Welcome! Log In Create A New Profile

Advanced

[njs] Improved memory footprint of RegExp.prototype.test().

Dmitry Volyntsev
October 04, 2023 04:56PM
details: https://hg.nginx.org/njs/rev/a6410aad6eab
branches:
changeset: 2213:a6410aad6eab
user: Dmitry Volyntsev <xeioex@nginx.com>
date: Tue Oct 03 00:39:04 2023 -0700
description:
Improved memory footprint of RegExp.prototype.test().

diffstat:

src/njs_regexp.c | 102 ++++++++++++++++++++++++++++++++++++++----------------
src/njs_regexp.h | 2 -
2 files changed, 71 insertions(+), 33 deletions(-)

diffs (228 lines):

diff -r f076da3c7a6f -r a6410aad6eab src/njs_regexp.c
--- a/src/njs_regexp.c Fri Sep 29 21:41:34 2023 -0700
+++ b/src/njs_regexp.c Tue Oct 03 00:39:04 2023 -0700
@@ -25,6 +25,9 @@ static u_char *njs_regexp_compile_trace_
njs_trace_data_t *td, u_char *start);
static u_char *njs_regexp_match_trace_handler(njs_trace_t *trace,
njs_trace_data_t *td, u_char *start);
+#define NJS_REGEXP_FLAG_TEST 1
+static njs_int_t njs_regexp_exec(njs_vm_t *vm, njs_value_t *r, njs_value_t *s,
+ unsigned flags, njs_value_t *retval);
static njs_array_t *njs_regexp_exec_result(njs_vm_t *vm, njs_value_t *r,
njs_utf8_t utf8, njs_string_prop_t *string, njs_regex_match_data_t *data);
static njs_int_t njs_regexp_string_create(njs_vm_t *vm, njs_value_t *value,
@@ -849,7 +852,7 @@ njs_regexp_prototype_test(njs_vm_t *vm,
return NJS_ERROR;
}

- ret = njs_regexp_exec(vm, r, string, &value);
+ ret = njs_regexp_exec(vm, r, string, NJS_REGEXP_FLAG_TEST, &value);
if (njs_slow_path(ret != NJS_OK)) {
return NJS_ERROR;
}
@@ -865,10 +868,11 @@ njs_regexp_prototype_test(njs_vm_t *vm,
*/
static njs_int_t
njs_regexp_builtin_exec(njs_vm_t *vm, njs_value_t *r, njs_value_t *s,
- njs_value_t *retval)
+ unsigned flags, njs_value_t *retval)
{
- size_t length, offset;
+ size_t c, length, offset;
int64_t last_index;
+ uint32_t index;
njs_int_t ret;
njs_utf8_t utf8;
njs_value_t value;
@@ -940,6 +944,32 @@ njs_regexp_builtin_exec(njs_vm_t *vm, nj
ret = njs_regexp_match(vm, &pattern->regex[type], string.start, offset,
string.size, match_data);
if (ret >= 0) {
+ if (pattern->global || pattern->sticky) {
+ c = njs_regex_capture(match_data, 1);
+
+ if (utf8 == NJS_STRING_UTF8) {
+ index = njs_string_index(&string, c);
+
+ } else {
+ index = c;
+ }
+
+ njs_set_number(&value, index);
+ ret = njs_value_property_set(vm, r,
+ njs_value_arg(&njs_string_lindex),
+ &value);
+ if (njs_slow_path(ret != NJS_OK)) {
+ njs_regex_match_data_free(match_data, vm->regex_generic_ctx);
+ return NJS_ERROR;
+ }
+ }
+
+ if (flags & NJS_REGEXP_FLAG_TEST) {
+ njs_regex_match_data_free(match_data, vm->regex_generic_ctx);
+ njs_set_boolean(retval, 1);
+ return NJS_OK;
+ }
+
result = njs_regexp_exec_result(vm, r, utf8, &string, match_data);

njs_regex_match_data_free(match_data, vm->regex_generic_ctx);
@@ -975,6 +1005,12 @@ not_found:
}


+static njs_exotic_slots_t njs_regexp_prototype_exotic_slots = {
+ .prop_handler = NULL,
+ .keys = NULL,
+};
+
+
static njs_array_t *
njs_regexp_exec_result(njs_vm_t *vm, njs_value_t *r, njs_utf8_t utf8,
njs_string_prop_t *string, njs_regex_match_data_t *match_data)
@@ -986,7 +1022,7 @@ njs_regexp_exec_result(njs_vm_t *vm, njs
njs_int_t ret;
njs_uint_t i, n;
njs_array_t *array;
- njs_value_t name, value;
+ njs_value_t name;
njs_object_t *groups;
njs_regexp_t *regexp;
njs_object_prop_t *prop;
@@ -1005,6 +1041,8 @@ njs_regexp_exec_result(njs_vm_t *vm, njs
goto fail;
}

+ array->object.slots = &njs_regexp_prototype_exotic_slots;
+
for (i = 0; i < pattern->ncaptures; i++) {
n = 2 * i;
c = njs_regex_capture(match_data, n);
@@ -1048,24 +1086,6 @@ njs_regexp_exec_result(njs_vm_t *vm, njs

njs_set_number(&prop->u.value, index);

- if (pattern->global || pattern->sticky) {
- c = njs_regex_capture(match_data, 1);
-
- if (utf8 == NJS_STRING_UTF8) {
- index = njs_string_index(string, c);
-
- } else {
- index = c;
- }
-
- njs_set_number(&value, index);
- ret = njs_value_property_set(vm, r, njs_value_arg(&njs_string_lindex),
- &value);
- if (njs_slow_path(ret != NJS_OK)) {
- goto fail;
- }
- }
-
lhq.key_hash = NJS_INDEX_HASH;
lhq.key = njs_str_value("index");
lhq.replace = 0;
@@ -1184,16 +1204,18 @@ njs_regexp_prototype_exec(njs_vm_t *vm,
return ret;
}

- return njs_regexp_builtin_exec(vm, r, s, retval);
+ return njs_regexp_builtin_exec(vm, r, s,
+ njs_number(njs_arg(args, nargs, 2)), retval);
}


-njs_int_t
-njs_regexp_exec(njs_vm_t *vm, njs_value_t *r, njs_value_t *s,
+static njs_int_t
+njs_regexp_exec(njs_vm_t *vm, njs_value_t *r, njs_value_t *s, unsigned flags,
njs_value_t *retval)
{
njs_int_t ret;
njs_value_t exec;
+ njs_value_t arguments[2];

static const njs_value_t string_exec = njs_string("exec");

@@ -1203,12 +1225,30 @@ njs_regexp_exec(njs_vm_t *vm, njs_value_
}

if (njs_is_function(&exec)) {
- ret = njs_function_call(vm, njs_function(&exec), r, s, 1, retval);
+ njs_value_assign(&arguments[0], s);
+
+ if (flags) {
+ njs_set_number(&arguments[1], flags);
+ }
+
+ ret = njs_function_call(vm, njs_function(&exec), r, arguments,
+ flags ? 2 : 1, retval);
if (njs_slow_path(ret == NJS_ERROR)) {
return NJS_ERROR;
}

- if (njs_slow_path(!njs_is_object(retval) && !njs_is_null(retval))) {
+ if (njs_slow_path(!njs_is_null(retval)
+ && (flags & NJS_REGEXP_FLAG_TEST
+ && !njs_is_boolean(retval))))
+ {
+ njs_type_error(vm, "unexpected \"%s\" retval in njs_regexp_exec()",
+ njs_type_string(retval->type));
+ return NJS_ERROR;
+ }
+
+ if (njs_slow_path(!njs_is_null(retval)
+ && (!flags && !njs_is_object(retval))))
+ {
njs_type_error(vm, "unexpected \"%s\" retval in njs_regexp_exec()",
njs_type_string(retval->type));
return NJS_ERROR;
@@ -1222,7 +1262,7 @@ njs_regexp_exec(njs_vm_t *vm, njs_value_
return NJS_ERROR;
}

- return njs_regexp_builtin_exec(vm, r, s, retval);
+ return njs_regexp_builtin_exec(vm, r, s, flags, retval);
}


@@ -1319,7 +1359,7 @@ njs_regexp_prototype_symbol_replace(njs_
goto exception;
}

- ret = njs_regexp_exec(vm, rx, string, r);
+ ret = njs_regexp_exec(vm, rx, string, 0, r);
if (njs_slow_path(ret != NJS_OK)) {
goto exception;
}
@@ -1630,7 +1670,7 @@ njs_regexp_prototype_symbol_split(njs_vm
length = njs_string_prop(&s, string);

if (njs_slow_path(s.size == 0)) {
- ret = njs_regexp_exec(vm, rx, string, &z);
+ ret = njs_regexp_exec(vm, rx, string, NJS_REGEXP_FLAG_TEST, &z);
if (njs_slow_path(ret != NJS_OK)) {
return NJS_ERROR;
}
@@ -1659,7 +1699,7 @@ njs_regexp_prototype_symbol_split(njs_vm
return NJS_ERROR;
}

- ret = njs_regexp_exec(vm, rx, string, &z);
+ ret = njs_regexp_exec(vm, rx, string, 0, &z);
if (njs_slow_path(ret != NJS_OK)) {
return NJS_ERROR;
}
diff -r f076da3c7a6f -r a6410aad6eab src/njs_regexp.h
--- a/src/njs_regexp.h Fri Sep 29 21:41:34 2023 -0700
+++ b/src/njs_regexp.h Tue Oct 03 00:39:04 2023 -0700
@@ -17,8 +17,6 @@ njs_regexp_pattern_t *njs_regexp_pattern
njs_int_t njs_regexp_match(njs_vm_t *vm, njs_regex_t *regex,
const u_char *subject, size_t off, size_t len, njs_regex_match_data_t *d);
njs_regexp_t *njs_regexp_alloc(njs_vm_t *vm, njs_regexp_pattern_t *pattern);
-njs_int_t njs_regexp_exec(njs_vm_t *vm, njs_value_t *r, njs_value_t *s,
- njs_value_t *retval);
njs_int_t njs_regexp_prototype_exec(njs_vm_t *vm, njs_value_t *args,
njs_uint_t nargs, njs_index_t unused, njs_value_t *retval);
njs_int_t njs_regexp_prototype_symbol_replace(njs_vm_t *vm,
_______________________________________________
nginx-devel mailing list
nginx-devel@nginx.org
https://mailman.nginx.org/mailman/listinfo/nginx-devel
Subject Author Views Posted

[njs] Improved memory footprint of RegExp.prototype.test().

Dmitry Volyntsev 216 October 04, 2023 04:56PM



Sorry, you do not have permission to post/reply in this forum.

Online Users

Guests: 289
Record Number of Users: 8 on April 13, 2023
Record Number of Guests: 421 on December 02, 2018
Powered by nginx      Powered by FreeBSD      PHP Powered      Powered by MariaDB      ipv6 ready