archive-gh-me/jerryscript
1
0
Fork 0
mirror of https://github.com/jerryscript-project/jerryscript.git synced 2025-12-15 16:29:21 +00:00
Code Issues Projects Releases Wiki Activity

Update lastIndex handling in RegExpBuiltinExec (#4010)

Browse Source 
JerryScript-DCO-1.0-Signed-off-by: Dániel Bátyai daniel.batyai@h-lab.eu
This commit is contained in:
Dániel Bátyai 2020-07-21 15:13:51 +02:00 committed by GitHub
parent 69f6bc6566
commit 1eef69f24d
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
5 changed files with 165 additions and 97 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

174
jerry-core/ecma/operations/ecma-regexp-object.c
View File

@ -1706,6 +1706,7 @@ ecma_regexp_cleanup_context (ecma_regexp_ctx_t *ctx_p) /**< regexp context */
*
* See also:
* ECMA-262 v5, 15.10.6.2
* ECMA-262 v11, 21.2.5.2.2
*
* @return array object - if matched
* null - otherwise
@ -1719,12 +1720,15 @@ ecma_regexp_exec_helper (ecma_object_t *regexp_object_p, /**< RegExp object */
{
ecma_value_t ret_value = ECMA_VALUE_EMPTY;
/* 1. */
JERRY_ASSERT (ecma_object_is_regexp_object (ecma_make_object_value (regexp_object_p)));
/* 9. */
ecma_extended_object_t *ext_object_p = (ecma_extended_object_t *) regexp_object_p;
re_compiled_code_t *bc_p = ECMA_GET_INTERNAL_VALUE_ANY_POINTER (re_compiled_code_t,
ext_object_p->u.class_prop.u.value);
/* 3. */
lit_utf8_size_t input_size;
lit_utf8_size_t input_length;
uint8_t input_flags = ECMA_STRING_FLAG_IS_ASCII;
@ -1735,47 +1739,48 @@ ecma_regexp_exec_helper (ecma_object_t *regexp_object_p, /**< RegExp object */
&input_flags);
const lit_utf8_byte_t *input_curr_p = input_buffer_p;
const lit_utf8_byte_t *input_end_p = input_buffer_p + input_size;
ecma_regexp_ctx_t re_ctx;
ecma_regexp_initialize_context (&re_ctx,
bc_p,
input_buffer_p,
input_end_p);
/* 4. */
uint32_t index = 0;
if (bc_p->header.status_flags & (RE_FLAG_GLOBAL | RE_FLAG_STICKY))
{
ecma_value_t lastindex_value = ecma_op_object_get_by_magic_id (regexp_object_p, LIT_MAGIC_STRING_LASTINDEX_UL);
ecma_value_t lastindex_value = ecma_op_object_get_by_magic_id (regexp_object_p, LIT_MAGIC_STRING_LASTINDEX_UL);
ecma_number_t lastindex_num;
ret_value = ecma_op_to_integer (lastindex_value, &lastindex_num);
ecma_free_value (lastindex_value);
if (ECMA_IS_VALUE_ERROR (ret_value))
{
goto cleanup_string;
}
/* TODO: Replace with ToLength */
if (lastindex_num < 0.0f)
{
#if ENABLED (JERRY_ESNEXT)
lastindex_num = 0.0f;
ret_value = ecma_op_to_length (lastindex_value, &index);
ecma_free_value (lastindex_value);
#else /* !ENABLED (JERRY_ESNEXT) */
lastindex_num = input_length + 1;
#endif /* ENABLED (JERRY_ESNEXT) */
}
index = ecma_number_to_uint32 (lastindex_num);
ecma_number_t lastindex_num = 0.0f;
ret_value = ecma_op_to_integer (lastindex_value, &lastindex_num);
ecma_free_value (lastindex_value);
/* The ToInteger operation might have thrown an error, however in that case lastindex_num will still be zero,
* and the error will be handled later after the uint32 coercion. */
if ((re_ctx.flags & RE_FLAG_GLOBAL) && lastindex_num < 0.0f)
{
JERRY_ASSERT (!ECMA_IS_VALUE_ERROR (ret_value));
goto fail_put_lastindex;
}
index = ecma_number_to_uint32 (lastindex_num);
#endif /* ENABLED (JERRY_ESNEXT) */
if (ECMA_IS_VALUE_ERROR (ret_value))
{
goto cleanup_context;
}
if (re_ctx.flags & (RE_FLAG_GLOBAL | RE_FLAG_STICKY))
{
/* 12.a */
if (index > input_length)
{
ret_value = ecma_op_object_put (regexp_object_p,
ecma_get_magic_string (LIT_MAGIC_STRING_LASTINDEX_UL),
ecma_make_integer_value (0),
true);
if (!ECMA_IS_VALUE_ERROR (ret_value))
{
JERRY_ASSERT (ecma_is_value_boolean (ret_value));
/* lastIndex is out of bounds, the match should fail. */
ret_value = ECMA_VALUE_NULL;
}
goto cleanup_string;
goto fail_put_lastindex;
}
if (index > 0)
@ -1793,18 +1798,19 @@ ecma_regexp_exec_helper (ecma_object_t *regexp_object_p, /**< RegExp object */
}
}
}
/* 8. */
else
{
index = 0;
}
const lit_utf8_byte_t *input_end_p = input_buffer_p + input_size;
ecma_regexp_ctx_t re_ctx;
ecma_regexp_initialize_context (&re_ctx,
bc_p,
input_buffer_p,
input_end_p);
/* 2. Try to match */
/* 9. */
uint8_t *bc_start_p = (uint8_t *) (bc_p + 1);
/* 11. */
const lit_utf8_byte_t *matched_p = NULL;
/* 12. */
JERRY_ASSERT (index <= input_length);
while (true)
{
@ -1812,58 +1818,37 @@ ecma_regexp_exec_helper (ecma_object_t *regexp_object_p, /**< RegExp object */
if (matched_p != NULL)
{
break;
goto match_found;
}
#if ENABLED (JERRY_ESNEXT)
/* 12.c.i */
if (re_ctx.flags & RE_FLAG_STICKY)
{
ecma_value_t put_result = ecma_op_object_put (regexp_object_p,
ecma_get_magic_string (LIT_MAGIC_STRING_LASTINDEX_UL),
ecma_make_uint32_value (0),
true);
if (ECMA_IS_VALUE_ERROR (put_result))
{
ret_value = put_result;
goto cleanup_context;
}
JERRY_ASSERT (ecma_is_value_boolean (put_result));
ret_value = ECMA_VALUE_NULL;
goto cleanup_context;
goto fail_put_lastindex;
}
#endif /* ENABLED (JERRY_ESNEXT) */
/* 12.a */
if (input_curr_p >= input_end_p)
{
if (re_ctx.flags & RE_FLAG_GLOBAL)
{
ecma_value_t put_result = ecma_op_object_put (regexp_object_p,
ecma_get_magic_string (LIT_MAGIC_STRING_LASTINDEX_UL),
ecma_make_uint32_value (0),
true);
if (ECMA_IS_VALUE_ERROR (put_result))
{
ret_value = put_result;
goto cleanup_context;
}
JERRY_ASSERT (ecma_is_value_boolean (put_result));
goto fail_put_lastindex;
}
/* Failed to match, return 'null'. */
ret_value = ECMA_VALUE_NULL;
goto cleanup_context;
goto match_failed;
}
JERRY_ASSERT (input_curr_p < input_end_p);
/* 12.c.ii */
index++;
#if ENABLED (JERRY_ESNEXT)
if (re_ctx.flags & RE_FLAG_UNICODE)
{
index++;
const lit_code_point_t cp = ecma_regexp_unicode_advance (&input_curr_p,
input_end_p);
const lit_code_point_t cp = ecma_regexp_unicode_advance (&input_curr_p, input_end_p);
if (cp > LIT_UTF16_CODE_UNIT_MAX)
{
@ -1874,10 +1859,33 @@ ecma_regexp_exec_helper (ecma_object_t *regexp_object_p, /**< RegExp object */
}
#endif /* ENABLED (JERRY_ESNEXT) */
index++;
lit_utf8_incr (&input_curr_p);
}
JERRY_UNREACHABLE ();
fail_put_lastindex:
/* We should only get here if the regexp is global or sticky */
JERRY_ASSERT ((re_ctx.flags & (RE_FLAG_GLOBAL | RE_FLAG_STICKY)) != 0);
ret_value = ecma_op_object_put (regexp_object_p,
ecma_get_magic_string (LIT_MAGIC_STRING_LASTINDEX_UL),
ecma_make_integer_value (0),
true);
if (ECMA_IS_VALUE_ERROR (ret_value))
{
goto cleanup_context;
}
JERRY_ASSERT (ecma_is_value_boolean (ret_value));
match_failed:
/* 12.a.ii */
ret_value = ECMA_VALUE_NULL;
goto cleanup_context;
match_found:
JERRY_ASSERT (matched_p != NULL);
if (ECMA_RE_STACK_LIMIT_REACHED (matched_p))
@ -1886,10 +1894,12 @@ ecma_regexp_exec_helper (ecma_object_t *regexp_object_p, /**< RegExp object */
goto cleanup_context;
}
/* 15. */
if (re_ctx.flags & (RE_FLAG_GLOBAL | RE_FLAG_STICKY))
{
JERRY_ASSERT (index <= input_length);
/* 13-14. */
lit_utf8_size_t match_length;
const lit_utf8_byte_t *match_begin_p = re_ctx.captures_p[0].begin_p;
const lit_utf8_byte_t *match_end_p = re_ctx.captures_p[0].end_p;
@ -1904,25 +1914,25 @@ ecma_regexp_exec_helper (ecma_object_t *regexp_object_p, /**< RegExp object */
(lit_utf8_size_t) (match_end_p - match_begin_p));
}
ecma_value_t put_result = ecma_op_object_put (regexp_object_p,
ecma_get_magic_string (LIT_MAGIC_STRING_LASTINDEX_UL),
ecma_make_uint32_value (index + match_length),
true);
if (ECMA_IS_VALUE_ERROR (put_result))
ret_value = ecma_op_object_put (regexp_object_p,
ecma_get_magic_string (LIT_MAGIC_STRING_LASTINDEX_UL),
ecma_make_uint32_value (index + match_length),
true);
if (ECMA_IS_VALUE_ERROR (ret_value))
{
ret_value = put_result;
goto cleanup_context;
}
JERRY_ASSERT (ecma_is_value_boolean (put_result));
JERRY_ASSERT (ecma_is_value_boolean (ret_value));
}
/* 16-27. */
ret_value = ecma_regexp_create_result_object (&re_ctx, input_string_p, index);
cleanup_context:
ecma_regexp_cleanup_context (&re_ctx);
cleanup_string:
if (input_flags & ECMA_STRING_FLAG_MUST_BE_FREED)
{
jmem_heap_free_block ((void *) input_buffer_p, input_size);

49
tests/jerry/es.next/regexp-lastindex.js
View File

@ -12,11 +12,56 @@
// See the License for the specific language governing permissions and
// limitations under the License.
var t = new RegExp ("abc","g");
t.lastIndex = -12;
var t = /abc/g;
t.lastIndex = -12.5;
result = t.exec("abc abc");
assert(result[0] === "abc");
assert(result.index === 0);
assert(t.lastIndex === 3);
assert(RegExp.prototype.lastIndex === undefined)
var r = /./y
r.lastIndex = -1;
assert (JSON.stringify(r.exec("abca")) === '["a"]');
assert (r.lastIndex === 1);
assert (JSON.stringify(r.exec("abca")) === '["b"]');
assert (r.lastIndex === 2);
r.lastIndex = 5;
assert (JSON.stringify(r.exec("abca")) === 'null');
assert (r.lastIndex === 0);
var r = /a/y
assert (JSON.stringify(r.exec("abca")) === '["a"]');
assert (r.lastIndex === 1);
assert (JSON.stringify(r.exec("abca")) === 'null');
assert (r.lastIndex === 0);
var r = /./g
r.lastIndex = -1;
assert (JSON.stringify(r.exec("abca")) === '["a"]');
assert (r.lastIndex === 1);
assert (JSON.stringify(r.exec("abca")) === '["b"]');
assert (r.lastIndex === 2);
r.lastIndex = 5;
assert (JSON.stringify(r.exec("abca")) === 'null');
assert (r.lastIndex === 0);
var r = /a/g
assert (JSON.stringify(r.exec("abca")) === '["a"]');
assert (r.lastIndex === 1);
assert (JSON.stringify(r.exec("abca")) === '["a"]');
assert (r.lastIndex === 4);
assert (JSON.stringify(r.exec("abca")) === 'null');
assert (r.lastIndex === 0);
var r = /./uim
r.lastIndex = 2;
assert (JSON.stringify(r.exec("abcd")) === '["a"]');
assert (r.lastIndex === 2);
r.lastIndex = "lastIndex";
assert (JSON.stringify(r.exec("abcd")) === '["a"]');
assert (r.lastIndex === "lastIndex");

4
tests/jerry/es5.1/regexp-lastindex.js
View File

@ -12,8 +12,8 @@
// See the License for the specific language governing permissions and
// limitations under the License.
var t = new RegExp ("abc","g");
t.lastIndex = -12;
var t = /abc/g;
t.lastIndex = -12.5;
result = t.exec("abc abc");
assert(!result);
assert(t.lastIndex === 0);

14
tests/jerry/regexp-routines.js
View File

@ -101,3 +101,17 @@ assert (re2.ignoreCase === re1.ignoreCase);
assert (re2.multiline === re1.multiline);
assert (re2.source === re1.source);
assert (re2.lastIndex === 0);
var r = /./
r.lastIndex = {
valueOf: function() {
throw "abrupt lastIndex"
}
}
try {
r.exec("a");
assert(false);
} catch (e) {
assert (e === "abrupt lastIndex");
}

21
tests/test262-es6-excludelist.xml
View File

@ -82,16 +82,15 @@
<test id="built-ins/Reflect/enumerate/return-iterator.js"><reason></reason></test>
<test id="built-ins/RegExp/prototype/exec/get-sticky-coerce.js"><reason>RegExpBuiltinExec no longer uses accessors to get flags</reason></test>
<test id="built-ins/RegExp/prototype/exec/get-sticky-err.js"><reason>RegExpBuiltinExec no longer uses accessors to get flags</reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-coerce-global.js"><reason></reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-coerce-lastindex-err.js"><reason></reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-coerce-sticky.js"><reason></reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-failure-set-lastindex-err.js"><reason></reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-failure-set-lastindex.js"><reason></reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-get-global-err.js"><reason></reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-get-sticky-err.js"><reason></reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-success-g-set-lastindex-err.js"><reason></reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-success-g-set-lastindex.js"><reason></reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/coerce-global.js"><reason></reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-coerce-global.js"><reason>RegExpBuiltinExec no longer uses accessors to get flags</reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-coerce-sticky.js"><reason>RegExpBuiltinExec no longer uses accessors to get flags</reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-failure-set-lastindex-err.js"><reason>lastIndex no longer needs to be updated if not sticky or global</reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-failure-set-lastindex.js"><reason>lastIndex no longer needs to be updated if not sticky or global</reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-get-global-err.js"><reason>RegExpBuiltinExec no longer uses accessors to get flags</reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-get-sticky-err.js"><reason>RegExpBuiltinExec no longer uses accessors to get flags</reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-success-g-set-lastindex-err.js"><reason>RegExpBuiltinExec no longer uses accessors to get flags</reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/builtin-success-g-set-lastindex.js"><reason>RegExpBuiltinExec no longer uses accessors to get flags</reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/coerce-global.js"><reason>RegExpBuiltinExec no longer uses accessors to get flags</reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/coerce-sticky.js"><reason>RegExpBuiltinExec no longer uses accessors to get flags</reason></test>
<test id="built-ins/RegExp/prototype/Symbol.match/get-sticky-err.js"><reason>RegExpBuiltinExec no longer uses accessors to get flags</reason></test>
<test id="built-ins/RegExp/prototype/Symbol.replace/get-sticky-coerce.js"><reason>RegExpBuiltinExec no longer uses accessors to get flags</reason></test>
@ -99,7 +98,7 @@
<test id="built-ins/RegExp/prototype/Symbol.search/get-sticky-coerce.js"><reason>RegExpBuiltinExec no longer uses accessors to get flags</reason></test>
<test id="built-ins/RegExp/prototype/Symbol.search/get-sticky-err.js"><reason>RegExpBuiltinExec no longer uses accessors to get flags</reason></test>
<test id="built-ins/RegExp/prototype/Symbol.split/coerce-limit.js"><reason>Test is outdated: ES11, 21.2.5.13 13</reason></test>
<test id="built-ins/RegExp/prototype/test/get-sticky-err.js"><reason></reason></test>
<test id="built-ins/RegExp/prototype/test/get-sticky-err.js"><reason>RegExpBuiltinExec no longer uses accessors to get flags</reason></test>
<test id="built-ins/String/prototype/normalize/form-is-not-valid-throws.js"><reason></reason></test>
<test id="built-ins/String/prototype/normalize/length.js"><reason></reason></test>
<test id="built-ins/String/prototype/normalize/name.js"><reason></reason></test>