ParserOutput: Add stateless transforms to getText()
authorBrad Jorsch <bjorsch@wikimedia.org>
Wed, 22 Nov 2017 18:12:23 +0000 (13:12 -0500)
committerBrad Jorsch <bjorsch@wikimedia.org>
Thu, 30 Nov 2017 19:27:46 +0000 (14:27 -0500)
The stateful transforms are deprecated.

Inspired by Krinkle's If2fb32fc.

Bug: T171797
Change-Id: Ied5fe1a6159c2d4fa48170042b44d735ce7b6f9b

RELEASE-NOTES-1.31
docs/hooks.txt
includes/parser/ParserOptions.php
includes/parser/ParserOutput.php
tests/phpunit/includes/parser/ParserOutputTest.php

index b32e3e7..139773b 100644 (file)
@@ -31,6 +31,8 @@ production.
   [[iw:User:Example|iw>Example]].
 * (T111605) The 'ImportHandleUnknownUser' hook allows extensions to auto-create
   users during an import.
   [[iw:User:Example|iw>Example]].
 * (T111605) The 'ImportHandleUnknownUser' hook allows extensions to auto-create
   users during an import.
+* Added a hook, ParserOutputPostCacheTransform, to allow extensions to affect
+  the ParserOutput::getText() post-cache transformations.
 
 === External library changes in 1.31 ===
 
 
 === External library changes in 1.31 ===
 
@@ -119,6 +121,18 @@ changes to languages because of Phabricator reports.
 * The Block class will no longer accept usable-but-missing usernames for
   'byText' or ->setBlocker(). Callers should either ensure the blocker exists
   locally or use a new interwiki-format username like "iw>Example".
 * The Block class will no longer accept usable-but-missing usernames for
   'byText' or ->setBlocker(). Callers should either ensure the blocker exists
   locally or use a new interwiki-format username like "iw>Example".
+* The following methods that get and set ParserOutput state are deprecated.
+  Callers should use the new stateless $options parameter to
+  ParserOutput::getText() instead.
+  * ParserOptions::getEditSection()
+  * ParserOptions::setEditSection()
+  * ParserOutput::getEditSectionTokens()
+  * ParserOutput::setEditSectionTokens()
+  * ParserOutput::getTOCEnabled()
+  * ParserOutput::setTOCEnabled()
+  * OutputPage::enableSectionEditLinks()
+  * OutputPage::sectionEditLinksEnabled()
+  * The public ParserOutput state fields $mTOCEnabled and $mEditSectionTokens are also deprecated.
 
 == Compatibility ==
 MediaWiki 1.31 requires PHP 5.5.9 or later. There is experimental support for
 
 == Compatibility ==
 MediaWiki 1.31 requires PHP 5.5.9 or later. There is experimental support for
index 685a182..29883b2 100644 (file)
@@ -2594,6 +2594,12 @@ RejectParserCacheValue hook) because MediaWiki won't do it for you.
   callable here. The callable is passed the ParserOptions object and the option
   name.
 
   callable here. The callable is passed the ParserOptions object and the option
   name.
 
+'ParserOutputPostCacheTransform': Called from ParserOutput::getText() to do
+post-cache transforms.
+$parserOutput: The ParserOutput object.
+&$text: The text being transformed, before core transformations are done.
+&$options: The options array being used for the transformation.
+
 'ParserSectionCreate': Called each time the parser creates a document section
 from wikitext. Use this to apply per-section modifications to HTML (like
 wrapping the section in a DIV).  Caveat: DIVs are valid wikitext, and a DIV
 'ParserSectionCreate': Called each time the parser creates a document section
 from wikitext. Use this to apply per-section modifications to HTML (like
 wrapping the section in a DIV).  Caveat: DIVs are valid wikitext, and a DIV
index 5e2845f..f99089b 100644 (file)
@@ -869,6 +869,7 @@ class ParserOptions {
 
        /**
         * Create "edit section" links?
 
        /**
         * Create "edit section" links?
+        * @deprecated since 1.31, use ParserOutput::getText() options instead.
         * @return bool
         */
        public function getEditSection() {
         * @return bool
         */
        public function getEditSection() {
@@ -877,6 +878,7 @@ class ParserOptions {
 
        /**
         * Create "edit section" links?
 
        /**
         * Create "edit section" links?
+        * @deprecated since 1.31, use ParserOutput::getText() options instead.
         * @param bool|null $x New value (null is no change)
         * @return bool Old value
         */
         * @param bool|null $x New value (null is no change)
         * @return bool Old value
         */
index 3480a51..59c27e5 100644 (file)
@@ -144,6 +144,7 @@ class ParserOutput extends CacheTime {
        public $mSections = [];
 
        /**
        public $mSections = [];
 
        /**
+        * @deprecated since 1.31 Use getText() options.
         * @var bool $mEditSectionTokens prefix/suffix markers if edit sections were output as tokens.
         */
        public $mEditSectionTokens = false;
         * @var bool $mEditSectionTokens prefix/suffix markers if edit sections were output as tokens.
         */
        public $mEditSectionTokens = false;
@@ -164,6 +165,7 @@ class ParserOutput extends CacheTime {
        public $mTimestamp;
 
        /**
        public $mTimestamp;
 
        /**
+        * @deprecated since 1.31 Use getText() options.
         * @var bool $mTOCEnabled Whether TOC should be shown, can't override __NOTOC__.
         */
        public $mTOCEnabled = true;
         * @var bool $mTOCEnabled Whether TOC should be shown, can't override __NOTOC__.
         */
        public $mTOCEnabled = true;
@@ -250,9 +252,38 @@ class ParserOutput extends CacheTime {
                return $this->mText;
        }
 
                return $this->mText;
        }
 
-       public function getText() {
+       /**
+        * Get the output HTML
+        *
+        * @param array $options (since 1.31) Transformations to apply to the HTML
+        *  - allowTOC: (bool) Show the TOC, assuming there were enough headings
+        *     to generate one and `__NOTOC__` wasn't used. Default is true,
+        *     but might be statefully overridden.
+        *  - enableSectionEditLinks: (bool) Include section edit links, assuming
+        *    section edit link tokens are present in the HTML. Default is true,
+        *     but might be statefully overridden.
+        * @return string HTML
+        */
+       public function getText( $options = [] ) {
+               // @todo Warn if !array_key_exists( 'allowTOC', $options ) && empty( $this->mTOCEnabled )
+
+               // @todo Warn if !array_key_exists( 'enableSectionEditLinks', $options )
+               //     && !$this->mEditSectionTokens
+               //  Note that while $this->mEditSectionTokens defaults to false,
+               //  ParserOptions->getEditSection() defaults to true and Parser copies
+               //  that to us so true makes more sense as the stateless default.
+
+               $options += [
+                       // empty() here because old cached versions might lack the field somehow.
+                       // In that situation, the historical behavior (possibly buggy) is to remove the TOC.
+                       'allowTOC' => !empty( $this->mTOCEnabled ),
+                       'enableSectionEditLinks' => $this->mEditSectionTokens,
+               ];
                $text = $this->mText;
                $text = $this->mText;
-               if ( $this->mEditSectionTokens ) {
+
+               Hooks::runWithoutAbort( 'ParserOutputPostCacheTransform', [ $this, &$text, &$options ] );
+
+               if ( $options['enableSectionEditLinks'] ) {
                        $text = preg_replace_callback(
                                self::EDITSECTION_REGEX,
                                function ( $m ) {
                        $text = preg_replace_callback(
                                self::EDITSECTION_REGEX,
                                function ( $m ) {
@@ -278,8 +309,7 @@ class ParserOutput extends CacheTime {
                        $text = preg_replace( self::EDITSECTION_REGEX, '', $text );
                }
 
                        $text = preg_replace( self::EDITSECTION_REGEX, '', $text );
                }
 
-               // If you have an old cached version of this class - sorry, you can't disable the TOC
-               if ( isset( $this->mTOCEnabled ) && $this->mTOCEnabled ) {
+               if ( $options['allowTOC'] ) {
                        $text = str_replace( [ Parser::TOC_START, Parser::TOC_END ], '', $text );
                } else {
                        $text = preg_replace(
                        $text = str_replace( [ Parser::TOC_START, Parser::TOC_END ], '', $text );
                } else {
                        $text = preg_replace(
@@ -288,6 +318,7 @@ class ParserOutput extends CacheTime {
                                $text
                        );
                }
                                $text
                        );
                }
+
                return $text;
        }
 
                return $text;
        }
 
@@ -339,6 +370,9 @@ class ParserOutput extends CacheTime {
                return $this->mSections;
        }
 
                return $this->mSections;
        }
 
+       /**
+        * @deprecated since 1.31 Use getText() options.
+        */
        public function getEditSectionTokens() {
                return $this->mEditSectionTokens;
        }
        public function getEditSectionTokens() {
                return $this->mEditSectionTokens;
        }
@@ -426,6 +460,9 @@ class ParserOutput extends CacheTime {
                return $this->mLimitReportJSData;
        }
 
                return $this->mLimitReportJSData;
        }
 
+       /**
+        * @deprecated since 1.31 Use getText() options.
+        */
        public function getTOCEnabled() {
                return $this->mTOCEnabled;
        }
        public function getTOCEnabled() {
                return $this->mTOCEnabled;
        }
@@ -454,6 +491,9 @@ class ParserOutput extends CacheTime {
                return wfSetVar( $this->mSections, $toc );
        }
 
                return wfSetVar( $this->mSections, $toc );
        }
 
+       /**
+        * @deprecated since 1.31 Use getText() options.
+        */
        public function setEditSectionTokens( $t ) {
                return wfSetVar( $this->mEditSectionTokens, $t );
        }
        public function setEditSectionTokens( $t ) {
                return wfSetVar( $this->mEditSectionTokens, $t );
        }
@@ -470,6 +510,9 @@ class ParserOutput extends CacheTime {
                return wfSetVar( $this->mTimestamp, $timestamp );
        }
 
                return wfSetVar( $this->mTimestamp, $timestamp );
        }
 
+       /**
+        * @deprecated since 1.31 Use getText() options.
+        */
        public function setTOCEnabled( $flag ) {
                return wfSetVar( $this->mTOCEnabled, $flag );
        }
        public function setTOCEnabled( $flag ) {
                return wfSetVar( $this->mTOCEnabled, $flag );
        }
index ec8f0d0..441d60d 100644 (file)
@@ -1,5 +1,7 @@
 <?php
 
 <?php
 
+use Wikimedia\TestingAccessWrapper;
+
 /**
  * @group Database
  *        ^--- trigger DB shadowing because we are using Title magic
 /**
  * @group Database
  *        ^--- trigger DB shadowing because we are using Title magic
@@ -89,4 +91,231 @@ class ParserOutputTest extends MediaWikiTestCase {
                $this->assertArrayNotHasKey( 'foo', $properties );
        }
 
                $this->assertArrayNotHasKey( 'foo', $properties );
        }
 
+       /**
+        * @covers ParserOutput::getText
+        * @dataProvider provideGetText
+        * @param array $options Options to getText()
+        * @param array $poState ParserOptions state fields to set
+        * @param string $text Parser text
+        * @param string $expect Expected output
+        */
+       public function testGetText( $options, $poState, $text, $expect ) {
+               $this->setMwGlobals( [
+                       'wgArticlePath' => '/wiki/$1',
+                       'wgScriptPath' => '/w',
+                       'wgScript' => '/w/index.php',
+               ] );
+
+               $po = new ParserOutput( $text );
+
+               // Emulate Parser
+               $po->setEditSectionTokens( true );
+
+               if ( $poState ) {
+                       $wrap = TestingAccessWrapper::newFromObject( $po );
+                       foreach ( $poState as $key => $value ) {
+                               $wrap->$key = $value;
+                       }
+               }
+
+               $actual = $po->getText( $options );
+               $this->assertSame( $expect, $actual );
+       }
+
+       public static function provideGetText() {
+               // @codingStandardsIgnoreStart Generic.Files.LineLength
+               $text = <<<EOF
+<p>Test document.
+</p>
+<mw:toc><div id="toc" class="toc"><div class="toctitle"><h2>Contents</h2></div>
+<ul>
+<li class="toclevel-1 tocsection-1"><a href="#Section_1"><span class="tocnumber">1</span> <span class="toctext">Section 1</span></a></li>
+<li class="toclevel-1 tocsection-2"><a href="#Section_2"><span class="tocnumber">2</span> <span class="toctext">Section 2</span></a>
+<ul>
+<li class="toclevel-2 tocsection-3"><a href="#Section_2.1"><span class="tocnumber">2.1</span> <span class="toctext">Section 2.1</span></a></li>
+</ul>
+</li>
+<li class="toclevel-1 tocsection-4"><a href="#Section_3"><span class="tocnumber">3</span> <span class="toctext">Section 3</span></a></li>
+</ul>
+</div>
+</mw:toc>
+<h2><span class="mw-headline" id="Section_1">Section 1</span><mw:editsection page="Test Page" section="1">Section 1</mw:editsection></h2>
+<p>One
+</p>
+<h2><span class="mw-headline" id="Section_2">Section 2</span><mw:editsection page="Test Page" section="2">Section 2</mw:editsection></h2>
+<p>Two
+</p>
+<h3><span class="mw-headline" id="Section_2.1">Section 2.1</span><mw:editsection page="Test Page" section="3">Section 2.1</mw:editsection></h3>
+<p>Two point one
+</p>
+<h2><span class="mw-headline" id="Section_3">Section 3</span><mw:editsection page="Test Page" section="4">Section 3</mw:editsection></h2>
+<p>Three
+</p>
+EOF;
+
+               return [
+                       'No stateless options, default state' => [
+                               [], [], $text, <<<EOF
+<p>Test document.
+</p>
+<div id="toc" class="toc"><div class="toctitle"><h2>Contents</h2></div>
+<ul>
+<li class="toclevel-1 tocsection-1"><a href="#Section_1"><span class="tocnumber">1</span> <span class="toctext">Section 1</span></a></li>
+<li class="toclevel-1 tocsection-2"><a href="#Section_2"><span class="tocnumber">2</span> <span class="toctext">Section 2</span></a>
+<ul>
+<li class="toclevel-2 tocsection-3"><a href="#Section_2.1"><span class="tocnumber">2.1</span> <span class="toctext">Section 2.1</span></a></li>
+</ul>
+</li>
+<li class="toclevel-1 tocsection-4"><a href="#Section_3"><span class="tocnumber">3</span> <span class="toctext">Section 3</span></a></li>
+</ul>
+</div>
+
+<h2><span class="mw-headline" id="Section_1">Section 1</span><span class="mw-editsection"><span class="mw-editsection-bracket">[</span><a href="/w/index.php?title=Test_Page&amp;action=edit&amp;section=1" title="Edit section: Section 1">edit</a><span class="mw-editsection-bracket">]</span></span></h2>
+<p>One
+</p>
+<h2><span class="mw-headline" id="Section_2">Section 2</span><span class="mw-editsection"><span class="mw-editsection-bracket">[</span><a href="/w/index.php?title=Test_Page&amp;action=edit&amp;section=2" title="Edit section: Section 2">edit</a><span class="mw-editsection-bracket">]</span></span></h2>
+<p>Two
+</p>
+<h3><span class="mw-headline" id="Section_2.1">Section 2.1</span><span class="mw-editsection"><span class="mw-editsection-bracket">[</span><a href="/w/index.php?title=Test_Page&amp;action=edit&amp;section=3" title="Edit section: Section 2.1">edit</a><span class="mw-editsection-bracket">]</span></span></h3>
+<p>Two point one
+</p>
+<h2><span class="mw-headline" id="Section_3">Section 3</span><span class="mw-editsection"><span class="mw-editsection-bracket">[</span><a href="/w/index.php?title=Test_Page&amp;action=edit&amp;section=4" title="Edit section: Section 3">edit</a><span class="mw-editsection-bracket">]</span></span></h2>
+<p>Three
+</p>
+EOF
+                       ],
+                       'No stateless options, TOC statefully disabled' => [
+                               [], [ 'mTOCEnabled' => false ], $text, <<<EOF
+<p>Test document.
+</p>
+
+<h2><span class="mw-headline" id="Section_1">Section 1</span><span class="mw-editsection"><span class="mw-editsection-bracket">[</span><a href="/w/index.php?title=Test_Page&amp;action=edit&amp;section=1" title="Edit section: Section 1">edit</a><span class="mw-editsection-bracket">]</span></span></h2>
+<p>One
+</p>
+<h2><span class="mw-headline" id="Section_2">Section 2</span><span class="mw-editsection"><span class="mw-editsection-bracket">[</span><a href="/w/index.php?title=Test_Page&amp;action=edit&amp;section=2" title="Edit section: Section 2">edit</a><span class="mw-editsection-bracket">]</span></span></h2>
+<p>Two
+</p>
+<h3><span class="mw-headline" id="Section_2.1">Section 2.1</span><span class="mw-editsection"><span class="mw-editsection-bracket">[</span><a href="/w/index.php?title=Test_Page&amp;action=edit&amp;section=3" title="Edit section: Section 2.1">edit</a><span class="mw-editsection-bracket">]</span></span></h3>
+<p>Two point one
+</p>
+<h2><span class="mw-headline" id="Section_3">Section 3</span><span class="mw-editsection"><span class="mw-editsection-bracket">[</span><a href="/w/index.php?title=Test_Page&amp;action=edit&amp;section=4" title="Edit section: Section 3">edit</a><span class="mw-editsection-bracket">]</span></span></h2>
+<p>Three
+</p>
+EOF
+                       ],
+                       'No stateless options, section edits statefully disabled' => [
+                               [], [ 'mEditSectionTokens' => false ], $text, <<<EOF
+<p>Test document.
+</p>
+<div id="toc" class="toc"><div class="toctitle"><h2>Contents</h2></div>
+<ul>
+<li class="toclevel-1 tocsection-1"><a href="#Section_1"><span class="tocnumber">1</span> <span class="toctext">Section 1</span></a></li>
+<li class="toclevel-1 tocsection-2"><a href="#Section_2"><span class="tocnumber">2</span> <span class="toctext">Section 2</span></a>
+<ul>
+<li class="toclevel-2 tocsection-3"><a href="#Section_2.1"><span class="tocnumber">2.1</span> <span class="toctext">Section 2.1</span></a></li>
+</ul>
+</li>
+<li class="toclevel-1 tocsection-4"><a href="#Section_3"><span class="tocnumber">3</span> <span class="toctext">Section 3</span></a></li>
+</ul>
+</div>
+
+<h2><span class="mw-headline" id="Section_1">Section 1</span></h2>
+<p>One
+</p>
+<h2><span class="mw-headline" id="Section_2">Section 2</span></h2>
+<p>Two
+</p>
+<h3><span class="mw-headline" id="Section_2.1">Section 2.1</span></h3>
+<p>Two point one
+</p>
+<h2><span class="mw-headline" id="Section_3">Section 3</span></h2>
+<p>Three
+</p>
+EOF
+                       ],
+                       'Stateless options override stateful settings' => [
+                               [ 'allowTOC' => true, 'enableSectionEditLinks' => true ],
+                               [ 'mTOCEnabled' => false, 'mEditSectionTokens' => false ],
+                               $text, <<<EOF
+<p>Test document.
+</p>
+<div id="toc" class="toc"><div class="toctitle"><h2>Contents</h2></div>
+<ul>
+<li class="toclevel-1 tocsection-1"><a href="#Section_1"><span class="tocnumber">1</span> <span class="toctext">Section 1</span></a></li>
+<li class="toclevel-1 tocsection-2"><a href="#Section_2"><span class="tocnumber">2</span> <span class="toctext">Section 2</span></a>
+<ul>
+<li class="toclevel-2 tocsection-3"><a href="#Section_2.1"><span class="tocnumber">2.1</span> <span class="toctext">Section 2.1</span></a></li>
+</ul>
+</li>
+<li class="toclevel-1 tocsection-4"><a href="#Section_3"><span class="tocnumber">3</span> <span class="toctext">Section 3</span></a></li>
+</ul>
+</div>
+
+<h2><span class="mw-headline" id="Section_1">Section 1</span><span class="mw-editsection"><span class="mw-editsection-bracket">[</span><a href="/w/index.php?title=Test_Page&amp;action=edit&amp;section=1" title="Edit section: Section 1">edit</a><span class="mw-editsection-bracket">]</span></span></h2>
+<p>One
+</p>
+<h2><span class="mw-headline" id="Section_2">Section 2</span><span class="mw-editsection"><span class="mw-editsection-bracket">[</span><a href="/w/index.php?title=Test_Page&amp;action=edit&amp;section=2" title="Edit section: Section 2">edit</a><span class="mw-editsection-bracket">]</span></span></h2>
+<p>Two
+</p>
+<h3><span class="mw-headline" id="Section_2.1">Section 2.1</span><span class="mw-editsection"><span class="mw-editsection-bracket">[</span><a href="/w/index.php?title=Test_Page&amp;action=edit&amp;section=3" title="Edit section: Section 2.1">edit</a><span class="mw-editsection-bracket">]</span></span></h3>
+<p>Two point one
+</p>
+<h2><span class="mw-headline" id="Section_3">Section 3</span><span class="mw-editsection"><span class="mw-editsection-bracket">[</span><a href="/w/index.php?title=Test_Page&amp;action=edit&amp;section=4" title="Edit section: Section 3">edit</a><span class="mw-editsection-bracket">]</span></span></h2>
+<p>Three
+</p>
+EOF
+                       ],
+                       'Statelessly disable section edit links' => [
+                               [ 'enableSectionEditLinks' => false ], [], $text, <<<EOF
+<p>Test document.
+</p>
+<div id="toc" class="toc"><div class="toctitle"><h2>Contents</h2></div>
+<ul>
+<li class="toclevel-1 tocsection-1"><a href="#Section_1"><span class="tocnumber">1</span> <span class="toctext">Section 1</span></a></li>
+<li class="toclevel-1 tocsection-2"><a href="#Section_2"><span class="tocnumber">2</span> <span class="toctext">Section 2</span></a>
+<ul>
+<li class="toclevel-2 tocsection-3"><a href="#Section_2.1"><span class="tocnumber">2.1</span> <span class="toctext">Section 2.1</span></a></li>
+</ul>
+</li>
+<li class="toclevel-1 tocsection-4"><a href="#Section_3"><span class="tocnumber">3</span> <span class="toctext">Section 3</span></a></li>
+</ul>
+</div>
+
+<h2><span class="mw-headline" id="Section_1">Section 1</span></h2>
+<p>One
+</p>
+<h2><span class="mw-headline" id="Section_2">Section 2</span></h2>
+<p>Two
+</p>
+<h3><span class="mw-headline" id="Section_2.1">Section 2.1</span></h3>
+<p>Two point one
+</p>
+<h2><span class="mw-headline" id="Section_3">Section 3</span></h2>
+<p>Three
+</p>
+EOF
+                       ],
+                       'Statelessly disable TOC' => [
+                               [ 'allowTOC' => false ], [], $text, <<<EOF
+<p>Test document.
+</p>
+
+<h2><span class="mw-headline" id="Section_1">Section 1</span><span class="mw-editsection"><span class="mw-editsection-bracket">[</span><a href="/w/index.php?title=Test_Page&amp;action=edit&amp;section=1" title="Edit section: Section 1">edit</a><span class="mw-editsection-bracket">]</span></span></h2>
+<p>One
+</p>
+<h2><span class="mw-headline" id="Section_2">Section 2</span><span class="mw-editsection"><span class="mw-editsection-bracket">[</span><a href="/w/index.php?title=Test_Page&amp;action=edit&amp;section=2" title="Edit section: Section 2">edit</a><span class="mw-editsection-bracket">]</span></span></h2>
+<p>Two
+</p>
+<h3><span class="mw-headline" id="Section_2.1">Section 2.1</span><span class="mw-editsection"><span class="mw-editsection-bracket">[</span><a href="/w/index.php?title=Test_Page&amp;action=edit&amp;section=3" title="Edit section: Section 2.1">edit</a><span class="mw-editsection-bracket">]</span></span></h3>
+<p>Two point one
+</p>
+<h2><span class="mw-headline" id="Section_3">Section 3</span><span class="mw-editsection"><span class="mw-editsection-bracket">[</span><a href="/w/index.php?title=Test_Page&amp;action=edit&amp;section=4" title="Edit section: Section 3">edit</a><span class="mw-editsection-bracket">]</span></span></h2>
+<p>Three
+</p>
+EOF
+                       ],
+               ];
+               // @codingStandardsIgnoreEnd
+       }
+
 }
 }