From 66e58d1ea783e7b02ef64fce1acce03297cbdfd0 Mon Sep 17 00:00:00 2001
From: Petr Pchelko <ppchelko@wikimedia.org>
Date: Fri, 8 Oct 2021 11:38:33 -0700
Subject: [PATCH] wfParseUrl: rely on parse_url for proto-relative urls

On July 12 2011 in df2306b97 Roan added this workaround
for parse_url lacking support for protocol-relative URLs.

In 2012 PHP 5.4.7 came out, that added support for
protocol-relative URLs to parse_url, see [1].

It's 2021, and after a decade of honorably serving our
community, this workaround can be now be retired.

It's not a simple deletion since we have custom 'delimiter'
bit that wfParseUrl returns, and we need to keep supporting
that.

[1] https://bugs.php.net/bug.php?id=62844
Change-Id: If9b3bb2a51afa7af1ba8e674eadaa8db4f6f33f9
---
 includes/GlobalFunctions.php                  | 24 +++++++------------
 .../GlobalFunctions/wfParseUrlTest.php        |  8 +++++++
 2 files changed, 16 insertions(+), 16 deletions(-)

diff --git a/includes/GlobalFunctions.php b/includes/GlobalFunctions.php
index 3760c563595..61bf30318f8 100644
--- a/includes/GlobalFunctions.php
+++ b/includes/GlobalFunctions.php
@@ -776,22 +776,13 @@ function wfUrlProtocolsWithoutProtRel() {
 function wfParseUrl( $url ) {
 	global $wgUrlProtocols; // Allow all protocols defined in DefaultSettings/LocalSettings.php
 
-	// Protocol-relative URLs are handled really badly by parse_url(). It's so
-	// bad that the easiest way to handle them is to just prepend 'http:' and
-	// strip the protocol out later.
-	$wasRelative = substr( $url, 0, 2 ) == '//';
-	if ( $wasRelative ) {
-		$url = "http:$url";
-	}
 	$bits = parse_url( $url );
-	// parse_url() returns an array without scheme for some invalid URLs, e.g.
-	// parse_url("%0Ahttp://example.com") == [ 'host' => '%0Ahttp', 'path' => 'example.com' ]
-	if ( !$bits || !isset( $bits['scheme'] ) ) {
+	if ( !$bits ) {
 		return false;
 	}
 
 	// parse_url() incorrectly handles schemes case-sensitively. Convert it to lowercase.
-	$bits['scheme'] = strtolower( $bits['scheme'] );
+	$bits['scheme'] = strtolower( $bits['scheme'] ?? '' );
 
 	// most of the protocols are followed by ://, but mailto: and sometimes news: not, check for it
 	if ( in_array( $bits['scheme'] . '://', $wgUrlProtocols ) ) {
@@ -804,7 +795,13 @@ function wfParseUrl( $url ) {
 			$bits['host'] = $bits['path'];
 			$bits['path'] = '';
 		}
+	} elseif ( !strlen( $bits['scheme'] ) && isset( $bits['host'] ) ) {
+		// This means $url was protocol-relative, e.g. //example.com
+		$bits['delimiter'] = '//';
 	} else {
+		// Option 1: scheme is not in $wgUrlProtocols
+		// Option 2: parse_url() returns an array without scheme or host for some invalid URLs,
+		// e.g. parse_url("%0Ahttp://example.com") == [ 'path' => '%0Ahttp://example.com' ]
 		return false;
 	}
 
@@ -823,11 +820,6 @@ function wfParseUrl( $url ) {
 		}
 	}
 
-	// If the URL was protocol-relative, fix scheme and delimiter
-	if ( $wasRelative ) {
-		$bits['scheme'] = '';
-		$bits['delimiter'] = '//';
-	}
 	return $bits;
 }
 
diff --git a/tests/phpunit/includes/GlobalFunctions/wfParseUrlTest.php b/tests/phpunit/includes/GlobalFunctions/wfParseUrlTest.php
index 8774bb20218..c8f6275f9ea 100644
--- a/tests/phpunit/includes/GlobalFunctions/wfParseUrlTest.php
+++ b/tests/phpunit/includes/GlobalFunctions/wfParseUrlTest.php
@@ -192,6 +192,14 @@ class WfParseUrlTest extends MediaWikiIntegrationTestCase {
 					'fragment' => 'quux',
 				]
 			],
+			[
+				'%0Ahttp://example.com',
+				false,
+			],
+			[
+				'http:///test.com',
+				false,
+			]
 		];
 	}
 }