[RFC] Add RFC 3986 and WHATWG compliant URL parsing support #14461

kocsismate · 2024-06-03T20:46:49Z

2nd take after the failed experiment with #11315

RFC: https://wiki.php.net/rfc/url_parsing_api

kocsismate · 2024-06-03T20:48:40Z

ext/standard/basic_functions.stub.php

@@ -3715,7 +3715,6 @@ function uniqid(string $prefix = "", bool $more_entropy = false): string {}

 /**
 * @return int|string|array<string, int|string>|null|false
- * @compile-time-eval


Removed for benchmarking purposes

ext/url/php_url.c

kocsismate · 2024-06-03T20:59:10Z

ext/url/php_url.c

+
+static void cleanup_parser(void)
+{
+	if (++URL_G(urls) % 500 == 0) {


Approach is copy-pasted from lexbor/lexbor#206

ext/url/php_url.c

ext/url/config.m4

In preparation of php#14461 (https://wiki.php.net/rfc/url_parsing_api)

DanielEScherzer · 2025-05-15T20:26:37Z

ext/uri/php_uri_arginfo.h

+
+	zval property_errors_default_value;
+	ZVAL_UNDEF(&property_errors_default_value);
+	zend_string *property_errors_name = zend_string_init("errors", sizeof("errors") - 1, 1);


please add the new strings from zend_string.h to PropertyInfo::PHP_85_KNOWN in gen_stub.php so they can be used here

DanielEScherzer · 2025-05-15T20:28:45Z

ext/uri/php_lexbor.c

+		lexbor_str_init_append(&str, lexbor_parser->mraw, (const lxb_char_t *) Z_STRVAL_P(value), Z_STRLEN_P(value)); \
+	} else if (Z_TYPE_P(value) == IS_LONG && Z_LVAL_P(value) != 0) { \
+        ZVAL_STR(value, zend_long_to_str(Z_LVAL_P(value))); \
+    	lexbor_str_init_append(&str, lexbor_parser->mraw, (const lxb_char_t *) Z_STRVAL_P(value), Z_STRLEN_P(value)); \


a few lines here indented with spaces rather than tabs

DanielEScherzer · 2025-05-15T20:29:06Z

ext/uri/php_lexbor.c

+		case URI_COMPONENT_READ_NORMALIZED_UNICODE: /* Intentional fallthrough */ \
+		case URI_COMPONENT_READ_NORMALIZED_ASCII: { \
+			ZVAL_STRINGL(retval, (const char *) start, len);  \
+        	break; \


Suggested change

break; \

break; \

DanielEScherzer · 2025-05-15T20:32:21Z

ext/uri/php_lexbor.c

+
+	if (lexbor_uri->path.opaque) {
+		LEXBOR_READ_ASCII_URI_COMPONENT(lexbor_uri->path.str.data, lexbor_uri->path.str.length, read_mode, retval);
+	} else if (lexbor_uri->path.str.length) {


these two blocks look the same - can they be combined?

DanielEScherzer · 2025-05-15T20:34:37Z

ext/uri/php_lexbor.c

+	return smart_str_extract(&uri_str);
+}
+
+static void lexbor_free_uri(void *uri)


if this function isn't needed, suggest removing the handler and making it optional in the lexbor_uri_handler specification above

DanielEScherzer · 2025-05-16T04:00:18Z

ext/standard/url.c

+	if (uri_handler_register(&parse_url_uri_handler) == FAILURE) {
+		return FAILURE;
+	}
+


how about just return uri_handler_register(&parse_url_uri_handler); ?

DanielEScherzer · 2025-05-16T04:02:33Z

ext/uri/php_lexbor.c

+			ZVAL_STRINGL(retval, (const char *) start, len);  \
+        	break; \
+		} \
+		EMPTY_SWITCH_DEFAULT_CASE() \


given the EMPTY_SWITCH_DEFAULT_CASE() and the fact that all of the cases fall through, what is the point of even having a switch?

DanielEScherzer · 2025-05-16T04:10:19Z

ext/uri/php_uri.c

+	zend_string *that_str = that_internal_uri->handler->uri_to_string(
+			that_internal_uri->uri, URI_RECOMPOSITION_NORMALIZED_ASCII, exclude_fragment);
+	if (that_str == NULL) {
+		zend_throw_exception_ex(NULL, 0, "Cannot recompose %s to string", ZSTR_VAL(that_object->ce->name));


should this_str be released here?

DanielEScherzer · 2025-05-16T04:15:16Z

ext/uri/php_uri_arginfo.h

+
+ZEND_BEGIN_ARG_WITH_RETURN_TYPE_INFO_EX(arginfo_class_Uri_Rfc3986_Uri_parse, 0, 1, IS_STATIC, 1)
+	ZEND_ARG_TYPE_INFO(0, uri, IS_STRING, 0)
+	ZEND_ARG_OBJ_INFO_WITH_DEFAULT_VALUE(0, baseUrl, Uri\\Rfc3986\\\125ri, 1, "null")


this has the parameter type as Uri\\Rfc3986\\\125ri - any idea why it isn't Uri\\Rfc3986\\Uri? Does it matter? Similar below

DanielEScherzer · 2025-05-16T04:18:47Z

ext/uri/php_uriparser.c

+};
+
+#define URIPARSER_PARSE_STR(uri_str, internal_uri, errors) do { \
+    zend_string *str = smart_str_extract(&uri_str); \


line got indented with spaces rather than a tab

In preparation of php#14461 (https://wiki.php.net/rfc/url_parsing_api)

kocsismate added the Status: Requires RFC label Jun 3, 2024

github-actions bot added Category: Build System Category: Engine Extension: dom Extension: standard labels Jun 3, 2024

kocsismate force-pushed the ext-url2 branch from ee8cc69 to 0e29115 Compare June 3, 2024 20:57

kocsismate commented Jun 3, 2024

View reviewed changes

kocsismate force-pushed the ext-url2 branch 5 times, most recently from 823e4c6 to e65deb7 Compare June 6, 2024 08:23

kocsismate changed the title ~~Add ext/url based on Lexbor~~ [RFC] Add ext/url based on Lexbor Jun 12, 2024

nielsdos reviewed Jun 15, 2024

View reviewed changes

ext/url/php_url.c Outdated Show resolved Hide resolved

bukka reviewed Jul 8, 2024

View reviewed changes

ext/url/config.m4 Outdated Show resolved Hide resolved

kocsismate force-pushed the ext-url2 branch from f897c66 to dcb0299 Compare August 1, 2024 06:30

kocsismate changed the title ~~[RFC] Add ext/url based on Lexbor~~ [RFC] Add RFC 3986 and WHATWG compliant URL parsing support Aug 1, 2024

kocsismate force-pushed the ext-url2 branch 2 times, most recently from 0ccffa0 to 04d4e6e Compare August 13, 2024 12:30

github-actions bot added Extension: openssl Extension: intl Extension: com_dotnet Extension: curl Extension: date Extension: dba Extension: enchant Extension: ffi Extension: fileinfo Extension: filter Extension: ftp labels Aug 13, 2024

kocsismate added 12 commits April 14, 2025 08:56

Add normalization support

6764c12

SOAP test fixes

1b6233c

Fix some memory leaks

ff05ee1

Some cleanups

ea7848d

Changes based on discussion

45542b3

Removal of Uri\Uri

066dd2a

A lot of fixes and API changes

b485171

Updates

b3a48bf

Add new tests, path fixes

ea81d80

Add more tests for verifying the behavior of withers

17c7300

Fix code review comments

aafa027

A few fixes and improvements after feedback

449fda3

kocsismate force-pushed the ext-url2 branch from e25327b to 449fda3 Compare April 14, 2025 07:03

kocsismate added 4 commits April 14, 2025 18:27

Test fixes

96691bf

Remove WHATWG non-raw getters

3ac4884

Rename WHATWG getters again

88c8a46

Add UriComparisonMode

b8a8ba7

kocsismate force-pushed the ext-url2 branch from 0fab842 to b8a8ba7 Compare April 30, 2025 20:23

Expose $softErrors for Uri\WhatWg\Url::resolve()

0a1d1ec

kocsismate force-pushed the ext-url2 branch from f9cc74d to 0a1d1ec Compare May 3, 2025 20:36

Add SensitiveParameter support

35f737d

kocsismate mentioned this pull request May 3, 2025

Clarification needed with regards to percent-encoding of non-UTF 8 characters inside the host uriparser/uriparser#223

Open

kocsismate added a commit to kocsismate/php-src that referenced this pull request May 11, 2025

Create separate lexbor extension

c2b1e8d

In preparation of php#14461 (https://wiki.php.net/rfc/url_parsing_api)

kocsismate mentioned this pull request May 11, 2025

Create separate lexbor extension #18538

Open

kocsismate added a commit to kocsismate/php-src that referenced this pull request May 11, 2025

Create separate lexbor extension

6a309fe

In preparation of php#14461 (https://wiki.php.net/rfc/url_parsing_api)

kocsismate added a commit to kocsismate/php-src that referenced this pull request May 13, 2025

Create separate lexbor extension

9e9cb4c

In preparation of php#14461 (https://wiki.php.net/rfc/url_parsing_api)

kocsismate added a commit to kocsismate/php-src that referenced this pull request May 13, 2025

Create separate lexbor extension

c228b4a

In preparation of php#14461 (https://wiki.php.net/rfc/url_parsing_api)

kocsismate added a commit to kocsismate/php-src that referenced this pull request May 13, 2025

Create separate lexbor extension

accf126

In preparation of php#14461 (https://wiki.php.net/rfc/url_parsing_api)

DanielEScherzer reviewed May 16, 2025

View reviewed changes

nielsdos pushed a commit to kocsismate/php-src that referenced this pull request May 16, 2025

Create separate lexbor extension

e1936fe

In preparation of php#14461 (https://wiki.php.net/rfc/url_parsing_api)

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[RFC] Add RFC 3986 and WHATWG compliant URL parsing support #14461

[RFC] Add RFC 3986 and WHATWG compliant URL parsing support #14461

kocsismate commented Jun 3, 2024 •

edited

Loading

kocsismate Jun 3, 2024

kocsismate Jun 3, 2024

DanielEScherzer May 15, 2025

DanielEScherzer May 15, 2025

DanielEScherzer May 15, 2025

DanielEScherzer May 15, 2025

DanielEScherzer May 15, 2025

DanielEScherzer May 16, 2025

DanielEScherzer May 16, 2025

DanielEScherzer May 16, 2025

DanielEScherzer May 16, 2025

DanielEScherzer May 16, 2025

[RFC] Add RFC 3986 and WHATWG compliant URL parsing support #14461

Are you sure you want to change the base?

[RFC] Add RFC 3986 and WHATWG compliant URL parsing support #14461

Conversation

kocsismate commented Jun 3, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

kocsismate commented Jun 3, 2024 •

edited

Loading