/**
* Recursive function for RegExp matching. Tests for a regular expression
* match and returns a MatchResult value.
*
* See also:
* ECMA-262 v5, 15.10.2.1
*
* @return completion value
* Returned value must be freed with ecma_free_completion_value
*/
static ecma_completion_value_t
re_match_regexp (re_matcher_ctx_t *re_ctx_p, /**< RegExp matcher context */
re_bytecode_t *bc_p, /**< pointer to the current RegExp bytecode */
lit_utf8_iterator_t iter, /**< input string iterator */
lit_utf8_iterator_t *out_iter_p) /**< Output: matching substring iterator */
{
ecma_completion_value_t ret_value = ecma_make_empty_completion_value ();
re_opcode_t op;
while ((op = re_get_opcode (&bc_p)))
{
switch (op)
{
case RE_OP_MATCH:
{
JERRY_DDLOG ("Execute RE_OP_MATCH: match\n");
*out_iter_p = iter;
ret_value = ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_TRUE);
return ret_value; /* match */
}
case RE_OP_CHAR:
{
if (lit_utf8_iterator_is_eos (&iter))
{
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
bool is_ignorecase = re_ctx_p->flags & RE_FLAG_IGNORE_CASE;
ecma_char_t ch1 = (ecma_char_t) re_get_value (&bc_p); /* Already canonicalized. */
ecma_char_t ch2 = re_canonicalize (lit_utf8_iterator_read_next (&iter), is_ignorecase);
JERRY_DDLOG ("Character matching %d to %d: ", ch1, ch2);
if (ch1 != ch2)
{
JERRY_DDLOG ("fail\n");
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
JERRY_DDLOG ("match\n");
break; /* tail merge */
}
case RE_OP_PERIOD:
{
if (lit_utf8_iterator_is_eos (&iter))
{
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
ecma_char_t ch = lit_utf8_iterator_read_next (&iter);
JERRY_DDLOG ("Period matching '.' to %d: ", (uint32_t) ch);
if (lit_char_is_line_terminator (ch))
{
JERRY_DDLOG ("fail\n");
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
JERRY_DDLOG ("match\n");
break; /* tail merge */
}
case RE_OP_ASSERT_START:
{
JERRY_DDLOG ("Execute RE_OP_ASSERT_START: ");
if ((iter.buf_p + iter.buf_pos.offset) <= re_ctx_p->input_start_p)
{
JERRY_DDLOG ("match\n");
break;
}
if (!(re_ctx_p->flags & RE_FLAG_MULTILINE))
{
JERRY_DDLOG ("fail\n");
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
if (lit_char_is_line_terminator (lit_utf8_iterator_peek_prev (&iter)))
{
JERRY_DDLOG ("match\n");
break;
}
JERRY_DDLOG ("fail\n");
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
case RE_OP_ASSERT_END:
{
JERRY_DDLOG ("Execute RE_OP_ASSERT_END: ");
if ((iter.buf_p + iter.buf_pos.offset) >= re_ctx_p->input_end_p)
{
JERRY_DDLOG ("match\n");
break; /* tail merge */
}
if (!(re_ctx_p->flags & RE_FLAG_MULTILINE))
{
JERRY_DDLOG ("fail\n");
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
if (lit_char_is_line_terminator (lit_utf8_iterator_peek_next (&iter)))
{
JERRY_DDLOG ("match\n");
break; /* tail merge */
}
JERRY_DDLOG ("fail\n");
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
case RE_OP_ASSERT_WORD_BOUNDARY:
case RE_OP_ASSERT_NOT_WORD_BOUNDARY:
{
bool is_wordchar_left, is_wordchar_right;
if ((iter.buf_p + iter.buf_pos.offset) <= re_ctx_p->input_start_p)
{
is_wordchar_left = false; /* not a wordchar */
}
else
{
is_wordchar_left = lit_char_is_word_char (lit_utf8_iterator_peek_prev (&iter));
}
if ((iter.buf_p + iter.buf_pos.offset) >= re_ctx_p->input_end_p)
{
is_wordchar_right = false; /* not a wordchar */
}
else
{
is_wordchar_right = lit_char_is_word_char (lit_utf8_iterator_peek_next (&iter));
}
if (op == RE_OP_ASSERT_WORD_BOUNDARY)
{
JERRY_DDLOG ("Execute RE_OP_ASSERT_WORD_BOUNDARY: ");
if (is_wordchar_left == is_wordchar_right)
{
JERRY_DDLOG ("fail\n");
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
}
else
{
JERRY_ASSERT (op == RE_OP_ASSERT_NOT_WORD_BOUNDARY);
JERRY_DDLOG ("Execute RE_OP_ASSERT_NOT_WORD_BOUNDARY: ");
if (is_wordchar_left != is_wordchar_right)
{
JERRY_DDLOG ("fail\n");
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
}
JERRY_DDLOG ("match\n");
break; /* tail merge */
}
case RE_OP_LOOKAHEAD_POS:
case RE_OP_LOOKAHEAD_NEG:
{
ecma_completion_value_t match_value = ecma_make_empty_completion_value ();
lit_utf8_iterator_t sub_iter = lit_utf8_iterator_create (NULL, 0);
uint32_t array_size = re_ctx_p->num_of_captures + re_ctx_p->num_of_non_captures;
MEM_DEFINE_LOCAL_ARRAY (saved_bck_p, array_size, lit_utf8_iterator_t);
size_t size = (size_t) (array_size) * sizeof (lit_utf8_iterator_t);
memcpy (saved_bck_p, re_ctx_p->saved_p, size);
do
{
uint32_t offset = re_get_value (&bc_p);
if (!sub_iter.buf_p)
{
match_value = re_match_regexp (re_ctx_p, bc_p, iter, &sub_iter);
if (ecma_is_completion_value_throw (match_value))
{
break;
}
}
bc_p += offset;
}
while (re_get_opcode (&bc_p) == RE_OP_ALTERNATIVE);
if (!ecma_is_completion_value_throw (match_value))
{
JERRY_DDLOG ("Execute RE_OP_LOOKAHEAD_POS/NEG: ");
ecma_free_completion_value (match_value);
if ((op == RE_OP_LOOKAHEAD_POS && sub_iter.buf_p)
|| (op == RE_OP_LOOKAHEAD_NEG && !sub_iter.buf_p))
{
JERRY_DDLOG ("match\n");
match_value = re_match_regexp (re_ctx_p, bc_p, iter, &sub_iter);
}
else
{
JERRY_DDLOG ("fail\n");
match_value = ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
}
if (!ecma_is_completion_value_throw (match_value))
{
if (ecma_is_value_true (match_value))
{
*out_iter_p = sub_iter;
}
else
{
JERRY_ASSERT (ecma_is_value_boolean (match_value));
/* restore saved */
memcpy (re_ctx_p->saved_p, saved_bck_p, size);
}
}
MEM_FINALIZE_LOCAL_ARRAY (saved_bck_p);
return match_value;
}
case RE_OP_CHAR_CLASS:
case RE_OP_INV_CHAR_CLASS:
{
uint32_t num_of_ranges;
bool is_match;
JERRY_DDLOG ("Execute RE_OP_CHAR_CLASS/RE_OP_INV_CHAR_CLASS, ");
if (lit_utf8_iterator_is_eos (&iter))
{
JERRY_DDLOG ("fail\n");
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
bool is_ignorecase = re_ctx_p->flags & RE_FLAG_IGNORE_CASE;
ecma_char_t curr_ch = re_canonicalize (lit_utf8_iterator_read_next (&iter), is_ignorecase);
num_of_ranges = re_get_value (&bc_p);
is_match = false;
while (num_of_ranges)
{
ecma_char_t ch1 = re_canonicalize ((ecma_char_t) re_get_value (&bc_p), is_ignorecase);
ecma_char_t ch2 = re_canonicalize ((ecma_char_t) re_get_value (&bc_p), is_ignorecase);
JERRY_DDLOG ("num_of_ranges=%d, ch1=%d, ch2=%d, curr_ch=%d; ",
num_of_ranges, ch1, ch2, curr_ch);
if (curr_ch >= ch1 && curr_ch <= ch2)
{
/* We must read all the ranges from bytecode. */
is_match = true;
}
num_of_ranges--;
}
if (op == RE_OP_CHAR_CLASS)
{
if (!is_match)
{
JERRY_DDLOG ("fail\n");
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
}
else
{
JERRY_ASSERT (op == RE_OP_INV_CHAR_CLASS);
if (is_match)
{
JERRY_DDLOG ("fail\n");
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
}
JERRY_DDLOG ("match\n");
break; /* tail merge */
}
case RE_OP_BACKREFERENCE:
{
uint32_t backref_idx;
backref_idx = re_get_value (&bc_p);
JERRY_DDLOG ("Execute RE_OP_BACKREFERENCE (idx: %d): ", backref_idx);
backref_idx *= 2; /* backref n -> saved indices [n*2, n*2+1] */
JERRY_ASSERT (backref_idx >= 2 && backref_idx + 1 < re_ctx_p->num_of_captures);
if (!re_ctx_p->saved_p[backref_idx].buf_p || !re_ctx_p->saved_p[backref_idx + 1].buf_p)
{
JERRY_DDLOG ("match\n");
break; /* capture is 'undefined', always matches! */
}
lit_utf8_iterator_t sub_iter = re_ctx_p->saved_p[backref_idx];
while (sub_iter.buf_pos.offset < re_ctx_p->saved_p[backref_idx + 1].buf_pos.offset)
{
ecma_char_t ch1, ch2;
if ((iter.buf_p + iter.buf_pos.offset) >= re_ctx_p->input_end_p)
{
JERRY_DDLOG ("fail\n");
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
ch1 = lit_utf8_iterator_read_next (&sub_iter);
ch2 = lit_utf8_iterator_read_next (&iter);
if (ch1 != ch2)
{
JERRY_DDLOG ("fail\n");
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
}
JERRY_DDLOG ("match\n");
break; /* tail merge */
}
case RE_OP_SAVE_AT_START:
{
re_bytecode_t *old_bc_p;
JERRY_DDLOG ("Execute RE_OP_SAVE_AT_START\n");
lit_utf8_iterator_t old_start_p = re_ctx_p->saved_p[RE_GLOBAL_START_IDX];
re_ctx_p->saved_p[RE_GLOBAL_START_IDX] = iter;
do
{
uint32_t offset = re_get_value (&bc_p);
lit_utf8_iterator_t sub_iter = lit_utf8_iterator_create (NULL, 0);
ecma_completion_value_t match_value = re_match_regexp (re_ctx_p, bc_p, iter, &sub_iter);
if (ecma_is_value_true (match_value))
{
*out_iter_p = sub_iter;
return match_value; /* match */
}
else if (ecma_is_completion_value_throw (match_value))
{
return match_value;
}
bc_p += offset;
old_bc_p = bc_p;
}
while (re_get_opcode (&bc_p) == RE_OP_ALTERNATIVE);
bc_p = old_bc_p;
re_ctx_p->saved_p[RE_GLOBAL_START_IDX] = old_start_p;
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
case RE_OP_SAVE_AND_MATCH:
{
JERRY_DDLOG ("End of pattern is reached: match\n");
re_ctx_p->saved_p[RE_GLOBAL_END_IDX] = iter;
*out_iter_p = iter;
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_TRUE); /* match */
}
case RE_OP_ALTERNATIVE:
{
/*
* Alternatives should be jump over, when alternative opcode appears.
*/
uint32_t offset = re_get_value (&bc_p);
JERRY_DDLOG ("Execute RE_OP_ALTERNATIVE");
bc_p += offset;
while (*bc_p == RE_OP_ALTERNATIVE)
{
JERRY_DDLOG (", jump: %d");
bc_p++;
offset = re_get_value (&bc_p);
bc_p += offset;
}
JERRY_DDLOG ("\n");
break; /* tail merge */
}
case RE_OP_CAPTURE_NON_GREEDY_ZERO_GROUP_START:
case RE_OP_NON_CAPTURE_NON_GREEDY_ZERO_GROUP_START:
{
/*
* On non-greedy iterations we have to execute the bytecode
* after the group first, if zero iteration is allowed.
*/
uint32_t start_idx, iter_idx, offset;
lit_utf8_iterator_t old_start = lit_utf8_iterator_create (NULL, 0);
lit_utf8_iterator_t sub_iter = lit_utf8_iterator_create (NULL, 0);
re_bytecode_t *old_bc_p;
old_bc_p = bc_p; /* save the bytecode start position of the group start */
start_idx = re_get_value (&bc_p);
offset = re_get_value (&bc_p);
if (RE_IS_CAPTURE_GROUP (op))
{
JERRY_ASSERT (start_idx <= re_ctx_p->num_of_captures / 2);
iter_idx = start_idx - 1;
start_idx *= 2;
old_start = re_ctx_p->saved_p[start_idx];
re_ctx_p->saved_p[start_idx] = iter;
}
else
{
JERRY_ASSERT (start_idx < re_ctx_p->num_of_non_captures);
iter_idx = start_idx + (re_ctx_p->num_of_captures / 2) - 1;
start_idx += re_ctx_p->num_of_captures;
}
re_ctx_p->num_of_iterations_p[iter_idx] = 0;
/* Jump all over to the end of the END opcode. */
bc_p += offset;
/* Try to match after the close paren if zero is allowed */
ecma_completion_value_t match_value = re_match_regexp (re_ctx_p, bc_p, iter, &sub_iter);
if (ecma_is_value_true (match_value))
{
*out_iter_p = sub_iter;
return match_value; /* match */
}
else if (ecma_is_completion_value_throw (match_value))
{
return match_value;
}
if (RE_IS_CAPTURE_GROUP (op))
{
re_ctx_p->saved_p[start_idx] = old_start;
}
bc_p = old_bc_p;
/* FALLTHRU */
}
case RE_OP_CAPTURE_GROUP_START:
case RE_OP_CAPTURE_GREEDY_ZERO_GROUP_START:
case RE_OP_NON_CAPTURE_GROUP_START:
case RE_OP_NON_CAPTURE_GREEDY_ZERO_GROUP_START:
{
uint32_t start_idx, iter_idx, old_iteration_cnt, offset;
lit_utf8_iterator_t sub_iter = lit_utf8_iterator_create (NULL, 0);
re_bytecode_t *old_bc_p;
re_bytecode_t *end_bc_p = NULL;
start_idx = re_get_value (&bc_p);
if (op != RE_OP_CAPTURE_GROUP_START
&& op != RE_OP_NON_CAPTURE_GROUP_START)
{
offset = re_get_value (&bc_p);
end_bc_p = bc_p + offset;
}
if (RE_IS_CAPTURE_GROUP (op))
{
JERRY_ASSERT (start_idx <= re_ctx_p->num_of_captures / 2);
iter_idx = start_idx - 1;
start_idx *= 2;
}
else
{
JERRY_ASSERT (start_idx < re_ctx_p->num_of_non_captures);
iter_idx = start_idx + (re_ctx_p->num_of_captures / 2) - 1;
start_idx += re_ctx_p->num_of_captures;
}
lit_utf8_iterator_t old_start = re_ctx_p->saved_p[start_idx];
old_iteration_cnt = re_ctx_p->num_of_iterations_p[iter_idx];
re_ctx_p->saved_p[start_idx] = iter;
re_ctx_p->num_of_iterations_p[iter_idx] = 0;
do
{
offset = re_get_value (&bc_p);
ecma_completion_value_t match_value = re_match_regexp (re_ctx_p, bc_p, iter, &sub_iter);
if (ecma_is_value_true (match_value))
{
*out_iter_p = sub_iter;
return match_value; /* match */
}
else if (ecma_is_completion_value_throw (match_value))
{
return match_value;
}
bc_p += offset;
old_bc_p = bc_p;
}
while (re_get_opcode (&bc_p) == RE_OP_ALTERNATIVE);
bc_p = old_bc_p;
re_ctx_p->num_of_iterations_p[iter_idx] = old_iteration_cnt;
/* Try to match after the close paren if zero is allowed. */
if (op == RE_OP_CAPTURE_GREEDY_ZERO_GROUP_START
|| op == RE_OP_NON_CAPTURE_GREEDY_ZERO_GROUP_START)
{
JERRY_ASSERT (end_bc_p);
ecma_completion_value_t match_value = re_match_regexp (re_ctx_p, end_bc_p, iter, &sub_iter);
if (ecma_is_value_true (match_value))
{
*out_iter_p = sub_iter;
return match_value; /* match */
}
else if (ecma_is_completion_value_throw (match_value))
{
return match_value;
}
}
re_ctx_p->saved_p[start_idx] = old_start;
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
case RE_OP_CAPTURE_NON_GREEDY_GROUP_END:
case RE_OP_NON_CAPTURE_NON_GREEDY_GROUP_END:
{
uint32_t end_idx, iter_idx, min, max;
re_bytecode_t *old_bc_p;
/*
* On non-greedy iterations we have to execute the bytecode
* after the group first. Try to iterate only if it fails.
*/
old_bc_p = bc_p; /* save the bytecode start position of the group end */
end_idx = re_get_value (&bc_p);
min = re_get_value (&bc_p);
max = re_get_value (&bc_p);
re_get_value (&bc_p); /* start offset */
if (RE_IS_CAPTURE_GROUP (op))
{
JERRY_ASSERT (end_idx <= re_ctx_p->num_of_captures / 2);
iter_idx = end_idx - 1;
end_idx = (end_idx * 2) + 1;
}
else
{
JERRY_ASSERT (end_idx <= re_ctx_p->num_of_non_captures);
iter_idx = end_idx + (re_ctx_p->num_of_captures / 2) - 1;
end_idx += re_ctx_p->num_of_captures;
}
re_ctx_p->num_of_iterations_p[iter_idx]++;
if (re_ctx_p->num_of_iterations_p[iter_idx] >= min
&& re_ctx_p->num_of_iterations_p[iter_idx] <= max)
{
lit_utf8_iterator_t old_end = re_ctx_p->saved_p[end_idx];
re_ctx_p->saved_p[end_idx] = iter;
lit_utf8_iterator_t sub_iter = lit_utf8_iterator_create (NULL, 0);
ecma_completion_value_t match_value = re_match_regexp (re_ctx_p, bc_p, iter, &sub_iter);
if (ecma_is_value_true (match_value))
{
*out_iter_p = sub_iter;
return match_value; /* match */
}
else if (ecma_is_completion_value_throw (match_value))
{
return match_value;
}
re_ctx_p->saved_p[end_idx] = old_end;
}
re_ctx_p->num_of_iterations_p[iter_idx]--;
bc_p = old_bc_p;
/* If non-greedy fails and try to iterate... */
/* FALLTHRU */
}
case RE_OP_CAPTURE_GREEDY_GROUP_END:
case RE_OP_NON_CAPTURE_GREEDY_GROUP_END:
{
uint32_t start_idx, end_idx, iter_idx, min, max, offset;
lit_utf8_iterator_t old_start = lit_utf8_iterator_create (NULL, 0);
lit_utf8_iterator_t old_end = lit_utf8_iterator_create (NULL, 0);
lit_utf8_iterator_t sub_iter = lit_utf8_iterator_create (NULL, 0);
re_bytecode_t *old_bc_p;
end_idx = re_get_value (&bc_p);
min = re_get_value (&bc_p);
max = re_get_value (&bc_p);
offset = re_get_value (&bc_p);
if (RE_IS_CAPTURE_GROUP (op))
{
JERRY_ASSERT (end_idx <= re_ctx_p->num_of_captures / 2);
iter_idx = end_idx - 1;
start_idx = end_idx * 2;
end_idx = start_idx + 1;
}
else
{
JERRY_ASSERT (end_idx <= re_ctx_p->num_of_non_captures);
iter_idx = end_idx + (re_ctx_p->num_of_captures / 2) - 1;
end_idx += re_ctx_p->num_of_captures;
start_idx = end_idx;
}
/* Check the empty iteration if the minimum number of iterations is reached. */
if (re_ctx_p->num_of_iterations_p[iter_idx] >= min
&& iter.buf_p == re_ctx_p->saved_p[start_idx].buf_p
&& iter.buf_pos.offset == re_ctx_p->saved_p[start_idx].buf_pos.offset)
{
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
re_ctx_p->num_of_iterations_p[iter_idx]++;
old_bc_p = bc_p; /* Save the bytecode end position of the END opcodes for matching after it. */
old_end = re_ctx_p->saved_p[end_idx];
re_ctx_p->saved_p[end_idx] = iter;
if (re_ctx_p->num_of_iterations_p[iter_idx] < max)
{
bc_p -= offset;
offset = re_get_value (&bc_p);
old_start = re_ctx_p->saved_p[start_idx];
re_ctx_p->saved_p[start_idx] = iter;
ecma_completion_value_t match_value = re_match_regexp (re_ctx_p, bc_p, iter, &sub_iter);
if (ecma_is_value_true (match_value))
{
*out_iter_p = sub_iter;
return match_value; /* match */
}
else if (ecma_is_completion_value_throw (match_value))
{
return match_value;
}
re_ctx_p->saved_p[start_idx] = old_start;
/* Try to match alternatives if any. */
bc_p += offset;
while (*bc_p == RE_OP_ALTERNATIVE)
{
bc_p++; /* RE_OP_ALTERNATIVE */
offset = re_get_value (&bc_p);
old_start = re_ctx_p->saved_p[start_idx];
re_ctx_p->saved_p[start_idx] = iter;
ecma_completion_value_t match_value = re_match_regexp (re_ctx_p, bc_p, iter, &sub_iter);
if (ecma_is_value_true (match_value))
{
*out_iter_p = sub_iter;
return match_value; /* match */
}
else if (ecma_is_completion_value_throw (match_value))
{
return match_value;
}
re_ctx_p->saved_p[start_idx] = old_start;
bc_p += offset;
}
}
if (re_ctx_p->num_of_iterations_p[iter_idx] >= min
&& re_ctx_p->num_of_iterations_p[iter_idx] <= max)
{
/* Try to match the rest of the bytecode. */
ecma_completion_value_t match_value = re_match_regexp (re_ctx_p, old_bc_p, iter, &sub_iter);
if (ecma_is_value_true (match_value))
{
*out_iter_p = sub_iter;
return match_value; /* match */
}
else if (ecma_is_completion_value_throw (match_value))
{
return match_value;
}
}
/* restore if fails */
re_ctx_p->saved_p[end_idx] = old_end;
re_ctx_p->num_of_iterations_p[iter_idx]--;
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
case RE_OP_NON_GREEDY_ITERATOR:
{
uint32_t min, max, offset, num_of_iter;
lit_utf8_iterator_t sub_iter = lit_utf8_iterator_create (NULL, 0);
min = re_get_value (&bc_p);
max = re_get_value (&bc_p);
offset = re_get_value (&bc_p);
JERRY_DDLOG ("Non-greedy iterator, min=%lu, max=%lu, offset=%ld\n",
(unsigned long) min, (unsigned long) max, (long) offset);
num_of_iter = 0;
while (num_of_iter <= max)
{
if (num_of_iter >= min)
{
ecma_completion_value_t match_value = re_match_regexp (re_ctx_p, bc_p + offset, iter, &sub_iter);
if (ecma_is_value_true (match_value))
{
*out_iter_p = sub_iter;
return match_value; /* match */
}
else if (ecma_is_completion_value_throw (match_value))
{
return match_value;
}
}
ecma_completion_value_t match_value = re_match_regexp (re_ctx_p, bc_p, iter, &sub_iter);
if (!ecma_is_value_true (match_value))
{
if (ecma_is_completion_value_throw (match_value))
{
return match_value;
}
break;
}
iter = sub_iter;
num_of_iter++;
}
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
case RE_OP_GREEDY_ITERATOR:
{
uint32_t min, max, offset, num_of_iter;
lit_utf8_iterator_t sub_iter = lit_utf8_iterator_create (NULL, 0);
min = re_get_value (&bc_p);
max = re_get_value (&bc_p);
offset = re_get_value (&bc_p);
JERRY_DDLOG ("Greedy iterator, min=%lu, max=%lu, offset=%ld\n",
(unsigned long) min, (unsigned long) max, (long) offset);
num_of_iter = 0;
while (num_of_iter < max)
{
ecma_completion_value_t match_value = re_match_regexp (re_ctx_p, bc_p, iter, &sub_iter);
if (!ecma_is_value_true (match_value))
{
if (ecma_is_completion_value_throw (match_value))
{
return match_value;
}
break;
}
iter = sub_iter;
num_of_iter++;
}
while (num_of_iter >= min)
{
ecma_completion_value_t match_value = re_match_regexp (re_ctx_p, bc_p + offset, iter, &sub_iter);
if (ecma_is_value_true (match_value))
{
*out_iter_p = sub_iter;
return match_value; /* match */
}
else if (ecma_is_completion_value_throw (match_value))
{
return match_value;
}
if (num_of_iter == min)
{
break;
}
lit_utf8_iterator_read_prev (&iter);
num_of_iter--;
}
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
}
default:
{
JERRY_DDLOG ("UNKNOWN opcode (%d)!\n", (uint32_t) op);
return ecma_make_throw_obj_completion_value (ecma_new_standard_error (ECMA_ERROR_COMMON));
}
}
}
JERRY_UNREACHABLE ();
return ecma_make_simple_completion_value (ECMA_SIMPLE_VALUE_FALSE); /* fail */
} /* regexp_match */
/**
* RegExp bytecode dumper
*/
void
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);
uint8_t *bytecode_p = (uint8_t *) (compiled_code_p + 1);
re_opcode_t op;
while ((op = re_get_opcode (&bytecode_p)))
{
switch (op)
{
case RE_OP_MATCH:
{
JERRY_DEBUG_MSG ("MATCH, ");
break;
}
case RE_OP_CHAR:
{
JERRY_DEBUG_MSG ("CHAR ");
JERRY_DEBUG_MSG ("%c, ", (char) re_get_char (&bytecode_p));
break;
}
case RE_OP_CAPTURE_NON_GREEDY_ZERO_GROUP_START:
{
JERRY_DEBUG_MSG ("N");
/* FALLTHRU */
}
case RE_OP_CAPTURE_GREEDY_ZERO_GROUP_START:
{
JERRY_DEBUG_MSG ("GZ_START ");
JERRY_DEBUG_MSG ("%d ", re_get_value (&bytecode_p));
JERRY_DEBUG_MSG ("%d ", re_get_value (&bytecode_p));
JERRY_DEBUG_MSG ("%d, ", re_get_value (&bytecode_p));
break;
}
case RE_OP_CAPTURE_GROUP_START:
{
JERRY_DEBUG_MSG ("START ");
JERRY_DEBUG_MSG ("%d ", re_get_value (&bytecode_p));
JERRY_DEBUG_MSG ("%d, ", re_get_value (&bytecode_p));
break;
}
case RE_OP_CAPTURE_NON_GREEDY_GROUP_END:
{
JERRY_DEBUG_MSG ("N");
/* FALLTHRU */
}
case RE_OP_CAPTURE_GREEDY_GROUP_END:
{
JERRY_DEBUG_MSG ("G_END ");
JERRY_DEBUG_MSG ("%d ", re_get_value (&bytecode_p));
JERRY_DEBUG_MSG ("%d ", re_get_value (&bytecode_p));
JERRY_DEBUG_MSG ("%d ", re_get_value (&bytecode_p));
JERRY_DEBUG_MSG ("%d, ", re_get_value (&bytecode_p));
break;
}
case RE_OP_NON_CAPTURE_NON_GREEDY_ZERO_GROUP_START:
{
JERRY_DEBUG_MSG ("N");
/* FALLTHRU */
}
case RE_OP_NON_CAPTURE_GREEDY_ZERO_GROUP_START:
{
JERRY_DEBUG_MSG ("GZ_NC_START ");
JERRY_DEBUG_MSG ("%d ", re_get_value (&bytecode_p));
JERRY_DEBUG_MSG ("%d ", re_get_value (&bytecode_p));
JERRY_DEBUG_MSG ("%d, ", re_get_value (&bytecode_p));
break;
}
case RE_OP_NON_CAPTURE_GROUP_START:
{
JERRY_DEBUG_MSG ("NC_START ");
JERRY_DEBUG_MSG ("%d ", re_get_value (&bytecode_p));
JERRY_DEBUG_MSG ("%d, ", re_get_value (&bytecode_p));
break;
}
case RE_OP_NON_CAPTURE_NON_GREEDY_GROUP_END:
{
JERRY_DEBUG_MSG ("N");
/* FALLTHRU */
}
case RE_OP_NON_CAPTURE_GREEDY_GROUP_END:
{
JERRY_DEBUG_MSG ("G_NC_END ");
JERRY_DEBUG_MSG ("%d ", re_get_value (&bytecode_p));
JERRY_DEBUG_MSG ("%d ", re_get_value (&bytecode_p));
JERRY_DEBUG_MSG ("%d ", re_get_value (&bytecode_p));
JERRY_DEBUG_MSG ("%d, ", re_get_value (&bytecode_p));
break;
}
case RE_OP_SAVE_AT_START:
{
JERRY_DEBUG_MSG ("RE_START ");
JERRY_DEBUG_MSG ("%d, ", re_get_value (&bytecode_p));
break;
}
case RE_OP_SAVE_AND_MATCH:
{
JERRY_DEBUG_MSG ("RE_END, ");
break;
}
case RE_OP_GREEDY_ITERATOR:
{
JERRY_DEBUG_MSG ("GREEDY_ITERATOR ");
JERRY_DEBUG_MSG ("%d ", re_get_value (&bytecode_p));
JERRY_DEBUG_MSG ("%d ", re_get_value (&bytecode_p));
JERRY_DEBUG_MSG ("%d, ", re_get_value (&bytecode_p));
break;
}
case RE_OP_NON_GREEDY_ITERATOR:
{
JERRY_DEBUG_MSG ("NON_GREEDY_ITERATOR ");
JERRY_DEBUG_MSG ("%d, ", re_get_value (&bytecode_p));
JERRY_DEBUG_MSG ("%d, ", re_get_value (&bytecode_p));
JERRY_DEBUG_MSG ("%d, ", re_get_value (&bytecode_p));
break;
}
case RE_OP_PERIOD:
{
JERRY_DEBUG_MSG ("PERIOD ");
break;
}
case RE_OP_ALTERNATIVE:
{
JERRY_DEBUG_MSG ("ALTERNATIVE ");
JERRY_DEBUG_MSG ("%d, ", re_get_value (&bytecode_p));
break;
}
case RE_OP_ASSERT_START:
{
JERRY_DEBUG_MSG ("ASSERT_START ");
break;
}
case RE_OP_ASSERT_END:
{
JERRY_DEBUG_MSG ("ASSERT_END ");
break;
}
case RE_OP_ASSERT_WORD_BOUNDARY:
{
JERRY_DEBUG_MSG ("ASSERT_WORD_BOUNDARY ");
break;
}
case RE_OP_ASSERT_NOT_WORD_BOUNDARY:
{
JERRY_DEBUG_MSG ("ASSERT_NOT_WORD_BOUNDARY ");
break;
}
case RE_OP_LOOKAHEAD_POS:
{
JERRY_DEBUG_MSG ("LOOKAHEAD_POS ");
JERRY_DEBUG_MSG ("%d, ", re_get_value (&bytecode_p));
break;
}
case RE_OP_LOOKAHEAD_NEG:
{
JERRY_DEBUG_MSG ("LOOKAHEAD_NEG ");
JERRY_DEBUG_MSG ("%d, ", re_get_value (&bytecode_p));
break;
}
case RE_OP_BACKREFERENCE:
{
JERRY_DEBUG_MSG ("BACKREFERENCE ");
JERRY_DEBUG_MSG ("%d, ", re_get_value (&bytecode_p));
break;
}
case RE_OP_INV_CHAR_CLASS:
{
JERRY_DEBUG_MSG ("INV_");
/* FALLTHRU */
}
case RE_OP_CHAR_CLASS:
{
JERRY_DEBUG_MSG ("CHAR_CLASS ");
uint32_t num_of_class = re_get_value (&bytecode_p);
JERRY_DEBUG_MSG ("%d", num_of_class);
while (num_of_class)
{
JERRY_DEBUG_MSG (" %d", re_get_char (&bytecode_p));
JERRY_DEBUG_MSG ("-%d", re_get_char (&bytecode_p));
num_of_class--;
}
JERRY_DEBUG_MSG (", ");
break;
}
default:
{
JERRY_DEBUG_MSG ("UNKNOWN(%d), ", (uint32_t) op);
break;
}
}
}
JERRY_DEBUG_MSG ("EOF\n");
} /* re_dump_bytecode */
