mirror of
https://gerrit.wikimedia.org/r/mediawiki/extensions/TemplateStyles
synced 2024-11-28 02:00:20 +00:00
Linting and slight tweak to parser
- add a tweak to the parser (keep the value tokens in a declaration separate in the parsed tree to ease matching at render time - add error checking to encoding/decoding the templatestyles property - pick some lint suggested by Brion and Bryan Change-Id: If60b91e119102c0f0f559fe7e5a4c421c94b7ff4
This commit is contained in:
Coren
parent
b35763edbf
commit
dfa7e27b5a
|
|
@ -27,47 +27,71 @@ class CSSParser {
|
|||
function __construct( $css ) {
|
||||
preg_match_all( '/(
|
||||
[ \n\t]+
|
||||
(?# Sequences of whitespace )
|
||||
| \/\* (?: [^*]+ | \*[^\/] )* \*\/ [ \n\t]*
|
||||
(?# Comments and any trailing whitespace )
|
||||
| " (?: [^"\\\\\n]+ | \\\\\. )* ["\n]
|
||||
(?# Double-quoted string literals (to newline when unclosed )
|
||||
| \' (?: [^\'\\\\\n]+ | \\\\\. )* [\'\n]
|
||||
(#? Single-quoted string literals (to newline when unclosed )
|
||||
| [+-]? (?: [0-9]* \. )? [0-9]+ (?: [_a-z][_a-z0-9-]* | % )?
|
||||
| url [ \n\t]* \(
|
||||
(#? Numerical literals - including optional trailing units or percent sign )
|
||||
| @? -? (?: [_a-z] | \\\\[0-9a-f]{1,6} [ \n\t]? )
|
||||
(?: [_a-z0-9-]+ | \\\\[0-9a-f]{1,6} [ \n\t]? | [^\0-\177] )*
|
||||
(?: [_a-z0-9-]+ | \\\\[0-9a-f]{1,6} [ \n\t]? | [^\0-\177] )* (?: [ \n\t]* \( )?
|
||||
(#? Identifiers - including leading `@` for at-rule blocks )
|
||||
(#? Trailing open captures are captured to match functional values )
|
||||
| \# (?: [_a-z0-9-]+ | \\\\[0-9a-f]{1,6} [ \n\t]? | [^\0-\177] )*
|
||||
(#? So-called hatch literals )
|
||||
| u\+ [0-9a-f]{1,6} (?: - [0-9a-f]{1,6} )?
|
||||
(#? Unicode range literals )
|
||||
| u\+ [0-9a-f?]{1,6}
|
||||
| <!--
|
||||
| -->
|
||||
| .)/xis', $css, $match );
|
||||
(#? Unicode mask literals )
|
||||
| .)
|
||||
(#? Any unmatched token is reduced to single characters )
|
||||
/xis', $css, $match );
|
||||
|
||||
$space = false;
|
||||
foreach ( $match[0] as $t ) {
|
||||
if ( preg_match( '/^(?:[ \n\t]|\/\*|<!--|-->)/', $t ) ) {
|
||||
if ( preg_match( '/^(?:[ \n\t]|\/\*)/', $t ) ) {
|
||||
|
||||
// Fold any sequence of whitespace to a single space token
|
||||
|
||||
if ( !$space ) {
|
||||
$space = true;
|
||||
$this->tokens[] = ' ';
|
||||
continue;
|
||||
}
|
||||
|
||||
} else {
|
||||
|
||||
// decode any hexadecimal escape character into their corresponding UTF-8
|
||||
// sequence - our output is UTF-8 so the escaping is unnecessary and this
|
||||
// prevents trying to obfuscate ASCII in identifiers to prevent matches.
|
||||
|
||||
$t = preg_replace_callback( '/\\\\([0-9a-f]{1,6})[ \n\t]?/', function( $match ) {
|
||||
return html_entity_decode( '&#'.$match[1].';', ENT_NOQUOTES, 'UTF-8' );
|
||||
}, $t );
|
||||
$space = false;
|
||||
$this->tokens[] = $t;
|
||||
|
||||
}
|
||||
}
|
||||
$this->index = 0;
|
||||
}
|
||||
|
||||
private function peek( $i ) {
|
||||
if ( $this->index+$i >= count( $this->tokens ) )
|
||||
if ( $this->index+$i >= count( $this->tokens ) ) {
|
||||
return null;
|
||||
}
|
||||
return $this->tokens[$this->index+$i];
|
||||
}
|
||||
|
||||
private function consume( $num = 1 ) {
|
||||
if ( $num > 0 ) {
|
||||
if ( $this->index+$num >= count( $this->tokens ) )
|
||||
if ( $this->index+$num >= count( $this->tokens ) ) {
|
||||
$num = count( $this->tokens ) - $this->index;
|
||||
$text = implode( array_slice( $this->tokens, $this->index, $num ) );
|
||||
}
|
||||
$text = array_slice( $this->tokens, $this->index, $num );
|
||||
$this->index += $num;
|
||||
return $text;
|
||||
}
|
||||
|
|
@ -76,15 +100,17 @@ class CSSParser {
|
|||
|
||||
private function consumeTo( $delim ) {
|
||||
$consume = 0;
|
||||
while ( !in_array( $this->peek( $consume ), $delim ) )
|
||||
while ( !in_array( $this->peek( $consume ), $delim ) ) {
|
||||
$consume++;
|
||||
}
|
||||
return $this->consume( $consume );
|
||||
}
|
||||
|
||||
private function consumeWS() {
|
||||
$consume = 0;
|
||||
while ( $this->peek( $consume ) === ' ' )
|
||||
while ( $this->peek( $consume ) === ' ' ) {
|
||||
$consume++;
|
||||
}
|
||||
return $this->consume( $consume );
|
||||
}
|
||||
|
||||
|
|
@ -99,7 +125,7 @@ class CSSParser {
|
|||
*/
|
||||
private function parseDecl() {
|
||||
$this->consumeWS();
|
||||
$name = $this->consume();
|
||||
$name = $this->consume()[0];
|
||||
$this->consumeWS();
|
||||
if ( $this->peek( 0 )!=':' ) {
|
||||
$this->consumeTo( [';', '}', null] );
|
||||
|
|
@ -113,7 +139,7 @@ class CSSParser {
|
|||
$this->consumeWS();
|
||||
$value = $this->consumeTo( [';', '}', null] );
|
||||
if ( $this->peek( 0 ) == ';' ) {
|
||||
$value .= $this->consume();
|
||||
$this->consume();
|
||||
$this->consumeWS();
|
||||
}
|
||||
return [ $name => $value ];
|
||||
|
|
@ -132,12 +158,15 @@ class CSSParser {
|
|||
$decls = [];
|
||||
while ( $this->peek( 0 ) !== null and $this->peek( 0 ) != '}' ) {
|
||||
$decl = $this->parseDecl();
|
||||
if ( $decl )
|
||||
foreach ( $decl as $k => $d )
|
||||
if ( $decl ) {
|
||||
foreach ( $decl as $k => $d ) {
|
||||
$decls[$k] = $d;
|
||||
}
|
||||
if ( $this->peek( 0 ) == '}' )
|
||||
}
|
||||
}
|
||||
if ( $this->peek( 0 ) == '}' ) {
|
||||
$this->consume();
|
||||
}
|
||||
return $decls;
|
||||
}
|
||||
|
||||
|
|
@ -163,8 +192,9 @@ class CSSParser {
|
|||
$this->consume();
|
||||
$this->consumeWS();
|
||||
$text = '';
|
||||
} else
|
||||
$text .= $this->consume();
|
||||
} else {
|
||||
$text .= $this->consume()[0];
|
||||
}
|
||||
}
|
||||
$selectors[] = $text;
|
||||
if ( $this->peek( 0 ) == '{' ) {
|
||||
|
|
@ -204,37 +234,44 @@ class CSSParser {
|
|||
$at = $this->consume();
|
||||
$this->consumeWS();
|
||||
$text = '';
|
||||
while ( !in_array( $this->peek( 0 ), ['{', ';', null] ) )
|
||||
$text .= $this->consume();
|
||||
while ( !in_array( $this->peek( 0 ), ['{', ';', null] ) ) {
|
||||
$text .= $this->consume()[0];
|
||||
}
|
||||
if ( $this->peek( 0 ) == '{' ) {
|
||||
$this->consume();
|
||||
$r = $this->rules( [ '}', null ] );
|
||||
if ( $r )
|
||||
if ( $r ) {
|
||||
$atrules[] = [ "name"=>$at, "text"=>$text, "rules"=>$r ];
|
||||
}
|
||||
} else {
|
||||
$atrules[] = [ "name"=>$at, "text"=>$text ];
|
||||
}
|
||||
} elseif ( $this->peek( 0 )[0] == '@' ) {
|
||||
$at = $this->consume();
|
||||
$text = '';
|
||||
while ( !in_array( $this->peek( 0 ), ['{', ';', null] ) )
|
||||
$text .= $this->consume();
|
||||
while ( !in_array( $this->peek( 0 ), ['{', ';', null] ) ) {
|
||||
$text .= $this->consume()[0];
|
||||
}
|
||||
if ( $this->peek( 0 ) == '{' ) {
|
||||
$this->consume();
|
||||
$decl = $this->parseDecls();
|
||||
if ( $decl )
|
||||
if ( $decl ) {
|
||||
$atrules[] = [ "name"=>$at, "text"=>$text, "rules"=>[ "selectors"=>'', "decls"=>$decl ] ];
|
||||
}
|
||||
} else {
|
||||
$atrules[] = [ "name"=>$at, "text"=>$text ];
|
||||
}
|
||||
} else
|
||||
} else {
|
||||
$rules[] = $this->parseRule();
|
||||
}
|
||||
$this->consumeWS();
|
||||
}
|
||||
if ( $rules )
|
||||
if ( $rules ) {
|
||||
$atrules[] = [ "name"=>'', "rules"=>$rules ];
|
||||
if ( $this->peek( 0 ) !== null )
|
||||
}
|
||||
if ( $this->peek( 0 ) !== null ) {
|
||||
$this->consume();
|
||||
}
|
||||
return $atrules;
|
||||
}
|
||||
|
||||
|
|
|
|||
|
|
@ -27,14 +27,16 @@ class CSSRenderer {
|
|||
* and defaults to the empty string.
|
||||
*/
|
||||
function add( $rules, $media = '' ) {
|
||||
if ( !array_key_exists( $media, $this->bymedia ) )
|
||||
if ( !array_key_exists( $media, $this->bymedia ) ) {
|
||||
$this->bymedia[$media] = [];
|
||||
}
|
||||
|
||||
foreach ( $rules as $at ) {
|
||||
switch ( $at['name'] ) {
|
||||
case '@media':
|
||||
if ( $media == '' )
|
||||
if ( $media == '' ) {
|
||||
$this->add( $at['rules'], "@media ".$at['text'] );
|
||||
}
|
||||
break;
|
||||
case '':
|
||||
$this->bymedia[$media] = array_merge( $this->bymedia[$media], $at['rules'] );
|
||||
|
|
@ -54,18 +56,20 @@ class CSSRenderer {
|
|||
$css = '';
|
||||
|
||||
foreach ( $this->bymedia as $at => $rules ) {
|
||||
if ( $at != '' )
|
||||
if ( $at != '' ) {
|
||||
$css .= "$at {\n";
|
||||
}
|
||||
foreach ( $rules as $rule ) {
|
||||
$css .= implode( ',', $rule['selectors'] ) . "{";
|
||||
foreach ( $rule['decls'] as $key => $value ) {
|
||||
$css .= "$key:$value";
|
||||
$css .= "$key:" . implode( '', $value ) . ';';
|
||||
}
|
||||
$css .= "} ";
|
||||
}
|
||||
if ( $at != '' )
|
||||
if ( $at != '' ) {
|
||||
$css .= "} ";
|
||||
}
|
||||
}
|
||||
|
||||
return $css;
|
||||
}
|
||||
|
|
|
|||
|
|
@ -11,27 +11,40 @@ class TemplateStylesHooks {
|
|||
* Register parser hooks
|
||||
*/
|
||||
public static function onParserFirstCallInit( &$parser ) {
|
||||
$parser->setHook( 'templatestyles', array( 'TemplateStylesHooks', 'render' ) );
|
||||
$parser->setHook( 'templatestyles', 'TemplateStylesHooks::render' );
|
||||
return true;
|
||||
}
|
||||
|
||||
private static function decodeFromBlob( $blob ) {
|
||||
$tree = gzdecode( $blob );
|
||||
if ( $tree ) {
|
||||
$tree = unserialize( $tree );
|
||||
}
|
||||
return $tree;
|
||||
}
|
||||
|
||||
private static function encodeToBlob( $tree ) {
|
||||
return gzencode( serialize( $tree ) );
|
||||
}
|
||||
|
||||
public static function onOutputPageParserOutput( &$out, $parseroutput ) {
|
||||
global $wgTemplateStylesNamespaces;
|
||||
if ( $wgTemplateStylesNamespaces )
|
||||
if ( $wgTemplateStylesNamespaces ) {
|
||||
$namespaces = $wgTemplateStylesNamespaces;
|
||||
else
|
||||
} else {
|
||||
$namespaces = [ NS_TEMPLATE ];
|
||||
}
|
||||
|
||||
$renderer = new CSSRenderer();
|
||||
$pages = [];
|
||||
|
||||
if ( $out->canUseWikiPage() )
|
||||
$pages[$out->getWikiPage()->getID()] = 'self';
|
||||
|
||||
foreach ( $namespaces as $ns )
|
||||
if ( array_key_exists( $ns, $parseroutput->getTemplates() ) )
|
||||
foreach ( $parseroutput->getTemplates()[$ns] as $title => $pageid )
|
||||
foreach ( $namespaces as $ns ) {
|
||||
if ( array_key_exists( $ns, $parseroutput->getTemplates() ) ) {
|
||||
foreach ( $parseroutput->getTemplates()[$ns] as $title => $pageid ) {
|
||||
$pages[$pageid] = $title;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if ( count( $pages ) ) {
|
||||
$db = wfGetDB( DB_SLAVE );
|
||||
|
|
@ -43,22 +56,27 @@ class TemplateStylesHooks {
|
|||
[ 'ORDER BY', 'pp_page' ]
|
||||
);
|
||||
foreach ( $res as $row ) {
|
||||
$css = unserialize( gzdecode( $row->pp_value ) );
|
||||
$css = self::decodeFromBlob( $row->pp_value );
|
||||
if ( $css ) {
|
||||
$renderer->add( $css );
|
||||
}
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
$selfcss = $out->getProperty( 'templatestyles' );
|
||||
if ( $selfcss ) {
|
||||
$selfcss = unserialize( gzdecode( $selfcss ) );
|
||||
$selfcss = self::decodeFromBlob( unserialize( gzdecode( $selfcss ) ) );
|
||||
if ( $selfcss ) {
|
||||
$renderer->add( $selfcss );
|
||||
}
|
||||
}
|
||||
|
||||
$css = $renderer->render();
|
||||
if ( $css )
|
||||
if ( $css ) {
|
||||
$out->addInlineStyle( $css );
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Parser hook for <templatestyles>.
|
||||
|
|
@ -77,9 +95,13 @@ class TemplateStylesHooks {
|
|||
public static function render( $input, $args, $parser, $frame ) {
|
||||
$css = new CSSParser( $input );
|
||||
|
||||
if ( $css )
|
||||
$parser->getOutput()->setProperty( 'templatestyles', gzencode( serialize( $css->rules() ) ) );
|
||||
if ( $css ) {
|
||||
$parser->getOutput()->setProperty( 'templatestyles', self::encodeToBlob( $css->rules() ) );
|
||||
}
|
||||
|
||||
// TODO: The UX would benefit from the CSS being run through the
|
||||
// hook for syntax highlighting rather that simply being presented
|
||||
// as a preformatted block.
|
||||
$html =
|
||||
Html::openElement( 'div', [ 'class' => 'mw-templatestyles-doc' ] )
|
||||
. Html::rawElement(
|
||||
|
|
|
|||
|
|
@ -2,7 +2,7 @@
|
|||
"license": "LGPL-2.1+",
|
||||
"require-dev": {
|
||||
"jakub-onderka/php-parallel-lint": "0.9",
|
||||
"mediawiki/mediawiki-codesniffer": "0.4.0"
|
||||
"mediawiki/mediawiki-codesniffer": "0.5.0"
|
||||
},
|
||||
"scripts": {
|
||||
"test": [
|
||||
|
|
|
|||
Loading…
Reference in a new issue