fries/quickjs
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Implement RegExp 'd' flag (#86)

Browse source
This commit is contained in:
Ben Noordhuis 2023-11-20 09:45:44 +01:00 committed by GitHub
parent e2bc6441f8
commit d1960d1bfe
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
5 changed files with 152 additions and 62 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

11
libregexp.c
View file

@ -2582,25 +2582,26 @@ void *lre_realloc(void *opaque, void *ptr, size_t size)
int main(int argc, char **argv) int main(int argc, char **argv)
{ {
int len, ret, i; int len, flags, ret, i;
uint8_t *bc; uint8_t *bc;
char error_msg[64]; char error_msg[64];
uint8_t *capture[CAPTURE_COUNT_MAX * 2]; uint8_t *capture[CAPTURE_COUNT_MAX * 2];
const char *input; const char *input;
int input_len, capture_count; int input_len, capture_count;
if (argc < 3) { if (argc < 4) {
printf("usage: %s regexp input\n", argv[0]); printf("usage: %s regexp flags input\n", argv[0]);
exit(1); exit(1);
} }
flags = atoi(argv[2]);
bc = lre_compile(&len, error_msg, sizeof(error_msg), argv[1], bc = lre_compile(&len, error_msg, sizeof(error_msg), argv[1],
strlen(argv[1]), 0, NULL); strlen(argv[1]), flags, NULL);
if (!bc) { if (!bc) {
fprintf(stderr, "error: %s\n", error_msg); fprintf(stderr, "error: %s\n", error_msg);
exit(1); exit(1);
} }
input = argv[2]; input = argv[3];
input_len = strlen(input); input_len = strlen(input);
ret = lre_exec(capture, bc, (uint8_t *)input, 0, input_len, 0, NULL); ret = lre_exec(capture, bc, (uint8_t *)input, 0, input_len, 0, NULL);

1
libregexp.h
View file

@ -36,6 +36,7 @@
#define LRE_FLAG_DOTALL (1 << 3) #define LRE_FLAG_DOTALL (1 << 3)
#define LRE_FLAG_UTF16 (1 << 4) #define LRE_FLAG_UTF16 (1 << 4)
#define LRE_FLAG_STICKY (1 << 5) #define LRE_FLAG_STICKY (1 << 5)
#define LRE_FLAG_INDICES (1 << 6) /* Unused by libregexp, just recorded. */
#define LRE_FLAG_NAMED_GROUPS (1 << 7) /* named groups are present in the regexp */ #define LRE_FLAG_NAMED_GROUPS (1 << 7) /* named groups are present in the regexp */

1
quickjs-atom.h
View file

@ -166,6 +166,7 @@ DEF(revoke, "revoke")
DEF(async, "async") DEF(async, "async")
DEF(exec, "exec") DEF(exec, "exec")
DEF(groups, "groups") DEF(groups, "groups")
DEF(indices, "indices")
DEF(status, "status") DEF(status, "status")
DEF(reason, "reason") DEF(reason, "reason")
DEF(globalThis, "globalThis") DEF(globalThis, "globalThis")

199
quickjs.c
View file

@ -1151,6 +1151,11 @@ static const JSClassExoticMethods js_proxy_exotic_methods;
static const JSClassExoticMethods js_module_ns_exotic_methods; static const JSClassExoticMethods js_module_ns_exotic_methods;
static JSClassID js_class_id_alloc = JS_CLASS_INIT_COUNT; static JSClassID js_class_id_alloc = JS_CLASS_INIT_COUNT;
static JSValue js_int32(int32_t v)
{
return JS_MKVAL(JS_TAG_INT, v);
}
static void js_trigger_gc(JSRuntime *rt, size_t size) static void js_trigger_gc(JSRuntime *rt, size_t size)
{ {
BOOL force_gc; BOOL force_gc;
@ -39831,6 +39836,9 @@ static JSValue js_compile_regexp(JSContext *ctx, JSValueConst pattern,
/* XXX: re_flags = LRE_FLAG_OCTAL unless strict mode? */ /* XXX: re_flags = LRE_FLAG_OCTAL unless strict mode? */
for (i = 0; i < len; i++) { for (i = 0; i < len; i++) {
switch(str[i]) { switch(str[i]) {
case 'd':
mask = LRE_FLAG_INDICES;
break;
case 'g': case 'g':
mask = LRE_FLAG_GLOBAL; mask = LRE_FLAG_GLOBAL;
break; break;
@ -40153,6 +40161,11 @@ static JSValue js_regexp_get_flags(JSContext *ctx, JSValueConst this_val)
if (JS_VALUE_GET_TAG(this_val) != JS_TAG_OBJECT) if (JS_VALUE_GET_TAG(this_val) != JS_TAG_OBJECT)
return JS_ThrowTypeErrorNotAnObject(ctx); return JS_ThrowTypeErrorNotAnObject(ctx);
res = JS_ToBoolFree(ctx, JS_GetPropertyStr(ctx, this_val, "hasIndices"));
if (res < 0)
goto exception;
if (res)
*p++ = 'd';
res = JS_ToBoolFree(ctx, JS_GetProperty(ctx, this_val, JS_ATOM_global)); res = JS_ToBoolFree(ctx, JS_GetProperty(ctx, this_val, JS_ATOM_global));
if (res < 0) if (res < 0)
goto exception; goto exception;
@ -40232,25 +40245,32 @@ static JSValue js_regexp_exec(JSContext *ctx, JSValueConst this_val,
{ {
JSRegExp *re = js_get_regexp(ctx, this_val, TRUE); JSRegExp *re = js_get_regexp(ctx, this_val, TRUE);
JSString *str; JSString *str;
JSValue str_val, obj, val, groups = JS_UNDEFINED; JSValue t, ret, str_val, obj, val, groups;
JSValue indices, indices_groups;
uint8_t *re_bytecode; uint8_t *re_bytecode;
int ret;
uint8_t **capture, *str_buf; uint8_t **capture, *str_buf;
int capture_count, shift, i, re_flags; int rc, capture_count, shift, i, re_flags;
int64_t last_index; int64_t last_index;
const char *group_name_ptr; const char *group_name_ptr;
if (!re) if (!re)
return JS_EXCEPTION; return JS_EXCEPTION;
str_val = JS_ToString(ctx, argv[0]); str_val = JS_ToString(ctx, argv[0]);
if (JS_IsException(str_val)) if (JS_IsException(str_val))
return str_val;
val = JS_GetProperty(ctx, this_val, JS_ATOM_lastIndex);
if (JS_IsException(val) ||
JS_ToLengthFree(ctx, &last_index, val)) {
JS_FreeValue(ctx, str_val);
return JS_EXCEPTION; return JS_EXCEPTION;
}
ret = JS_EXCEPTION;
obj = JS_NULL;
groups = JS_UNDEFINED;
indices = JS_UNDEFINED;
indices_groups = JS_UNDEFINED;
capture = NULL;
val = JS_GetProperty(ctx, this_val, JS_ATOM_lastIndex);
if (JS_IsException(val) || JS_ToLengthFree(ctx, &last_index, val))
goto fail;
re_bytecode = re->bytecode->u.str8; re_bytecode = re->bytecode->u.str8;
re_flags = lre_get_flags(re_bytecode); re_flags = lre_get_flags(re_bytecode);
if ((re_flags & (LRE_FLAG_GLOBAL | LRE_FLAG_STICKY)) == 0) { if ((re_flags & (LRE_FLAG_GLOBAL | LRE_FLAG_STICKY)) == 0) {
@ -40258,27 +40278,23 @@ static JSValue js_regexp_exec(JSContext *ctx, JSValueConst this_val,
} }
str = JS_VALUE_GET_STRING(str_val); str = JS_VALUE_GET_STRING(str_val);
capture_count = lre_get_capture_count(re_bytecode); capture_count = lre_get_capture_count(re_bytecode);
capture = NULL;
if (capture_count > 0) { if (capture_count > 0) {
capture = js_malloc(ctx, sizeof(capture[0]) * capture_count * 2); capture = js_malloc(ctx, sizeof(capture[0]) * capture_count * 2);
if (!capture) { if (!capture)
JS_FreeValue(ctx, str_val); goto fail;
return JS_EXCEPTION;
}
} }
shift = str->is_wide_char; shift = str->is_wide_char;
str_buf = str->u.str8; str_buf = str->u.str8;
if (last_index > str->len) { if (last_index > str->len) {
ret = 2; rc = 2;
} else { } else {
ret = lre_exec(capture, re_bytecode, rc = lre_exec(capture, re_bytecode,
str_buf, last_index, str->len, str_buf, last_index, str->len,
shift, ctx); shift, ctx);
} }
obj = JS_NULL; if (rc != 1) {
if (ret != 1) { if (rc >= 0) {
if (ret >= 0) { if (rc == 2 || (re_flags & (LRE_FLAG_GLOBAL | LRE_FLAG_STICKY))) {
if (ret == 2 || (re_flags & (LRE_FLAG_GLOBAL | LRE_FLAG_STICKY))) {
if (JS_SetProperty(ctx, this_val, JS_ATOM_lastIndex, if (JS_SetProperty(ctx, this_val, JS_ATOM_lastIndex,
JS_NewInt32(ctx, 0)) < 0) JS_NewInt32(ctx, 0)) < 0)
goto fail; goto fail;
@ -40287,7 +40303,6 @@ static JSValue js_regexp_exec(JSContext *ctx, JSValueConst this_val,
JS_ThrowInternalError(ctx, "out of memory in regexp execution"); JS_ThrowInternalError(ctx, "out of memory in regexp execution");
goto fail; goto fail;
} }
JS_FreeValue(ctx, str_val);
} else { } else {
int prop_flags; int prop_flags;
if (re_flags & (LRE_FLAG_GLOBAL | LRE_FLAG_STICKY)) { if (re_flags & (LRE_FLAG_GLOBAL | LRE_FLAG_STICKY)) {
@ -40305,52 +40320,123 @@ static JSValue js_regexp_exec(JSContext *ctx, JSValueConst this_val,
if (JS_IsException(groups)) if (JS_IsException(groups))
goto fail; goto fail;
} }
if (re_flags & LRE_FLAG_INDICES) {
for(i = 0; i < capture_count; i++) { indices = JS_NewArray(ctx);
int start, end; if (JS_IsException(indices))
JSValue val; goto fail;
if (capture[2 * i] == NULL || if (group_name_ptr) {
capture[2 * i + 1] == NULL) { indices_groups = JS_NewObjectProto(ctx, JS_NULL);
val = JS_UNDEFINED; if (JS_IsException(indices_groups))
} else {
start = (capture[2 * i] - str_buf) >> shift;
end = (capture[2 * i + 1] - str_buf) >> shift;
val = js_sub_string(ctx, str, start, end);
if (JS_IsException(val))
goto fail; goto fail;
} }
}
for(i = 0; i < capture_count; i++) {
const char *name = NULL;
uint8_t **match = &capture[2 * i];
int start = -1;
int end = -1;
if (group_name_ptr && i > 0) { if (group_name_ptr && i > 0) {
if (*group_name_ptr) { if (*group_name_ptr) name = group_name_ptr;
if (JS_DefinePropertyValueStr(ctx, groups, group_name_ptr, group_name_ptr += strlen(group_name_ptr) + 1;
JS_DupValue(ctx, val), }
prop_flags) < 0) {
if (match[0] && match[1]) {
start = (match[0] - str_buf) >> shift;
end = (match[1] - str_buf) >> shift;
}
if (!JS_IsUndefined(indices)) {
JSValue val = JS_UNDEFINED;
if (start != -1) {
val = JS_NewArray(ctx);
if (JS_IsException(val))
goto fail;
if (JS_DefinePropertyValueUint32(ctx, val, 0,
js_int32(start),
prop_flags) < 0) {
JS_FreeValue(ctx, val);
goto fail;
}
if (JS_DefinePropertyValueUint32(ctx, val, 1,
js_int32(end),
prop_flags) < 0) {
JS_FreeValue(ctx, val); JS_FreeValue(ctx, val);
goto fail; goto fail;
} }
} }
group_name_ptr += strlen(group_name_ptr) + 1; if (name && !JS_IsUndefined(indices_groups)) {
val = JS_DupValue(ctx, val);
if (JS_DefinePropertyValueStr(ctx, indices_groups,
name, val, prop_flags) < 0) {
JS_FreeValue(ctx, val);
goto fail;
}
}
if (JS_DefinePropertyValueUint32(ctx, indices, i, val,
prop_flags) < 0) {
goto fail;
}
} }
JSValue val = JS_UNDEFINED;
if (start != -1) {
val = js_sub_string(ctx, str, start, end);
if (JS_IsException(val))
goto fail;
}
if (name) {
if (JS_DefinePropertyValueStr(ctx, groups, name,
JS_DupValue(ctx, val),
prop_flags) < 0) {
JS_FreeValue(ctx, val);
goto fail;
}
}
if (JS_DefinePropertyValueUint32(ctx, obj, i, val, prop_flags) < 0) if (JS_DefinePropertyValueUint32(ctx, obj, i, val, prop_flags) < 0)
goto fail; goto fail;
} }
t = groups, groups = JS_UNDEFINED;
if (JS_DefinePropertyValue(ctx, obj, JS_ATOM_groups, if (JS_DefinePropertyValue(ctx, obj, JS_ATOM_groups,
groups, prop_flags) < 0) t, prop_flags) < 0) {
goto fail; goto fail;
if (JS_DefinePropertyValue(ctx, obj, JS_ATOM_index, }
JS_NewInt32(ctx, (capture[0] - str_buf) >> shift), prop_flags) < 0)
t = js_int32((capture[0] - str_buf) >> shift);
if (JS_DefinePropertyValue(ctx, obj, JS_ATOM_index, t, prop_flags) < 0)
goto fail; goto fail;
if (JS_DefinePropertyValue(ctx, obj, JS_ATOM_input, str_val, prop_flags) < 0)
goto fail1; t = str_val, str_val = JS_UNDEFINED;
if (JS_DefinePropertyValue(ctx, obj, JS_ATOM_input, t, prop_flags) < 0)
goto fail;
if (!JS_IsUndefined(indices)) {
t = indices_groups, indices_groups = JS_UNDEFINED;
if (JS_DefinePropertyValue(ctx, indices, JS_ATOM_groups,
t, prop_flags) < 0) {
goto fail;
}
t = indices, indices = JS_UNDEFINED;
if (JS_DefinePropertyValue(ctx, obj, JS_ATOM_indices,
t, prop_flags) < 0) {
goto fail;
}
}
} }
js_free(ctx, capture); ret = obj;
return obj; obj = JS_UNDEFINED;
fail: fail:
JS_FreeValue(ctx, groups); JS_FreeValue(ctx, indices_groups);
JS_FreeValue(ctx, indices);
JS_FreeValue(ctx, str_val); JS_FreeValue(ctx, str_val);
fail1: JS_FreeValue(ctx, groups);
JS_FreeValue(ctx, obj); JS_FreeValue(ctx, obj);
js_free(ctx, capture); js_free(ctx, capture);
return JS_EXCEPTION; return ret;
} }
/* delete portions of a string that match a given regex */ /* delete portions of a string that match a given regex */
@ -41185,12 +41271,13 @@ static const JSCFunctionListEntry js_regexp_funcs[] = {
static const JSCFunctionListEntry js_regexp_proto_funcs[] = { static const JSCFunctionListEntry js_regexp_proto_funcs[] = {
JS_CGETSET_DEF("flags", js_regexp_get_flags, NULL ), JS_CGETSET_DEF("flags", js_regexp_get_flags, NULL ),
JS_CGETSET_DEF("source", js_regexp_get_source, NULL ), JS_CGETSET_DEF("source", js_regexp_get_source, NULL ),
JS_CGETSET_MAGIC_DEF("global", js_regexp_get_flag, NULL, 1 ), JS_CGETSET_MAGIC_DEF("global", js_regexp_get_flag, NULL, LRE_FLAG_GLOBAL ),
JS_CGETSET_MAGIC_DEF("ignoreCase", js_regexp_get_flag, NULL, 2 ), JS_CGETSET_MAGIC_DEF("ignoreCase", js_regexp_get_flag, NULL, LRE_FLAG_IGNORECASE ),
JS_CGETSET_MAGIC_DEF("multiline", js_regexp_get_flag, NULL, 4 ), JS_CGETSET_MAGIC_DEF("multiline", js_regexp_get_flag, NULL, LRE_FLAG_MULTILINE ),
JS_CGETSET_MAGIC_DEF("dotAll", js_regexp_get_flag, NULL, 8 ), JS_CGETSET_MAGIC_DEF("dotAll", js_regexp_get_flag, NULL, LRE_FLAG_DOTALL ),
JS_CGETSET_MAGIC_DEF("unicode", js_regexp_get_flag, NULL, 16 ), JS_CGETSET_MAGIC_DEF("unicode", js_regexp_get_flag, NULL, LRE_FLAG_UTF16 ),
JS_CGETSET_MAGIC_DEF("sticky", js_regexp_get_flag, NULL, 32 ), JS_CGETSET_MAGIC_DEF("sticky", js_regexp_get_flag, NULL, LRE_FLAG_STICKY ),
JS_CGETSET_MAGIC_DEF("hasIndices", js_regexp_get_flag, NULL, LRE_FLAG_INDICES ),
JS_CFUNC_DEF("exec", 1, js_regexp_exec ), JS_CFUNC_DEF("exec", 1, js_regexp_exec ),
JS_CFUNC_DEF("compile", 2, js_regexp_compile ), JS_CFUNC_DEF("compile", 2, js_regexp_compile ),
JS_CFUNC_DEF("test", 1, js_regexp_test ), JS_CFUNC_DEF("test", 1, js_regexp_test ),

2
test262.conf
View file

@ -154,7 +154,7 @@ Reflect.setPrototypeOf
regexp-dotall regexp-dotall
regexp-duplicate-named-groups=skip regexp-duplicate-named-groups=skip
regexp-lookbehind regexp-lookbehind
regexp-match-indices=skip regexp-match-indices
regexp-named-groups regexp-named-groups
regexp-unicode-property-escapes regexp-unicode-property-escapes
regexp-v-flag=skip regexp-v-flag=skip