Refactor parsing logic

• Extract parser state into its own instance
• Every type should be able to parse its own structure

lib/Sabberworm/CSS/CSSList/CSSList.php

@@ -2,11 +2,22 @@
 
 namespace Sabberworm\CSS\CSSList;
 
+use Sabberworm\CSS\Comment\Commentable;
+use Sabberworm\CSS\Parsing\ParserState;
+use Sabberworm\CSS\Parsing\SourceException;
+use Sabberworm\CSS\Parsing\UnexpectedTokenException;
+use Sabberworm\CSS\Property\AtRule;
+use Sabberworm\CSS\Property\Charset;
+use Sabberworm\CSS\Property\CSSNamespace;
+use Sabberworm\CSS\Property\Import;
+use Sabberworm\CSS\Property\Selector;
 use Sabberworm\CSS\Renderable;
+use Sabberworm\CSS\RuleSet\AtRuleSet;
 use Sabberworm\CSS\RuleSet\DeclarationBlock;
 use Sabberworm\CSS\RuleSet\RuleSet;
-use Sabberworm\CSS\Property\Selector;
-use Sabberworm\CSS\Comment\Commentable;
+use Sabberworm\CSS\Value\CSSString;
+use Sabberworm\CSS\Value\URL;
+use Sabberworm\CSS\Value\Value;
 
 /**
  * A CSSList is the most generic container available. Its contents include RuleSet as well as other CSSList objects.
@@ -24,6 +35,140 @@ public function __construct($iLineNo = 0) {
 		$this->iLineNo = $iLineNo;
 	}
 
+	public static function parseList(ParserState $oParserState, CSSList $oList) {
+		$bIsRoot = $oList instanceof Document;
+		if(is_string($oParserState)) {
+			$oParserState = new ParserState($oParserState);
+		}
+		$bLenientParsing = $oParserState->getSettings()->bLenientParsing;
+		while(!$oParserState->isEnd()) {
+			$comments = $oParserState->consumeWhiteSpace();
+			$oListItem = null;
+			if($bLenientParsing) {
+				try {
+					$oListItem = self::parseListItem($oParserState, $oList);
+				} catch (UnexpectedTokenException $e) {
+					$oListItem = false;
+				}
+			} else {
+				$oListItem = self::parseListItem($oParserState, $oList);
+			}
+			if($oListItem === null) {
+				// List parsing finished
+				return;
+			}
+			if($oListItem) {
+				$oListItem->setComments($comments);
+				$oList->append($oListItem);
+			}
+			$oParserState->consumeWhiteSpace();
+		}
+		if(!$bIsRoot && !$bLenientParsing) {
+			throw new SourceException("Unexpected end of document", $oParserState->currentLine());
+		}
+	}
+
+	private static function parseListItem(ParserState $oParserState, CSSList $oList) {
+		$bIsRoot = $oList instanceof Document;
+		if ($oParserState->comes('@')) {
+			$oAtRule = self::parseAtRule($oParserState);
+			if($oAtRule instanceof Charset) {
+				if(!$bIsRoot) {
+					throw new UnexpectedTokenException('@charset may only occur in root document', '', 'custom', $oParserState->currentLine());
+				}
+				if(count($oList->getContents()) > 0) {
+					throw new UnexpectedTokenException('@charset must be the first parseable token in a document', '', 'custom', $oParserState->currentLine());
+				}
+				$oParserState->setCharset($oAtRule->getCharset()->getString());
+			}
+			return $oAtRule;
+		} else if ($oParserState->comes('}')) {
+			$oParserState->consume('}');
+			if ($bIsRoot) {
+				if ($oParserState->getSettings()->bLenientParsing) {
+					while ($oParserState->comes('}')) $oParserState->consume('}');
+					return DeclarationBlock::parse($oParserState);
+				} else {
+					throw new SourceException("Unopened {", $oParserState->currentLine());
+				}
+			} else {
+				return null;
+			}
+		} else {
+			return DeclarationBlock::parse($oParserState);
+		}
+	}
+
+	private static function parseAtRule(ParserState $oParserState) {
+		$oParserState->consume('@');
+		$sIdentifier = $oParserState->parseIdentifier();
+		$iIdentifierLineNum = $oParserState->currentLine();
+		$oParserState->consumeWhiteSpace();
+		if ($sIdentifier === 'import') {
+			$oLocation = URL::parse($oParserState);
+			$oParserState->consumeWhiteSpace();
+			$sMediaQuery = null;
+			if (!$oParserState->comes(';')) {
+				$sMediaQuery = $oParserState->consumeUntil(';');
+			}
+			$oParserState->consume(';');
+			return new Import($oLocation, $sMediaQuery, $iIdentifierLineNum);
+		} else if ($sIdentifier === 'charset') {
+			$sCharset = CSSString::parse($oParserState);
+			$oParserState->consumeWhiteSpace();
+			$oParserState->consume(';');
+			return new Charset($sCharset, $iIdentifierLineNum);
+		} else if (self::identifierIs($sIdentifier, 'keyframes')) {
+			$oResult = new KeyFrame($iIdentifierLineNum);
+			$oResult->setVendorKeyFrame($sIdentifier);
+			$oResult->setAnimationName(trim($oParserState->consumeUntil('{', false, true)));
+			CSSList::parseList($oParserState, $oResult);
+			return $oResult;
+		} else if ($sIdentifier === 'namespace') {
+			$sPrefix = null;
+			$mUrl = Value::parsePrimitiveValue($oParserState);
+			if (!$oParserState->comes(';')) {
+				$sPrefix = $mUrl;
+				$mUrl = Value::parsePrimitiveValue($oParserState);
+			}
+			$oParserState->consume(';');
+			if ($sPrefix !== null && !is_string($sPrefix)) {
+				throw new UnexpectedTokenException('Wrong namespace prefix', $sPrefix, 'custom', $iIdentifierLineNum);
+			}
+			if (!($mUrl instanceof CSSString || $mUrl instanceof URL)) {
+				throw new UnexpectedTokenException('Wrong namespace url of invalid type', $mUrl, 'custom', $iIdentifierLineNum);
+			}
+			return new CSSNamespace($mUrl, $sPrefix, $iIdentifierLineNum);
+		} else {
+			//Unknown other at rule (font-face or such)
+			$sArgs = trim($oParserState->consumeUntil('{', false, true));
+			$bUseRuleSet = true;
+			foreach(explode('/', AtRule::BLOCK_RULES) as $sBlockRuleName) {
+				if(self::identifierIs($sIdentifier, $sBlockRuleName)) {
+					$bUseRuleSet = false;
+					break;
+				}
+			}
+			if($bUseRuleSet) {
+				$oAtRule = new AtRuleSet($sIdentifier, $sArgs, $iIdentifierLineNum);
+				RuleSet::parseRuleSet($oParserState, $oAtRule);
+			} else {
+				$oAtRule = new AtRuleBlockList($sIdentifier, $sArgs, $iIdentifierLineNum);
+				CSSList::parseList($oParserState, $oAtRule);
+			}
+			return $oAtRule;
+		}
+	}
+
+		/**
+	 * Tests an identifier for a given value. Since identifiers are all keywords, they can be vendor-prefixed. We need to check for these versions too.
+	 */
+	private static function identifierIs($sIdentifier, $sMatch) {
+		return (strcasecmp($sIdentifier, $sMatch) === 0)
+			?: preg_match("/^(-\\w+-)?$sMatch$/i", $sIdentifier) === 1;
+	}
+
+
 	/**
 	 * @return int
 	 */

lib/Sabberworm/CSS/CSSList/Document.php

@@ -2,6 +2,8 @@
 
 namespace Sabberworm\CSS\CSSList;
 
+use Sabberworm\CSS\Parsing\ParserState;
+
 /**
  * The root CSSList of a parsed file. Contains all top-level css contents, mostly declaration blocks, but also any @-rules encountered.
  */
@@ -14,6 +16,12 @@ public function __construct($iLineNo = 0) {
 		parent::__construct($iLineNo);
 	}
 
+	public static function parse(ParserState $oParserState) {
+		$oDocument = new Document($oParserState->currentLine());
+		CSSList::parseList($oParserState, $oDocument);
+		return $oDocument;
+	}
+
 	/**
 	 * Gets all DeclarationBlock objects recursively.
 	 */