_CHAR_LEFT_SQUARE && prev_char != LIT_CHAR_CIRCUMFLEX)
{
lit_utf8_decr (&parser_ctx_p->input_curr_p);
lit_utf8_decr (&parser_ctx_p->input_curr_p);
}
do
{
if (parser_ctx_p->input_curr_p >= parser_ctx_p->input_end_p)
{
return ecma_raise_syntax_error (ECMA_ERR_MSG ("invalid character class, end of string"));
}
ecma_char_t ch = lit_utf8_read_next (&parser_ctx_p->input_curr_p);
if (ch == LIT_CHAR_RIGHT_SQUARE)
{
if (start != LIT_CHAR_UNDEF)
{
re_append_char_class (re_ctx_p, start, start);
}
break;
}
else if (ch == LIT_CHAR_MINUS)
{
if (parser_ctx_p->input_curr_p >= parser_ctx_p->input_end_p)
{
return ecma_raise_syntax_error (ECMA_ERR_MSG ("invalid character class, end of string after '-'"));
}
if (start != LIT_CHAR_UNDEF
&& !is_range
&& *parser_ctx_p->input_curr_p != LIT_CHAR_RIGHT_SQUARE)
{
is_range = true;
continue;
}
}
else if (ch == LIT_CHAR_BACKSLASH)
{
if (parser_ctx_p->input_curr_p >= parser_ctx_p->input_end_p)
{
return ecma_raise_syntax_error (ECMA_ERR_MSG ("invalid character class, end of string after '\\'"));
}
ch = lit_utf8_read_next (&parser_ctx_p->input_curr_p);
if (ch == LIT_CHAR_LOWERCASE_B)
{
ch = LIT_CHAR_BS;
}
else if (ch == LIT_CHAR_LOWERCASE_F)
{
ch = LIT_CHAR_FF;
}
else if (ch == LIT_CHAR_LOWERCASE_N)
{
ch = LIT_CHAR_LF;
}
else if (ch == LIT_CHAR_LOWERCASE_T)
{
ch = LIT_CHAR_TAB;
}
else if (ch == LIT_CHAR_LOWERCASE_R)
{
ch = LIT_CHAR_CR;
}
else if (ch == LIT_CHAR_LOWERCASE_V)
{
ch = LIT_CHAR_VTAB;
}
else if (ch == LIT_CHAR_LOWERCASE_C)
{
if (parser_ctx_p->input_curr_p < parser_ctx_p->input_end_p)
{
ch = *parser_ctx_p->input_curr_p;
if ((ch >= LIT_CHAR_ASCII_UPPERCASE_LETTERS_BEGIN && ch <= LIT_CHAR_ASCII_UPPERCASE_LETTERS_END)
|| (ch >= LIT_CHAR_ASCII_LOWERCASE_LETTERS_BEGIN && ch <= LIT_CHAR_ASCII_LOWERCASE_LETTERS_END)
|| (ch >= LIT_CHAR_0 && ch <= LIT_CHAR_9))
{
/* See ECMA-262 v5, 15.10.2.10 (Point 3) */
ch = (ch % 32);
parser_ctx_p->input_curr_p++;
}
else
{
ch = LIT_CHAR_LOWERCASE_C;
}
}
}
else if (ch == LIT_CHAR_LOWERCASE_X && re_hex_lookup (parser_ctx_p, 2))
{
ecma_char_t code_unit;
if (!lit_read_code_unit_from_hex (parser_ctx_p->input_curr_p, 2, &code_unit))
{
return ecma_raise_syntax_error (ECMA_ERR_MSG ("invalid character class, end of string after '\\x'"));
}
parser_ctx_p->input_curr_p += 2;
if (parser_ctx_p->input_curr_p < parser_ctx_p->input_end_p
&& is_range == false
&& lit_utf8_peek_next (parser_ctx_p->input_curr_p) == LIT_CHAR_MINUS)
{
start = code_unit;
continue;
}
ch = code_unit;
}
else if (ch == LIT_CHAR_LOWERCASE_U && re_hex_lookup (parser_ctx_p, 4))
{
ecma_char_t code_unit;
if (!lit_read_code_unit_from_hex (parser_ctx_p->input_curr_p, 4, &code_unit))
{
return ecma_raise_syntax_error (ECMA_ERR_MSG ("invalid character class, end of string after '\\u'"));
}
parser_ctx_p->input_curr_p += 4;
if (parser_ctx_p->input_curr_p < parser_ctx_p->input_end_p
&& is_range == false
&& lit_utf8_peek_next (parser_ctx_p->input_curr_p) == LIT_CHAR_MINUS)
{
start = code_unit;
continue;
}
ch = code_unit;
}
else if (ch == LIT_CHAR_LOWERCASE_D)
{
/* See ECMA-262 v5, 15.10.2.12 */
re_append_char_class (re_ctx_p, LIT_CHAR_ASCII_DIGITS_BEGIN, LIT_CHAR_ASCII_DIGITS_END);
ch = LIT_CHAR_UNDEF;
}
else if (ch == LIT_CHAR_UPPERCASE_D)
{
/* See ECMA-262 v5, 15.10.2.12 */
re_append_char_class (re_ctx_p, LIT_CHAR_NULL, LIT_CHAR_ASCII_DIGITS_BEGIN - 1);
re_append_char_class (re_ctx_p, LIT_CHAR_ASCII_DIGITS_END + 1, LIT_UTF16_CODE_UNIT_MAX);
ch = LIT_CHAR_UNDEF;
}
else if (ch == LIT_CHAR_LOWERCASE_S)
{
/* See ECMA-262 v5, 15.10.2.12 */
re_append_char_class (re_ctx_p, LIT_CHAR_TAB, LIT_CHAR_CR);
re_append_char_class (re_ctx_p, LIT_CHAR_SP, LIT_CHAR_SP);
re_append_char_class (re_ctx_p, LIT_CHAR_NBSP, LIT_CHAR_NBSP);
re_append_char_class (re_ctx_p, 0x1680UL, 0x1680UL); /* Ogham Space Mark */
re_append_char_class (re_ctx_p, 0x180EUL, 0x180EUL); /* Mongolian Vowel Separator */
re_append_char_class (re_ctx_p, 0x2000UL, 0x200AUL); /* En Quad - Hair Space */
re_append_char_class (re_ctx_p, LIT_CHAR_LS, LIT_CHAR_PS);
re_append_char_class (re_ctx_p, 0x202FUL, 0x202FUL); /* Narrow No-Break Space */
re_append_char_class (re_ctx_p, 0x205FUL, 0x205FUL); /* Medium Mathematical Space */
re_append_char_class (re_ctx_p, 0x3000UL, 0x3000UL); /* Ideographic Space */
re_append_char_class (re_ctx_p, LIT_CHAR_BOM, LIT_CHAR_BOM);
ch = LIT_CHAR_UNDEF;
}
else if (ch == LIT_CHAR_UPPERCASE_S)
{
/* See ECMA-262 v5, 15.10.2.12 */
re_append_char_class (re_ctx_p, LIT_CHAR_NULL, LIT_CHAR_TAB - 1);
re_append_char_class (re_ctx_p, LIT_CHAR_CR + 1, LIT_CHAR_SP - 1);
re_append_char_class (re_ctx_p, LIT_CHAR_SP + 1, LIT_CHAR_NBSP - 1);
re_append_char_class (re_ctx_p, LIT_CHAR_NBSP + 1, 0x167FUL);
re_append_char_class (re_ctx_p, 0x1681UL, 0x180DUL);
re_append_char_class (re_ctx_p, 0x180FUL, 0x1FFFUL);
re_append_char_class (re_ctx_p, 0x200BUL, LIT_CHAR_LS - 1);
re_append_char_class (re_ctx_p, LIT_CHAR_PS + 1, 0x202EUL);
re_append_char_class (re_ctx_p, 0x2030UL, 0x205EUL);
re_append_char_class (re_ctx_p, 0x2060UL, 0x2FFFUL);
re_append_char_class (re_ctx_p, 0x3001UL, LIT_CHAR_BOM - 1);
re_append_char_class (re_ctx_p, LIT_CHAR_BOM + 1, LIT_UTF16_CODE_UNIT_MAX);
ch = LIT_CHAR_UNDEF;
}
else if (ch == LIT_CHAR_LOWERCASE_W)
{
/* See ECMA-262 v5, 15.10.2.12 */
re_append_char_class (re_ctx_p, LIT_CHAR_0, LIT_CHAR_9);
re_append_char_class (re_ctx_p, LIT_CHAR_UPPERCASE_A, LIT_CHAR_UPPERCASE_Z);
re_append_char_class (re_ctx_p, LIT_CHAR_UNDERSCORE, LIT_CHAR_UNDERSCORE);
re_append_char_class (re_ctx_p, LIT_CHAR_LOWERCASE_A, LIT_CHAR_LOWERCASE_Z);
ch = LIT_CHAR_UNDEF;
}
else if (ch == LIT_CHAR_UPPERCASE_W)
{
/* See ECMA-262 v5, 15.10.2.12 */
re_append_char_class (re_ctx_p, LIT_CHAR_NULL, LIT_CHAR_0 - 1);
re_append_char_class (re_ctx_p, LIT_CHAR_9 + 1, LIT_CHAR_UPPERCASE_A - 1);
re_append_char_class (re_ctx_p, LIT_CHAR_UPPERCASE_Z + 1, LIT_CHAR_UNDERSCORE - 1);
re_append_char_class (re_ctx_p, LIT_CHAR_UNDERSCORE + 1, LIT_CHAR_LOWERCASE_A - 1);
re_append_char_class (re_ctx_p, LIT_CHAR_LOWERCASE_Z + 1, LIT_UTF16_CODE_UNIT_MAX);
ch = LIT_CHAR_UNDEF;
}
else if (lit_char_is_octal_digit ((ecma_char_t) ch)
&& ch != LIT_CHAR_0)
{
lit_utf8_decr (&parser_ctx_p->input_curr_p);
ch = (ecma_char_t) re_parse_octal (parser_ctx_p);
}
} /* ch == LIT_CHAR_BACKSLASH */
if (start != LIT_CHAR_UNDEF)
{
if (is_range)
{
if (start > ch)
{
return ecma_raise_syntax_error (ECMA_ERR_MSG ("invalid character class, wrong order"));
}
else
{
re_append_char_class (re_ctx_p, start, ch);
start = LIT_CHAR_UNDEF;
is_range = false;
}
}
else
{
re_append_char_class (re_ctx_p, start, start);
start = ch;
}
}
else
{
start = ch;
}
}
while (is_char_class);
return re_parse_iterator (parser_ctx_p, out_token_p);
} /* re_parse_char_class */
/**
* Parse alternatives
*
@@ -251,18 +508,17 @@ re_parse_alternative (re_compiler_ctx_t *re_ctx_p, /**< RegExp compiler context
re_bytecode_ctx_t *bc_ctx_p = re_ctx_p->bytecode_ctx_p;
ecma_value_t ret_value = ECMA_VALUE_EMPTY;
uint32_t alterantive_offset = re_get_bytecode_length (re_ctx_p->bytecode_ctx_p);
bool should_loop = true;
uint32_t alternative_offset = re_get_bytecode_length (re_ctx_p->bytecode_ctx_p);
while (ecma_is_value_empty (ret_value) && should_loop)
while (ecma_is_value_empty (ret_value))
{
ecma_value_t next_token_result = re_parse_next_token (re_ctx_p->parser_ctx_p,
&(re_ctx_p->current_token));
if (ECMA_IS_VALUE_ERROR (next_token_result))
{
ret_value = next_token_result;
break;
return next_token_result;
}
JERRY_ASSERT (ecma_is_value_empty (next_token_result));
uint32_t new_atom_start_offset = re_get_bytecode_length (re_ctx_p->bytecode_ctx_p);
@@ -271,7 +527,7 @@ re_parse_alternative (re_compiler_ctx_t *re_ctx_p, /**< RegExp compiler context
{
case RE_TOK_START_CAPTURE_GROUP:
{
idx = re_ctx_p->num_of_captures++;
idx = re_ctx_p->captures_count++;
JERRY_TRACE_MSG ("Compile a capture group start (idx: %u)\n", (unsigned int) idx);
ret_value = re_parse_alternative (re_ctx_p, false);
@@ -285,7 +541,7 @@ re_parse_alternative (re_compiler_ctx_t *re_ctx_p, /**< RegExp compiler context
}
case RE_TOK_START_NON_CAPTURE_GROUP:
{
idx = re_ctx_p->num_of_non_captures++;
idx = re_ctx_p->non_captures_count++;
JERRY_TRACE_MSG ("Compile a non-capture group start (idx: %u)\n", (unsigned int) idx);
ret_value = re_parse_alternative (re_ctx_p, false);
@@ -304,8 +560,8 @@ re_parse_alternative (re_compiler_ctx_t *re_ctx_p, /**< RegExp compiler context
(unsigned int) re_ctx_p->current_token.qmax);
re_append_opcode (bc_ctx_p, RE_OP_CHAR);
re_append_char (bc_ctx_p, re_canonicalize ((ecma_char_t) re_ctx_p->current_token.value,
re_ctx_p->flags & RE_FLAG_IGNORE_CASE));
re_append_char (bc_ctx_p, ecma_regexp_canonicalize ((ecma_char_t) re_ctx_p->current_token.value,
re_ctx_p->flags & RE_FLAG_IGNORE_CASE));
ret_value = re_insert_simple_iterator (re_ctx_p, new_atom_start_offset);
break;
@@ -321,9 +577,9 @@ re_parse_alternative (re_compiler_ctx_t *re_ctx_p, /**< RegExp compiler context
case RE_TOK_ALTERNATIVE:
{
JERRY_TRACE_MSG ("Compile an alternative\n");
re_insert_u32 (bc_ctx_p, alterantive_offset, re_get_bytecode_length (bc_ctx_p) - alterantive_offset);
re_insert_u32 (bc_ctx_p, alternative_offset, re_get_bytecode_length (bc_ctx_p) - alternative_offset);
re_append_opcode (bc_ctx_p, RE_OP_ALTERNATIVE);
alterantive_offset = re_get_bytecode_length (re_ctx_p->bytecode_ctx_p);
alternative_offset = re_get_bytecode_length (re_ctx_p->bytecode_ctx_p);
break;
}
case RE_TOK_ASSERT_START:
@@ -353,7 +609,7 @@ re_parse_alternative (re_compiler_ctx_t *re_ctx_p, /**< RegExp compiler context
case RE_TOK_ASSERT_START_POS_LOOKAHEAD:
{
JERRY_TRACE_MSG ("Compile a positive lookahead assertion\n");
idx = re_ctx_p->num_of_non_captures++;
idx = re_ctx_p->non_captures_count++;
re_append_opcode (bc_ctx_p, RE_OP_LOOKAHEAD_POS);
ret_value = re_parse_alternative (re_ctx_p, false);
@@ -370,7 +626,7 @@ re_parse_alternative (re_compiler_ctx_t *re_ctx_p, /**< RegExp compiler context
case RE_TOK_ASSERT_START_NEG_LOOKAHEAD:
{
JERRY_TRACE_MSG ("Compile a negative lookahead assertion\n");
idx = re_ctx_p->num_of_non_captures++;
idx = re_ctx_p->non_captures_count++;
re_append_opcode (bc_ctx_p, RE_OP_LOOKAHEAD_NEG);
ret_value = re_parse_alternative (re_ctx_p, false);
@@ -387,7 +643,7 @@ re_parse_alternative (re_compiler_ctx_t *re_ctx_p, /**< RegExp compiler context
case RE_TOK_BACKREFERENCE:
{
uint32_t backref = (uint32_t) re_ctx_p->current_token.value;
idx = re_ctx_p->num_of_non_captures++;
idx = re_ctx_p->non_captures_count++;
if (backref > re_ctx_p->highest_backref)
{
@@ -417,14 +673,12 @@ re_parse_alternative (re_compiler_ctx_t *re_ctx_p, /**< RegExp compiler context
: RE_OP_CHAR_CLASS);
uint32_t offset = re_get_bytecode_length (re_ctx_p->bytecode_ctx_p);
ret_value = re_parse_char_class (re_ctx_p->parser_ctx_p,
re_append_char_class,
re_ctx_p,
ret_value = re_parse_char_class (re_ctx_p,
&(re_ctx_p->current_token));
if (!ECMA_IS_VALUE_ERROR (ret_value))
{
re_insert_u32 (bc_ctx_p, offset, re_ctx_p->parser_ctx_p->num_of_classes);
re_insert_u32 (bc_ctx_p, offset, re_ctx_p->parser_ctx_p->classes_count);
ret_value = re_insert_simple_iterator (re_ctx_p, new_atom_start_offset);
}
@@ -436,33 +690,25 @@ re_parse_alternative (re_compiler_ctx_t *re_ctx_p, /**< RegExp compiler context
if (expect_eof)
{
ret_value = ecma_raise_syntax_error (ECMA_ERR_MSG ("Unexpected end of paren."));
return ecma_raise_syntax_error (ECMA_ERR_MSG ("Unexpected end of paren."));
}
else
{
re_insert_u32 (bc_ctx_p, alterantive_offset, re_get_bytecode_length (bc_ctx_p) - alterantive_offset);
should_loop = false;
}
break;
re_insert_u32 (bc_ctx_p, alternative_offset, re_get_bytecode_length (bc_ctx_p) - alternative_offset);
return ECMA_VALUE_EMPTY;
}
case RE_TOK_EOF:
{
if (!expect_eof)
{
ret_value = ecma_raise_syntax_error (ECMA_ERR_MSG ("Unexpected end of pattern."));
}
else
{
re_insert_u32 (bc_ctx_p, alterantive_offset, re_get_bytecode_length (bc_ctx_p) - alterantive_offset);
should_loop = false;
return ecma_raise_syntax_error (ECMA_ERR_MSG ("Unexpected end of pattern."));
}
break;
re_insert_u32 (bc_ctx_p, alternative_offset, re_get_bytecode_length (bc_ctx_p) - alternative_offset);
return ECMA_VALUE_EMPTY;
}
default:
{
ret_value = ecma_raise_syntax_error (ECMA_ERR_MSG ("Unexpected RegExp token."));
break;
return ecma_raise_syntax_error (ECMA_ERR_MSG ("Unexpected RegExp token."));
}
}
}
@@ -488,7 +734,7 @@ re_find_bytecode_in_cache (ecma_string_t *pattern_str_p, /**< pattern string */
if (cached_bytecode_p != NULL)
{
ecma_string_t *cached_pattern_str_p = ecma_get_string_from_value (cached_bytecode_p->pattern);
ecma_string_t *cached_pattern_str_p = ecma_get_string_from_value (cached_bytecode_p->source);
if ((cached_bytecode_p->header.status_flags & RE_FLAGS_MASK) == flags
&& ecma_compare_ecma_strings (cached_pattern_str_p, pattern_str_p))
@@ -559,14 +805,11 @@ re_compile_bytecode (const re_compiled_code_t **out_bytecode_p, /**< [out] point
re_compiler_ctx_t re_ctx;
re_ctx.flags = flags;
re_ctx.highest_backref = 0;
re_ctx.num_of_non_captures = 0;
re_ctx.non_captures_count = 0;
re_bytecode_ctx_t bc_ctx;
bc_ctx.block_start_p = NULL;
bc_ctx.block_end_p = NULL;
bc_ctx.current_p = NULL;
re_ctx.bytecode_ctx_p = &bc_ctx;
re_initialize_regexp_bytecode (&bc_ctx);
ECMA_STRING_TO_UTF8_STRING (pattern_str_p, pattern_start_p, pattern_start_size);
@@ -574,23 +817,23 @@ re_compile_bytecode (const re_compiled_code_t **out_bytecode_p, /**< [out] point
parser_ctx.input_start_p = pattern_start_p;
parser_ctx.input_curr_p = (lit_utf8_byte_t *) pattern_start_p;
parser_ctx.input_end_p = pattern_start_p + pattern_start_size;
parser_ctx.num_of_groups = -1;
parser_ctx.groups_count = -1;
re_ctx.parser_ctx_p = &parser_ctx;
/* 1. Parse RegExp pattern */
re_ctx.num_of_captures = 1;
/* Parse RegExp pattern */
re_ctx.captures_count = 1;
re_append_opcode (&bc_ctx, RE_OP_SAVE_AT_START);
ecma_value_t parse_alt_result = re_parse_alternative (&re_ctx, true);
ecma_value_t result = re_parse_alternative (&re_ctx, true);
ECMA_FINALIZE_UTF8_STRING (pattern_start_p, pattern_start_size);
if (ECMA_IS_VALUE_ERROR (parse_alt_result))
if (ECMA_IS_VALUE_ERROR (result))
{
ret_value = parse_alt_result;
ret_value = result;
}
/* 2. Check for invalid backreference */
else if (re_ctx.highest_backref >= re_ctx.num_of_captures)
/* Check for invalid backreference */
else if (re_ctx.highest_backref >= re_ctx.captures_count)
{
ret_value = ecma_raise_syntax_error ("Invalid backreference.\n");
}
@@ -599,20 +842,14 @@ re_compile_bytecode (const re_compiled_code_t **out_bytecode_p, /**< [out] point
re_append_opcode (&bc_ctx, RE_OP_SAVE_AND_MATCH);
re_append_opcode (&bc_ctx, RE_OP_EOF);
/* 3. Insert extra informations for bytecode header */
re_compiled_code_t re_compiled_code;
re_compiled_code.header.refs = 1;
re_compiled_code.header.status_flags = re_ctx.flags;
/* Initialize bytecode header */
re_compiled_code_t *re_compiled_code_p = (re_compiled_code_t *) bc_ctx.block_start_p;
re_compiled_code_p->header.refs = 1;
re_compiled_code_p->header.status_flags = re_ctx.flags;
ecma_ref_ecma_string (pattern_str_p);
re_compiled_code.pattern = ecma_make_string_value (pattern_str_p);
re_compiled_code.num_of_captures = re_ctx.num_of_captures * 2;
re_compiled_code.num_of_non_captures = re_ctx.num_of_non_captures;
re_bytecode_list_insert (&bc_ctx,
0,
(uint8_t *) &re_compiled_code,
sizeof (re_compiled_code_t));
re_compiled_code_p->source = ecma_make_string_value (pattern_str_p);
re_compiled_code_p->captures_count = re_ctx.captures_count;
re_compiled_code_p->non_captures_count = re_ctx.non_captures_count;
}
size_t byte_code_size = (size_t) (bc_ctx.block_end_p - bc_ctx.block_start_p);
@@ -633,10 +870,7 @@ re_compile_bytecode (const re_compiled_code_t **out_bytecode_p, /**< [out] point
}
#endif /* ENABLED (JERRY_REGEXP_DUMP_BYTE_CODE) */
/* The RegExp bytecode contains at least a RE_OP_SAVE_AT_START opdoce, so it cannot be NULL. */
JERRY_ASSERT (bc_ctx.block_start_p != NULL);
*out_bytecode_p = (re_compiled_code_t *) bc_ctx.block_start_p;
((re_compiled_code_t *) bc_ctx.block_start_p)->header.size = (uint16_t) (byte_code_size >> JMEM_ALIGNMENT_LOG);
if (cache_idx == RE_CACHE_SIZE)
jerry-core/parser/regexp/re-compiler.h
+2 -2
View File
@@ -38,8 +38,8 @@
typedef struct
{
uint16_t flags; /**< RegExp flags */
uint32_t num_of_captures; /**< number of capture groups */
uint32_t num_of_non_captures; /**< number of non-capture groups */
uint32_t captures_count; /**< number of capture groups */
uint32_t non_captures_count; /**< number of non-capture groups */
uint32_t highest_backref; /**< highest backreference */
re_bytecode_ctx_t *bytecode_ctx_p; /**< pointer of RegExp bytecode context */
re_token_t current_token; /**< current token */
jerry-core/parser/regexp/re-parser.c
+16 -274
View File
@@ -40,26 +40,26 @@
* @return true - if lookup number of characters ahead are hex digits
* false - otherwise
*/
static bool
bool
re_hex_lookup (re_parser_ctx_t *parser_ctx_p, /**< RegExp parser context */
uint32_t lookup) /**< size of lookup */
{
bool is_digit = true;
const lit_utf8_byte_t *curr_p = parser_ctx_p->input_curr_p;
for (uint32_t i = 0; is_digit && i < lookup; i++)
if (JERRY_UNLIKELY (curr_p + lookup > parser_ctx_p->input_end_p))
{
if (curr_p < parser_ctx_p->input_end_p)
{
is_digit = lit_char_is_hex_digit (*curr_p++);
}
else
return false;
}
for (uint32_t i = 0; i < lookup; i++)
{
if (!lit_char_is_hex_digit (*curr_p++))
{
return false;
}
}
return is_digit;
return true;
} /* re_hex_lookup */
/**
@@ -86,7 +86,7 @@ re_parse_non_greedy_char (re_parser_ctx_t *parser_ctx_p) /**< RegExp parser cont
*
* @return uint32_t - parsed octal number
*/
static uint32_t
uint32_t
re_parse_octal (re_parser_ctx_t *parser_ctx_p) /**< RegExp parser context */
{
uint32_t number = 0;
@@ -110,7 +110,7 @@ re_parse_octal (re_parser_ctx_t *parser_ctx_p) /**< RegExp parser context */
*
* Returned value must be freed with ecma_free_value
*/
static ecma_value_t
ecma_value_t
re_parse_iterator (re_parser_ctx_t *parser_ctx_p, /**< RegExp parser context */
re_token_t *re_token_p) /**< [out] output token */
{
@@ -253,7 +253,7 @@ static void
re_count_num_of_groups (re_parser_ctx_t *parser_ctx_p) /**< RegExp parser context */
{
int char_class_in = 0;
parser_ctx_p->num_of_groups = 0;
parser_ctx_p->groups_count = 0;
const lit_utf8_byte_t *curr_p = parser_ctx_p->input_start_p;
while (curr_p < parser_ctx_p->input_end_p)
@@ -287,7 +287,7 @@ re_count_num_of_groups (re_parser_ctx_t *parser_ctx_p) /**< RegExp parser contex
&& *curr_p != LIT_CHAR_QUESTION
&& !char_class_in)
{
parser_ctx_p->num_of_groups++;
parser_ctx_p->groups_count++;
}
break;
}
@@ -295,264 +295,6 @@ re_count_num_of_groups (re_parser_ctx_t *parser_ctx_p) /**< RegExp parser contex
}
} /* re_count_num_of_groups */
/**
* Read the input pattern and parse the range of character class
*
* @return empty ecma value - if parsed successfully
* error ecma value - otherwise
*
* Returned value must be freed with ecma_free_value
*/
ecma_value_t
re_parse_char_class (re_parser_ctx_t *parser_ctx_p, /**< number of classes */
re_char_class_callback append_char_class, /**< callback function,
* which adds the char-ranges
* to the bytecode */
void *re_ctx_p, /**< regexp compiler context */
re_token_t *out_token_p) /**< [out] output token */
{
re_token_type_t token_type = ((re_compiler_ctx_t *) re_ctx_p)->current_token.type;
out_token_p->qmax = out_token_p->qmin = 1;
ecma_char_t start = LIT_CHAR_UNDEF;
bool is_range = false;
parser_ctx_p->num_of_classes = 0;
const ecma_char_t prev_char = lit_utf8_peek_prev (parser_ctx_p->input_curr_p);
if (prev_char != LIT_CHAR_LEFT_SQUARE && prev_char != LIT_CHAR_CIRCUMFLEX)
{
lit_utf8_decr (&parser_ctx_p->input_curr_p);
lit_utf8_decr (&parser_ctx_p->input_curr_p);
}
do
{
if (parser_ctx_p->input_curr_p >= parser_ctx_p->input_end_p)
{
return ecma_raise_syntax_error (ECMA_ERR_MSG ("invalid character class, end of string"));
}
ecma_char_t ch = lit_utf8_read_next (&parser_ctx_p->input_curr_p);
if (ch == LIT_CHAR_RIGHT_SQUARE)
{
if (start != LIT_CHAR_UNDEF)
{
append_char_class (re_ctx_p, start, start);
}
break;
}
else if (ch == LIT_CHAR_MINUS)
{
if (parser_ctx_p->input_curr_p >= parser_ctx_p->input_end_p)
{
return ecma_raise_syntax_error (ECMA_ERR_MSG ("invalid character class, end of string after '-'"));
}
if (start != LIT_CHAR_UNDEF
&& !is_range
&& *parser_ctx_p->input_curr_p != LIT_CHAR_RIGHT_SQUARE)
{
is_range = true;
continue;
}
}
else if (ch == LIT_CHAR_BACKSLASH)
{
if (parser_ctx_p->input_curr_p >= parser_ctx_p->input_end_p)
{
return ecma_raise_syntax_error (ECMA_ERR_MSG ("invalid character class, end of string after '\\'"));
}
ch = lit_utf8_read_next (&parser_ctx_p->input_curr_p);
if (ch == LIT_CHAR_LOWERCASE_B)
{
ch = LIT_CHAR_BS;
}
else if (ch == LIT_CHAR_LOWERCASE_F)
{
ch = LIT_CHAR_FF;
}
else if (ch == LIT_CHAR_LOWERCASE_N)
{
ch = LIT_CHAR_LF;
}
else if (ch == LIT_CHAR_LOWERCASE_T)
{
ch = LIT_CHAR_TAB;
}
else if (ch == LIT_CHAR_LOWERCASE_R)
{
ch = LIT_CHAR_CR;
}
else if (ch == LIT_CHAR_LOWERCASE_V)
{
ch = LIT_CHAR_VTAB;
}
else if (ch == LIT_CHAR_LOWERCASE_C)
{
if (parser_ctx_p->input_curr_p < parser_ctx_p->input_end_p)
{
ch = *parser_ctx_p->input_curr_p;
if ((ch >= LIT_CHAR_ASCII_UPPERCASE_LETTERS_BEGIN && ch <= LIT_CHAR_ASCII_UPPERCASE_LETTERS_END)
|| (ch >= LIT_CHAR_ASCII_LOWERCASE_LETTERS_BEGIN && ch <= LIT_CHAR_ASCII_LOWERCASE_LETTERS_END)
|| (ch >= LIT_CHAR_0 && ch <= LIT_CHAR_9))
{
/* See ECMA-262 v5, 15.10.2.10 (Point 3) */
ch = (ch % 32);
parser_ctx_p->input_curr_p++;
}
else
{
ch = LIT_CHAR_LOWERCASE_C;
}
}
}
else if (ch == LIT_CHAR_LOWERCASE_X && re_hex_lookup (parser_ctx_p, 2))
{
ecma_char_t code_unit;
if (!lit_read_code_unit_from_hex (parser_ctx_p->input_curr_p, 2, &code_unit))
{
return ecma_raise_syntax_error (ECMA_ERR_MSG ("invalid character class, end of string after '\\x'"));
}
parser_ctx_p->input_curr_p += 2;
if (parser_ctx_p->input_curr_p < parser_ctx_p->input_end_p
&& is_range == false
&& lit_utf8_peek_next (parser_ctx_p->input_curr_p) == LIT_CHAR_MINUS)
{
start = code_unit;
continue;
}
ch = code_unit;
}
else if (ch == LIT_CHAR_LOWERCASE_U && re_hex_lookup (parser_ctx_p, 4))
{
ecma_char_t code_unit;
if (!lit_read_code_unit_from_hex (parser_ctx_p->input_curr_p, 4, &code_unit))
{
return ecma_raise_syntax_error (ECMA_ERR_MSG ("invalid character class, end of string after '\\u'"));
}
parser_ctx_p->input_curr_p += 4;
if (parser_ctx_p->input_curr_p < parser_ctx_p->input_end_p
&& is_range == false
&& lit_utf8_peek_next (parser_ctx_p->input_curr_p) == LIT_CHAR_MINUS)
{
start = code_unit;
continue;
}
ch = code_unit;
}
else if (ch == LIT_CHAR_LOWERCASE_D)
{
/* See ECMA-262 v5, 15.10.2.12 */
append_char_class (re_ctx_p, LIT_CHAR_ASCII_DIGITS_BEGIN, LIT_CHAR_ASCII_DIGITS_END);
ch = LIT_CHAR_UNDEF;
}
else if (ch == LIT_CHAR_UPPERCASE_D)
{
/* See ECMA-262 v5, 15.10.2.12 */
append_char_class (re_ctx_p, LIT_CHAR_NULL, LIT_CHAR_ASCII_DIGITS_BEGIN - 1);
append_char_class (re_ctx_p, LIT_CHAR_ASCII_DIGITS_END + 1, LIT_UTF16_CODE_UNIT_MAX);
ch = LIT_CHAR_UNDEF;
}
else if (ch == LIT_CHAR_LOWERCASE_S)
{
/* See ECMA-262 v5, 15.10.2.12 */
append_char_class (re_ctx_p, LIT_CHAR_TAB, LIT_CHAR_CR);
append_char_class (re_ctx_p, LIT_CHAR_SP, LIT_CHAR_SP);
append_char_class (re_ctx_p, LIT_CHAR_NBSP, LIT_CHAR_NBSP);
append_char_class (re_ctx_p, 0x1680UL, 0x1680UL); /* Ogham Space Mark */
append_char_class (re_ctx_p, 0x180EUL, 0x180EUL); /* Mongolian Vowel Separator */
append_char_class (re_ctx_p, 0x2000UL, 0x200AUL); /* En Quad - Hair Space */
append_char_class (re_ctx_p, LIT_CHAR_LS, LIT_CHAR_PS);
append_char_class (re_ctx_p, 0x202FUL, 0x202FUL); /* Narrow No-Break Space */
append_char_class (re_ctx_p, 0x205FUL, 0x205FUL); /* Medium Mathematical Space */
append_char_class (re_ctx_p, 0x3000UL, 0x3000UL); /* Ideographic Space */
append_char_class (re_ctx_p, LIT_CHAR_BOM, LIT_CHAR_BOM);
ch = LIT_CHAR_UNDEF;
}
else if (ch == LIT_CHAR_UPPERCASE_S)
{
/* See ECMA-262 v5, 15.10.2.12 */
append_char_class (re_ctx_p, LIT_CHAR_NULL, LIT_CHAR_TAB - 1);
append_char_class (re_ctx_p, LIT_CHAR_CR + 1, LIT_CHAR_SP - 1);
append_char_class (re_ctx_p, LIT_CHAR_SP + 1, LIT_CHAR_NBSP - 1);
append_char_class (re_ctx_p, LIT_CHAR_NBSP + 1, 0x167FUL);
append_char_class (re_ctx_p, 0x1681UL, 0x180DUL);
append_char_class (re_ctx_p, 0x180FUL, 0x1FFFUL);
append_char_class (re_ctx_p, 0x200BUL, LIT_CHAR_LS - 1);
append_char_class (re_ctx_p, LIT_CHAR_PS + 1, 0x202EUL);
append_char_class (re_ctx_p, 0x2030UL, 0x205EUL);
append_char_class (re_ctx_p, 0x2060UL, 0x2FFFUL);
append_char_class (re_ctx_p, 0x3001UL, LIT_CHAR_BOM - 1);
append_char_class (re_ctx_p, LIT_CHAR_BOM + 1, LIT_UTF16_CODE_UNIT_MAX);
ch = LIT_CHAR_UNDEF;
}
else if (ch == LIT_CHAR_LOWERCASE_W)
{
/* See ECMA-262 v5, 15.10.2.12 */
append_char_class (re_ctx_p, LIT_CHAR_0, LIT_CHAR_9);
append_char_class (re_ctx_p, LIT_CHAR_UPPERCASE_A, LIT_CHAR_UPPERCASE_Z);
append_char_class (re_ctx_p, LIT_CHAR_UNDERSCORE, LIT_CHAR_UNDERSCORE);
append_char_class (re_ctx_p, LIT_CHAR_LOWERCASE_A, LIT_CHAR_LOWERCASE_Z);
ch = LIT_CHAR_UNDEF;
}
else if (ch == LIT_CHAR_UPPERCASE_W)
{
/* See ECMA-262 v5, 15.10.2.12 */
append_char_class (re_ctx_p, LIT_CHAR_NULL, LIT_CHAR_0 - 1);
append_char_class (re_ctx_p, LIT_CHAR_9 + 1, LIT_CHAR_UPPERCASE_A - 1);
append_char_class (re_ctx_p, LIT_CHAR_UPPERCASE_Z + 1, LIT_CHAR_UNDERSCORE - 1);
append_char_class (re_ctx_p, LIT_CHAR_UNDERSCORE + 1, LIT_CHAR_LOWERCASE_A - 1);
append_char_class (re_ctx_p, LIT_CHAR_LOWERCASE_Z + 1, LIT_UTF16_CODE_UNIT_MAX);
ch = LIT_CHAR_UNDEF;
}
else if (lit_char_is_octal_digit ((ecma_char_t) ch)
&& ch != LIT_CHAR_0)
{
lit_utf8_decr (&parser_ctx_p->input_curr_p);
ch = (ecma_char_t) re_parse_octal (parser_ctx_p);
}
} /* ch == LIT_CHAR_BACKSLASH */
if (start != LIT_CHAR_UNDEF)
{
if (is_range)
{
if (start > ch)
{
return ecma_raise_syntax_error (ECMA_ERR_MSG ("invalid character class, wrong order"));
}
else
{
append_char_class (re_ctx_p, start, ch);
start = LIT_CHAR_UNDEF;
is_range = false;
}
}
else
{
append_char_class (re_ctx_p, start, start);
start = ch;
}
}
else
{
start = ch;
}
}
while (token_type == RE_TOK_START_CHAR_CLASS || token_type == RE_TOK_START_INV_CHAR_CLASS);
return re_parse_iterator (parser_ctx_p, out_token_p);
} /* re_parse_char_class */
/**
* Read the input pattern and parse the next token for the RegExp compiler
*
@@ -730,12 +472,12 @@ re_parse_next_token (re_parser_ctx_t *parser_ctx_p, /**< RegExp parser context *
}
else
{
if (parser_ctx_p->num_of_groups == -1)
if (parser_ctx_p->groups_count == -1)
{
re_count_num_of_groups (parser_ctx_p);
}
if (parser_ctx_p->num_of_groups)
if (parser_ctx_p->groups_count)
{
parser_ctx_p->input_curr_p--;
uint32_t number = 0;
@@ -765,7 +507,7 @@ re_parse_next_token (re_parser_ctx_t *parser_ctx_p, /**< RegExp parser context *
}
while (true);
if ((int) number <= parser_ctx_p->num_of_groups)
if ((int) number <= parser_ctx_p->groups_count)
{
out_token_p->type = RE_TOK_BACKREFERENCE;
}
jerry-core/parser/regexp/re-parser.h
+6 -10
View File
@@ -94,18 +94,14 @@ typedef struct
const lit_utf8_byte_t *input_start_p; /**< start of input pattern */
const lit_utf8_byte_t *input_curr_p; /**< current position in input pattern */
const lit_utf8_byte_t *input_end_p; /**< end of input pattern */
int num_of_groups; /**< number of groups */
uint32_t num_of_classes; /**< number of character classes */
int groups_count; /**< number of groups */
uint32_t classes_count; /**< number of character classes */
} re_parser_ctx_t;
typedef void (*re_char_class_callback) (void *re_ctx_p, ecma_char_t start, ecma_char_t end);
ecma_value_t
re_parse_char_class (re_parser_ctx_t *parser_ctx_p, re_char_class_callback append_char_class, void *re_ctx_p,
re_token_t *out_token_p);
ecma_value_t
re_parse_next_token (re_parser_ctx_t *parser_ctx_p, re_token_t *out_token_p);
bool re_hex_lookup (re_parser_ctx_t *parser_ctx_p, uint32_t lookup);
uint32_t re_parse_octal (re_parser_ctx_t *parser_ctx_p);
ecma_value_t re_parse_iterator (re_parser_ctx_t *parser_ctx_p, re_token_t *re_token_p);
ecma_value_t re_parse_next_token (re_parser_ctx_t *parser_ctx_p, re_token_t *out_token_p);
/**
* @}