Author: burton Date: Sat Jan 29 16:51:57 2005 New Revision: 149089 URL: http://svn.apache.org/viewcvs?view=rev&rev=149089 Log: support for RSS 2.0 enclosures... Added: jakarta/commons/sandbox/feedparser/trunk/tests/feeds/rss-2.0-enclosure.rss Modified: jakarta/commons/sandbox/feedparser/trunk/TODO jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/AtomFeedParser.java jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/BaseParser.java jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/RSSFeedParser.java jakarta/commons/sandbox/feedparser/trunk/xdocs/index.xml
Modified: jakarta/commons/sandbox/feedparser/trunk/TODO Url: http://svn.apache.org/viewcvs/jakarta/commons/sandbox/feedparser/trunk/TODO?view=diff&rev=149089&p1=jakarta/commons/sandbox/feedparser/trunk/TODO&r1=149088&p2=jakarta/commons/sandbox/feedparser/trunk/TODO&r2=149089 ============================================================================== --- jakarta/commons/sandbox/feedparser/trunk/TODO (original) +++ jakarta/commons/sandbox/feedparser/trunk/TODO Sat Jan 29 16:51:57 2005 @@ -18,6 +18,10 @@ - (DONE) Remove ALL references to newsmonster IO layer. +- (DONE) Support MetaFeedParser.onCreated() for RSS 2.0 /rss/channel/item/pubdate + +- (DONE) Implement RSS 2.0 enclosure linkage ... this should be an onLink handler. + - Atom GUIDs and RSS 2.0 GUIDs - http://www.ietf.org/internet-drafts/draft-ietf-atompub-format-02.txt @@ -28,15 +32,13 @@ http://diveintomark.org/tests/client/autodiscovery/ -- Support Base64 Atom values and the ability to enable them. +- Support Base64 Atom values and the ability to enable them to be automatically decoded. - Atom's xml:base is NOT supported right now. We NEED to support this. - Do we support multiple content items in Atom? - We do not support multipart/alternative in the feedparser. - -- Implement RSS 2.0 enclosure linkage ... this should be an onLink handler. - Rework the factory mechanism to support multiple FeedParsers... should be an interface. Modified: jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/AtomFeedParser.java Url: http://svn.apache.org/viewcvs/jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/AtomFeedParser.java?view=diff&rev=149089&p1=jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/AtomFeedParser.java&r1=149088&p2=jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/AtomFeedParser.java&r2=149089 ============================================================================== --- jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/AtomFeedParser.java (original) +++ jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/AtomFeedParser.java Sat Jan 29 16:51:57 2005 @@ -375,28 +375,5 @@ } - //FIXME: unify this with RSSFeedParser.getChildElementTextByName - private static String selectText( String query, Element element ) throws Exception { - - XPath xpath = new XPath( query ); - xpath.setNamespaceContext( NS.context ); - - //perform onChannel method... (title, link, description) - Element e = (Element)xpath.selectSingleNode( element ); - - if ( e == null ) - return null; - - String result = e.getText(); - - //The normalize method of XML SHOULD take care of this but for some - //reason it doesnt. - if ( result != null ) - result = result.trim(); - - return result; - - } - } Modified: jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/BaseParser.java Url: http://svn.apache.org/viewcvs/jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/BaseParser.java?view=diff&rev=149089&p1=jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/BaseParser.java&r1=149088&p2=jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/BaseParser.java&r2=149089 ============================================================================== --- jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/BaseParser.java (original) +++ jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/BaseParser.java Sat Jan 29 16:51:57 2005 @@ -38,7 +38,7 @@ * Basic parser support common between RSS, Atom, and FOAF feed impls. * * @author <a href="mailto:[EMAIL PROTECTED]">Kevin A. Burton (burtonator)</a> - * @version $Id: BaseParser.java,v 1.2 2004/10/18 00:00:09 burton Exp $ + * @version $Id$ */ public class BaseParser { @@ -102,5 +102,57 @@ return null; } - + + // **** shared code for all parsers ***************************************** + + //FIXME: unify this with RSSFeedParser.getChildElementTextByName + protected static String selectText( String query, Element element ) throws Exception { + + XPath xpath = new XPath( query ); + xpath.setNamespaceContext( NS.context ); + + //perform onChannel method... (title, link, description) + Element e = (Element)xpath.selectSingleNode( element ); + + if ( e == null ) + return null; + + String result = e.getText(); + + //The normalize method of XML SHOULD take care of this but for some + //reason it doesnt. + if ( result != null ) + result = result.trim(); + + return result; + + } + + /** + * Regardless of namespace, get the child node text by name or null if it is not found. + * + * @author <a href="mailto:[EMAIL PROTECTED]">Kevin A. Burton</a> + */ + protected static String getChildElementTextByName( FeedParserState state, + String name ) throws Exception { + + //FIXME: this can be rewritten to use getChild() + + XPath xpath = new XPath( "descendant::*[local-name() = '" + name + "']" ); + Object resultNode = xpath.selectSingleNode( state.current ); + + String resultText = null; + + if ( resultNode != null ) + resultText = ((Element)resultNode).getText(); + + //The normalize method of XML SHOULD take care of this but for some + //reason it doesnt. + if ( resultText != null ) + resultText = resultText.trim(); + + return resultText; + + } + } Modified: jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/RSSFeedParser.java Url: http://svn.apache.org/viewcvs/jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/RSSFeedParser.java?view=diff&rev=149089&p1=jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/RSSFeedParser.java&r1=149088&p2=jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/RSSFeedParser.java&r2=149089 ============================================================================== --- jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/RSSFeedParser.java (original) +++ jakarta/commons/sandbox/feedparser/trunk/src/java/org/apache/commons/feedparser/RSSFeedParser.java Sat Jan 29 16:51:57 2005 @@ -36,7 +36,7 @@ * Handles parsing RSS . * * @author <a href="mailto:[EMAIL PROTECTED]">Kevin A. Burton (burtonator)</a> - * @version $Id: RSSFeedParser.java,v 1.15 2005/01/23 09:49:50 burton Exp $ + * @version $Id$ */ public class RSSFeedParser extends BaseParser { @@ -258,36 +258,39 @@ MetaFeedParser.parse( listener, state ); TagFeedParser.parse( listener, state ); + + doEnclosures( listener, state ); listener.onItemEnd(); } - - /** - * Regardless of namespace, get the child node text by name or null if it is not found. - * - * @author <a href="mailto:[EMAIL PROTECTED]">Kevin A. Burton</a> - */ - public static String getChildElementTextByName( FeedParserState state, - String name ) throws Exception { - //FIXME: this can be rewritten to use getChild() - - XPath xpath = new XPath( "descendant::*[local-name() = '" + name + "']" ); - Object resultNode = xpath.selectSingleNode( state.current ); + private static void doEnclosures( FeedParserListener listener, + FeedParserState state ) throws Exception { + + if ( listener instanceof LinkFeedParserListener == false ) + return; - String resultText = null; + Element element = state.current.getChild( "enclosure" ); - if ( resultNode != null ) - resultText = ((Element)resultNode).getText(); + if ( element == null ) + return; - //The normalize method of XML SHOULD take care of this but for some - //reason it doesnt. - if ( resultText != null ) - resultText = resultText.trim(); + LinkFeedParserListener linkFeedParserListener = (LinkFeedParserListener)listener; - return resultText; + String rel = null; + String type = element.getAttributeValue( "type" ); + String href = element.getAttributeValue( "url" ); + String title = null; + long length = Integer.parseInt( element.getAttributeValue( "length" ) ); + + linkFeedParserListener.onLink( state, + rel, + type, + href, + title, + length ); } - + } Added: jakarta/commons/sandbox/feedparser/trunk/tests/feeds/rss-2.0-enclosure.rss Url: http://svn.apache.org/viewcvs/jakarta/commons/sandbox/feedparser/trunk/tests/feeds/rss-2.0-enclosure.rss?view=auto&rev=149089 ============================================================================== --- (empty file) +++ jakarta/commons/sandbox/feedparser/trunk/tests/feeds/rss-2.0-enclosure.rss Sat Jan 29 16:51:57 2005 @@ -0,0 +1,30 @@ +<?xml version="1.0"?> +<!-- RSS generated by UserLand Frontier v9.0.1 on 10/17/2004; 2:40:35 PM Pacific --> +<rss version="2.0"> + <channel> + <title>Scripting News</title> + <link>http://www.scripting.com/</link> + <description>It's even worse than it appears.</description> + <language>en-us</language> + <copyright>Copyright 1997-2004 Dave Winer</copyright> + <pubDate>Sun, 17 Oct 2004 07:00:00 GMT</pubDate> + <lastBuildDate>Sun, 17 Oct 2004 21:40:35 GMT</lastBuildDate> + <docs>http://blogs.law.harvard.edu/tech/rss</docs> + <generator>UserLand Frontier v9.0.1</generator> + <managingEditor>[EMAIL PROTECTED]</managingEditor> + <webMaster>[EMAIL PROTECTED]</webMaster> + + <item> + <description><a href="http://static.podcatch.com/manila/gems/un/eps.mp3">This is a test</a>. For the next sixty seconds this station will conduct a test of the Emergency Podcast System. </description> + <pubDate>Sun, 17 Oct 2004 17:36:38 GMT</pubDate> + <guid>http://archive.scripting.com/2004/10/17#When:10:36:38AM</guid> + <enclosure url="http://static.podcatch.com/manila/gems/un/eps.mp3"; length="189455" type="audio/mpeg" /> + </item> + <item> + <description><a href="http://static.podcatch.com/manila/gems/un/anotherTestAudioBlogPost.mp3">I got another</a> test blog post. An audio test blog post. Pay no attention to the man behind the curtain.</description> + <pubDate>Sun, 17 Oct 2004 19:11:14 GMT</pubDate> + <guid>http://archive.scripting.com/2004/10/17#When:12:11:14PM</guid> + <enclosure url="http://static.podcatch.com/manila/gems/un/anotherTestAudioBlogPost.mp3"; length="106423" type="audio/mpeg" /> + </item> + </channel> +</rss> Modified: jakarta/commons/sandbox/feedparser/trunk/xdocs/index.xml Url: http://svn.apache.org/viewcvs/jakarta/commons/sandbox/feedparser/trunk/xdocs/index.xml?view=diff&rev=149089&p1=jakarta/commons/sandbox/feedparser/trunk/xdocs/index.xml&r1=149088&p2=jakarta/commons/sandbox/feedparser/trunk/xdocs/index.xml&r2=149089 ============================================================================== --- jakarta/commons/sandbox/feedparser/trunk/xdocs/index.xml (original) +++ jakarta/commons/sandbox/feedparser/trunk/xdocs/index.xml Sat Jan 29 16:51:57 2005 @@ -141,6 +141,10 @@ Helps enables tags within RSS feeds </dd> + <dd> + RSS 2.0 enclosures + </dd> + </dl> </section> --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
