Merge "fetchLexers: Pass '--json' to Pygmentize"

README

@@ -6,8 +6,8 @@ for updated information:
 
 == Requirements ==
 
-This version of the extension has been tested with Pygments 2.8.0 and
-MediaWiki 1.35. To get releases of this extension compatible
+This version of the extension has been tested with Pygments 2.11.2 and
+MediaWiki 1.36. To get releases of this extension compatible
 with earlier versions of MediaWiki, visit:
 
     https://www.mediawiki.org/wiki/Special:ExtensionDistributor/SyntaxHighlight_GeSHi
@@ -18,7 +18,7 @@ Add this line to your LocalSettings.php:
 
     wfLoadExtension( 'SyntaxHighlight_GeSHi' );
 
-By default, this extension will use a bundled copy of Pygments 2.8.0. If you
+By default, this extension will use a bundled copy of Pygments 2.11.2. If you
 would like to use a different copy of the library, you can set
 $wgPygmentizePath to point to the path to the 'pygmentize' binary.
 

includes/Pygmentize.php

@@ -188,6 +188,16 @@ class Pygmentize {
 		);
 	}
 
+	/**
+	 * Determine if the pygments command line supports the --json option
+	 *
+	 * @return bool
+	 */
+	private static function pygmentsSupportsJsonOutput(): bool {
+		$version = self::getVersion();
+		return ( version_compare( $version, '2.11.0' ) !== -1 );
+	}
+
 	/**
 	 * Shell out to get supported lexers by pygments
 	 *
@@ -195,8 +205,13 @@ class Pygmentize {
 	 * @return array
 	 */
 	public static function fetchLexers(): array {
+		$cliParams = [ self::getPath(), '-L', 'lexer' ];
+		if ( self::pygmentsSupportsJsonOutput() ) {
+			$cliParams[] = '--json';
+		}
+
 		$result = self::boxedCommand()
-			->params( self::getPath(), '-L', 'lexer' )
+			->params( $cliParams )
 			->includeStderr()
 			->execute();
 		self::recordShellout( 'fetch_lexers' );
@@ -205,9 +220,50 @@ class Pygmentize {
 			throw new PygmentsException( $output );
 		}
 
-		// Post-process the output, ideally pygments would output this in a
-		// machine-readable format (https://github.com/pygments/pygments/issues/1437)
-		$output = $result->getStdout();
+		if ( self::pygmentsSupportsJsonOutput() ) {
+			$lexers = self::parseLexersFromJson( $output );
+		} else {
+			$lexers = self::parseLexersFromText( $output );
+		}
+
+		$lexers = array_unique( $lexers );
+		sort( $lexers );
+		$data = [];
+		foreach ( $lexers as $lexer ) {
+			$data[$lexer] = true;
+		}
+
+		return $data;
+	}
+
+	/**
+	 * Parse json output of the pygments lexers list and return as php array
+	 *
+	 * @param string $output JSON formatted output of pygments lexers list
+	 * @return array
+	 */
+	private static function parseLexersFromJson( $output ): array {
+		$data = json_decode( $output, true );
+		if ( $data === null ) {
+			throw new PygmentsException(
+				'Got invalid JSON from Pygments: ' . $output );
+		}
+		$lexers = [];
+		foreach ( array_values( $data['lexers'] ) as $lexer ) {
+			$lexers = array_merge( $lexers, $lexer['aliases'] );
+		}
+		return $lexers;
+	}
+
+	/**
+	 * Parse original stdout of the pygments lexers list
+	 * This was the only format available before pygments 2.11.0
+	 * NOTE: Should be removed when pygments 2.11 is the minimum version expected to be installed
+	 *
+	 * @param string $output Textual list of pygments lexers
+	 * @return array
+	 */
+	private static function parseLexersFromText( $output ): array {
 		$lexers = [];
 		foreach ( explode( "\n", $output ) as $line ) {
 			if ( substr( $line, 0, 1 ) === '*' ) {
@@ -219,14 +275,7 @@ class Pygmentize {
 				}
 			}
 		}
-		$lexers = array_unique( $lexers );
-		sort( $lexers );
-		$data = [];
-		foreach ( $lexers as $lexer ) {
-			$data[$lexer] = true;
-		}
-
-		return $data;
+		return $lexers;
 	}
 
 	/**