394 lines
No EOL
56 KiB
HTML
394 lines
No EOL
56 KiB
HTML
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
|
|
<!-- Generated by Apache Maven Doxia Site Renderer 1.4 at 2013-10-04 -->
|
|
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
|
|
<head>
|
|
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />
|
|
<title>ROME - Change Log</title>
|
|
<style type="text/css" media="all">
|
|
@import url("./css/maven-base.css");
|
|
@import url("./css/maven-theme.css");
|
|
@import url("./css/site.css");
|
|
</style>
|
|
<link rel="stylesheet" href="./css/print.css" type="text/css" media="print" />
|
|
<meta name="author" content="mkurz" />
|
|
<meta name="Date-Creation-yyyymmdd" content="20110815" />
|
|
<meta name="Date-Revision-yyyymmdd" content="20131004" />
|
|
<meta http-equiv="Content-Language" content="en" />
|
|
|
|
</head>
|
|
<body class="composite">
|
|
<div id="banner">
|
|
<a href="http://github.com/rometools/" id="bannerLeft">
|
|
<img src="images/romelogo.png" alt="ROME" />
|
|
</a>
|
|
<div class="clear">
|
|
<hr/>
|
|
</div>
|
|
</div>
|
|
<div id="breadcrumbs">
|
|
|
|
|
|
<div class="xright">
|
|
|
|
<span id="publishDate">Last Published: 2013-10-04</span>
|
|
| <span id="projectVersion">Version: 2.0.0-SNAPSHOT</span>
|
|
</div>
|
|
<div class="clear">
|
|
<hr/>
|
|
</div>
|
|
</div>
|
|
<div id="leftColumn">
|
|
<div id="navcolumn">
|
|
|
|
|
|
<h5>Rome</h5>
|
|
<ul>
|
|
<li class="none">
|
|
<a href="index.html" title="Overview">Overview</a>
|
|
</li>
|
|
<li class="collapsed">
|
|
<a href="HowRomeWorks/index.html" title="How Rome Works">How Rome Works</a>
|
|
</li>
|
|
<li class="none">
|
|
<a href="RssAndAtOMUtilitiEsROMEV0.5AndAboveTutorialsAndArticles/index.html" title="Tutorials And Articles">Tutorials And Articles</a>
|
|
</li>
|
|
<li class="collapsed">
|
|
<a href="ROMEReleases/index.html" title="Releases">Releases</a>
|
|
</li>
|
|
<li class="none">
|
|
<a href="ROMEDevelopmentProposals/index.html" title="ROME Development Proposals">ROME Development Proposals</a>
|
|
</li>
|
|
</ul>
|
|
<h5>Project Documentation</h5>
|
|
<ul>
|
|
<li class="collapsed">
|
|
<a href="project-info.html" title="Project Information">Project Information</a>
|
|
</li>
|
|
<li class="collapsed">
|
|
<a href="project-reports.html" title="Project Reports">Project Reports</a>
|
|
</li>
|
|
</ul>
|
|
<a href="http://maven.apache.org/" title="Built by Maven" class="poweredBy">
|
|
<img class="poweredBy" alt="Built by Maven" src="./images/logos/maven-feather.png" />
|
|
</a>
|
|
|
|
|
|
</div>
|
|
</div>
|
|
<div id="bodyColumn">
|
|
<div id="contentBox">
|
|
<div class="section">
|
|
<h2>Change Log<a name="Change_Log"></a></h2>
|
|
<div class="section">
|
|
<h3>Changes made since v1.0<a name="Changes_made_since_v1.0"></a></h3>
|
|
<ol style="list-style-type: decimal">
|
|
<li><a class="externalLink" href="http://java.net/jira/browse/ROME-127">Issue 127</a>: Rome 1.0 not JDK 1.4 compatible</li></ol></div>
|
|
<div class="section">
|
|
<h3>Changes made since v1.0RC2<a name="Changes_made_since_v1.0RC2"></a></h3>
|
|
<ol style="list-style-type: decimal">
|
|
<li><a class="externalLink" href="http://java.net/jira/browse/ROME-121">Issue 121</a>: RSS item category iteration should try to reflect document order</li>
|
|
<li>New property preserveWireFeed available on SyndFeedInput<br />WireFeeds will be preserved if the property preserveWireFeed is set on the SyndFeedInput object it is built from. Atom/RSS Entry/Item objects are also available from SyndEntry objects if the WireFeed is preserved using the new getWireEntry() method. See <a href="./PreservingWireFeeds.html">Preserving WireFeeds (rome)</a> for details.</li></ol></div>
|
|
<div class="section">
|
|
<h3>Changes made since v1.0RC1<a name="Changes_made_since_v1.0RC1"></a></h3>
|
|
<ol style="list-style-type: decimal">
|
|
<li>Fix. Date parsing for Atom10 entry and additional W3C masks<br />Item date elements were being parsed with the W3C parser instead the lenient one (RFC822 + W3C + custom masks).<br />The following masks were added to W3C masks to handle RFC822 timezone (ie '-800'):
|
|
<div class="source">
|
|
<pre>
|
|
yyyy-MM-dd'T'HH:mm:ssZ yyyy-MM-dd't'HH:mm:sszZ
|
|
</pre></div></li>
|
|
<li>Fix. Contributors properties in SyndEntry were not implementing the semantics of list properties.<br />They were returning NULL instead, now they return an empty list if not values are set.</li>
|
|
<li>Fix. Contributors properties in SyndEntry and SyndFeed were not being converted to/from WireFeed</li>
|
|
<li>Fix. Syndication Module Generator was failing if some of its values were null.<br />Checks for nulll have been added it to the generator to prevent NullPointerExceptions</li>
|
|
<li>New. Added new constructor to XmlReader
|
|
<div class="source">
|
|
<pre>
|
|
public XmlReader(InputStream is, boolean lenient, String defaultEncoding)
|
|
</pre></div></li>
|
|
<li>New. Support atom person construct extensions, using the Extendable interface on SyndPerson:<br />Patch from James Roper. See <a class="externalLink" href="http://java.net/jira/browse/ROME-110">Issue 1101</a> for details</li>
|
|
<li>New. Maven 2 build for main project<br />ROME can now be built with Maven 2</li>
|
|
<li>New. OSGi support<br />OSGi headers to MANIFEST.MF so that rome.jar can also be used in an OSGi environment. See <a class="externalLink" href="http://java.net/jira/browse/ROME-117">Issue 117</a> for details.</li>
|
|
<li>New. Allow pretty printing to be turned on and off<br />see <a class="externalLink" href="http://java.net/jira/browse/ROME-114">Issue 114</a> for details. Thanks to Martin Kurz for the patch.</li>
|
|
<li>Configurable classloading behavior for OSGi compatibility.<br />We have received a report of some issues with plugin loading in an OSGi environment (<a class="externalLink" href="http://java.net/jira/browse/ROME-118">Issue 118</a>). The fix appears to be to change Class.forName to classLoader.loadClass, but the semantics for this are subtly different, so we have made this new behavior user selectable. Set the "rome.pluginmanager.useloadclass" system property to "true" to enable it.</li>
|
|
<li>More lenient number parsing<br />There were a number of problems with feeds providing blank or invalid values in fields which would be numbers. ROME will now handles these better. See issues <a class="externalLink" href="http://java.net/jira/browse/ROME-104">104</a>, <a class="externalLink" href="http://java.net/jira/browse/ROME-107">107</a> and <a class="externalLink" href="http://java.net/jira/browse/ROME-108">108</a> for details.</li></ol></div>
|
|
<div class="section">
|
|
<h3>Changes made from v0.9 to v1.0RC1<a name="Changes_made_from_v0.9_to_v1.0RC1"></a></h3>
|
|
<ol style="list-style-type: decimal">
|
|
<li>New. XmlReader support for default encoding<br />The XmlReader can be set with an alternate default encoding in case no encoding has been detected from the transport (HTTP), the stream or the XML prolog. if no value is set the default fallback rules based on the content-type will be used. The alternate default encoding can be set/viewed via a static methods, <b>setDefaultEncoding()</b> and <b>getDefaultEncoding()</b>.</li>
|
|
<li>Fix. Atom 1.0 links were generated without title and length attributes.<br />The Atom 1.0 Generator was not generating title and length attributes when values are present.</li>
|
|
<li>Fix. XmlReader, multi-line prolog encoding detection.<br />XmlReader handles properly xml-prolog detection when prolog goes over multiple lies (such as G groups feeds).</li>
|
|
<li>Fix. Base64 decoding was failing under certain padding conditions.</li>
|
|
<li>Fix. XmlReader fixes<br />Fixed bug that if BOM is UTF8 was not being set to UTF8. Changed logic to use Buffered stream instead pushback stream for all encoding detection. Changed logic of xml prolog detection to avoid having a buffer with half of a unicode character (instead filling up the buffer looking up to first '>' which means it a valid buffer).</li>
|
|
<li>New. XmlReader supports default encoding at instance level.<br />Via a new constructor is possible to indicate a default encoding different than the default encoding at class level.</li>
|
|
<li>Fix. Making the EqualsBean to follow equals contract.<br />For X.equals(null) it was throwing a NullPointerException, now it returns FALSE.</li>
|
|
<li>Fix. Render Atom icon and logo attributes.<br />AtomGenerator now adds icon and logo elements to xml tree</li>
|
|
<li>Fix. Updated AtomPub namespace to its permenent home.<br />AtomService namespace updated to <a class="externalLink" href="http://www.w3.org/2007/app">http://www.w3.org/2007/app</a></li>
|
|
<li>New. Added support for configuration per classloader level.<br />The PluginManager (handles Parsers and Generators) now is singleton at classloader level allowing different configurations in different classloaders.</li>
|
|
<li>Atom parser: better relative URI handling<br />Instead of simply resolving each relative URI at runtime and saving only the resolved one, we now save both the relative URI and the resolve one. We introduced the following new methods to provide access to the resolved URI.
|
|
<ul>
|
|
<li>Link.getLinkResolved()</li>
|
|
<li>Link.setLinkResolved()</li>
|
|
<li>Category.getSchemeResolved()</li>
|
|
<li>Category.setSchemeResolved()</li>
|
|
<li>Person.getUriResolved()</li>
|
|
<li>Person.setUriResolved()</li></ul></li>
|
|
<li>Utility methods useful in working with Atom protocol feeds<br />Added a couple of methods to make it easier to deal with Atompub feeds.
|
|
<ul>
|
|
<li>Entry.isMediaEntry()</li>
|
|
<li>Atom10Parser.parseEntry()</li>
|
|
<li>Atom10Generator.serializeEntry()</li></ul></li>
|
|
<li>Bugs fixed<br />Fixed the following bugs:
|
|
<ul>
|
|
<li>49 Better content/summary mapping</li>
|
|
<li>53 Content.setType not working with subtitles atom 1.0</li>
|
|
<li>56 fix of bug #39 leads to invalid atom feeds</li>
|
|
<li>63 Missing link attribute when generating Atom 1.0</li>
|
|
<li>64 ROME's Atom parser doesn't pick up multiple alt links</li>
|
|
<li>65 Atom feeds not including logo image</li>
|
|
<li>71 encoding problem in XmlReader.getXmlProlog()</li>
|
|
<li>79 Feed.setIcon()/setLogo() ignored by Atom10Generator</li>
|
|
<li>81 SyndFeedImpl.equals() does not obey equals contract</li></ul></li>
|
|
<li>Fix. Parsers where ignoring namespaced prefixed Attributes.<br />If an XML feed uses a prefix for the Atom elements and the attributes of Atom elements use the prefix the parser was not picking up those attributes.<br />The fix makes the parser to look for prefixed and non-prefixed attributes.</li>
|
|
<li>Fix. Atom Feed and Entry beans author and category property getters<br />They were returning NULL when there were not authors or categories, they must return an empty list.</li>
|
|
<li>New. Switch to enable/disable relative URI resolution in Atom 1.0 Parser.<br />The Atom10Parser class has a static method, setResolveURIs(boolean) that enables/disables relative URI resolution.</li>
|
|
<li>New. XmlReader handling content-type charset values has been relaxed.<br />XmlReader handles content-type charset encoding value within single quotes and double quotes.</li>
|
|
<li>Fix. Links, authors and contributors properties in SyndFeed were not implementing the semantics of list properties.<br />They were returning NULL instead, now they return an empty list if not values are set.</li>
|
|
<li>Fix. RSS conversion of a SyndFeed was losing the link of the feed if the links property was used instead the link property.<br />Over time the SyndFeed has been modified to support more Atom specific properties and their cardinality, conversion to RSS of these properties was not always taken care.<br />The RSS converter has been changed so the link from SyndFeed is taken as channel link and if not set the first value of the links property is taken.</li>
|
|
<li>Fix. WireFeedInput throws IllegalArgumentException if the feed type is not recognized.<br />Previously the IllegalArgumentException was wrapped by a ParsingFeedException (Reported by <a class="externalLink" href="http://java.net/jira/browse/ROME-91">Issue 91</a>).</li>
|
|
<li>Fix. SyndFeedImpl.equals(other) checks for instance of other before casting.<br />The underlying ObjectBean does this check, but in this method a cast is done before to obtain the foreign markup, no the instance check is peformed before to avoid a class cast exception.</li>
|
|
<li>Fix. Atom content based elements related fixes
|
|
<ul>
|
|
<li>Atom 0.3 Parser/Generator
|
|
<ul>
|
|
<li>Changed title to be treated as a Content construct. (<a class="externalLink" href="http://www.mnot.net/drafts/draft-nottingham-atom-format-02.html#rfc.section.4.3">http://www.mnot.net/drafts/draft-nottingham-atom-format-02.html#rfc.section.4.3</a>)</li></ul></li>
|
|
<li>Atom 1.0 Generator:
|
|
<ul>
|
|
<li>changed feed title/subtitle and entry title to be treated as Content constructs. (Parser had this implemented already.)</li>
|
|
<li>added title attribute to links. (Parser had this implemented already.)</li>
|
|
<li>fixed content parsing for some XML content types. e.g. (application/xhtml+xml)</li></ul></li></ul></li>
|
|
<li>Fix. Atom link and enclosures handling
|
|
<ul>
|
|
<li>Atom 0.3 Converter
|
|
<ul>
|
|
<li>fixed link parsing code to parse all links (not just the first alternate link) and added enclosure support via link rel="enclosure".</li>
|
|
<li>changed title conversion to use Content instead of plain text.</li></ul></li>
|
|
<li>Atom 1.0 Converter
|
|
<ul>
|
|
<li>added SyndEnclosure to atom:link rel=enclosure conversion.</li></ul></li></ul></li>
|
|
<li>Fix. RSS 1.0 URI generation
|
|
<ul>
|
|
<li>RSS 1.0 Generator
|
|
<ul>
|
|
<li>channel/items/Seq/li/@resource now get's the item URI instead of the Link. (<a class="externalLink" href="http://web.resource.org/rss/1.0/spec#s5.3.5">http://web.resource.org/rss/1.0/spec#s5.3.5</a>)</li></ul></li></ul></li>
|
|
<li>Fix. Javadocs corrections.
|
|
<ul>
|
|
<li>Fixed some javadoc comments for SyndEntry.</li></ul></li>
|
|
<li>Fix. Atom content based elements were not parsed with XML mime types.<br />If the mime type was and XML mime type the content value was being lost on parsing.</li>
|
|
<li>Fix. duplication of content:encoded elements when reading/writing and RSS feed.<br />content:encoded elements are treated special, without a module, they have to be removed from the foreign markup to avoid duplication in case of read/write. Note that this fix will break if a content module is used.</li>
|
|
<li>New. XmlFixerReader converts '&' into '&' when there is no matching entity.<br />Feeds commonly use '&' instead '&' in their content, this change converts those orphant '&'s into '&'s.</li>
|
|
<li>Fix. RSS090Parser does not set the URI property.<br />The fix honors the documentation "For RSS 0.91, RSS 0.92, RSS 0.93 & RSS 1.0 ... the SyndEntry uri property will be set with the value of the link element..."</li>
|
|
<li>New. Removal of all unused namespaces from generated feeds.<br />The generators now remove all unused namespaces from the XML document before generating it.</li></ol></div>
|
|
<div class="section">
|
|
<h3>Changes made from v0.8 to v0.9 <a name="Changes_made_from_v0.8_to_v0.9"></a></h3>
|
|
<ol style="list-style-type: decimal">
|
|
<li>Design changes
|
|
<ul>
|
|
<li>Support Atom feed.title, feed.subtitle and entry.title <a class="externalLink" href="http://java.net/jira/browse/ROME-48">Issue 48</a><br />#48 fixed via better support for Atom text constructs title and subtitle. Added get/setTitleEx() and get/setSubtitleEx(), which get get SyndContent objects. Title and subtitle still available from old getters/setters.</li>
|
|
<li>Support for mapping Atom summary/content to RSS description/content <a class="externalLink" href="https://rome.dev.java.net/servlets/ReadMsg?list=dev&msgNo=1680">https://rome.dev.java.net/servlets/ReadMsg?list=dev&msgNo=1680</a></li>
|
|
<li>Fixed by introduced Content object in RSS model. ROME now parses as RSS Content. That makes parsing easier and allows us to support a more logical summary/content mapping:
|
|
<ul>
|
|
<li>RSS to/from Atom</li>
|
|
<li>RSS to/from Atom</li></ul></li></ul></li>
|
|
<li>General parsing fixes
|
|
<ul>
|
|
<li>XmlReader xml prolog regular expression does not allow for single quotes <a class="externalLink" href="http://java.net/jira/browse/ROME-36">Issue 36</a><br />The XmlReader was only parsing prolog encodings within double quotes, the regular expression to detect the encoding has been change to detect single or double quotes.</li>
|
|
<li>Fix. XML prolog parsing now support whitespaces around '='<br />If the XML prolog contained spaces around the '=' between the encoding attribute name and the encoding attribute value the encoding was not being detected. The fix accepts all valid whitespace characters (as defined in the XML spec).</li>
|
|
<li>RSS parser does not recognize version="2.00" <a class="externalLink" href="http://java.net/jira/browse/ROME-33">Issue 33</a></li>
|
|
<li>Atom 1.0 Text Types Not Set Correctly <a class="externalLink" href="http://java.net/jira/browse/ROME-39">Issue 39</a></li>
|
|
<li>Security issue <a class="externalLink" href="http://java.net/jira/browse/ROME-46">Issue 46</a></li>
|
|
<li>Fix for the potential problem outlined in <a class="externalLink" href="http://www.securiteam.com/securitynews/6D0100A5PU.html">http://www.securiteam.com/securitynews/6D0100A5PU.html</a>. Thanks to Nelson Minar for bringing this to our attention.</li>
|
|
<li>Fix. Wrong default description type for RSS 2.0 Fix for <a class="externalLink" href="http://java.net/jira/browse/ROME-26">Issue 26</a></li>
|
|
<li>Change default description type for RSS 2.0 from text/plain to text/html as per RSS 2.0 spec</li>
|
|
<li>Fix to add all HTML4 entities, according to <a class="externalLink" href="http://www.w3.org/TR/REC-html40/sgml/entities.html">http://www.w3.org/TR/REC-html40/sgml/entities.html</a> specially for the HTMLsymbol set (Mathematical, Greek and Symbolic characters for HTML) and the HTMLspecial set (Special characters for HTML).</li></ul></li>
|
|
<li>Date parsing fixes
|
|
<ul>
|
|
<li>Additional version and date leniency could extract more information <a class="externalLink" href="http://java.net/jira/browse/ROME-24">Issue 24</a></li>
|
|
<li>Non RFC822 Dates not processed in RSS pubDate field <a class="externalLink" href="http://java.net/jira/browse/ROME-27">Issue 27</a>
|
|
<ul>
|
|
<li>RSS feed parsers were were only parsing RFC822 dates because they were not using the proper date-time parsing function for the date-time elements.</li>
|
|
<li>If a W3C date-time element had no time component it was being parsed as local time instead of GMT, ROME DateParser class has been modified to use GMT in this situation.</li>
|
|
<li>Current JDKs do not handle 'UT' timezone indicator, ROME DateParser class has been modified to handle it.</li>
|
|
<li>Use Atom updated instead of published <a class="externalLink" href="http://java.net/jira/browse/ROME-41">Issue 41</a></li>
|
|
<li>Atom 1.0 Date (Updated or Published) Not Set <a class="externalLink" href="http://java.net/jira/browse/ROME-42">Issue 42</a></li>
|
|
<li>lastBuildDate does not populate publishedDate <a class="externalLink" href="http://java.net/jira/browse/ROME-43">Issue 43</a> Provides a feed date for RSS 0.91 feeds that specify lastBuildDate but not pubDate.</li></ul></li>
|
|
<li>Fix. Parsing some numeric elements was failing due to whitespaces The image.width and image.height of RSS091U, the frequency of SyModule and the cloud.port of RSS092 elements are now being trimmed before doing the integer parsing.</li></ul></li>
|
|
<li>Atom link and URI fixes
|
|
<ul>
|
|
<li>Improper relative link resolution in Atom10Parser <a class="externalLink" href="http://java.net/jira/browse/ROME-37">Issue 37</a></li>
|
|
<li>ATOM 1.0 Entry links parsing <a class="externalLink" href="http://java.net/jira/browse/ROME-38">Issue 38</a></li>
|
|
<li>ConverterForRSS10.java does not set URI for item <a class="externalLink" href="http://java.net/jira/browse/ROME-25">Issue 25</a></li>
|
|
<li>Valid IRI href attributes are stripped for atom:link <a class="externalLink" href="http://java.net/jira/browse/ROME-34">Issue 34</a></li></ul></li>
|
|
<li>Module fixes
|
|
<ul>
|
|
<li>iTunes Module has incorrect author and category support <a class="externalLink" href="http://java.net/jira/browse/ROME-35">Issue 35</a></li>
|
|
<li>mediarss.io.MediaModuleParser NumberFormatException <a class="externalLink" href="http://java.net/jira/browse/ROME-45">Issue 45</a></li>
|
|
<li>Slash module not serializable for FeedFetcher <a class="externalLink" href="http://java.net/jira/browse/ROME-44">Issue 44</a></li></ul></li></ol></div>
|
|
<div class="section">
|
|
<h3>Changes made from v0.7 to v0.8<a name="Changes_made_from_v0.7_to_v0.8"></a></h3>
|
|
<ol style="list-style-type: decimal">
|
|
<li>Change. Added enclosure support at Synd* level<br />A new bean for handling enclosures at Synd* level has been created (SyndEnclosure/SyndEnclosureImpl, interface/implementation).<br />The SyndEntry/SyndEntryImpl bean has a new 'enclosures' property which returns the list of enclosures for that item.<br />The Wire* to Synd* converters for RSS propagate enclosures in both directions.<br />This enables handling enclosures from RSS 0.92, 0.93, 0.94 and 2.0 at Synd* level<br />Test cases have been modified to cover enclosures at Synd* level.</li>
|
|
<li>Change/Fix. Synd* - Atom entry dates mapping
|
|
<ul>
|
|
<li>(Change) Atom entries have 3 dates, 'modified', 'issued' and 'created'. Synd entries have only 1 date property 'publishedDate'. When converting from Atom to Synd the first not null date in the order above will be the one set in the Synd entry bean.</li>
|
|
<li>(Fix) When converting from Synd to Atom the Synd entries 'publishedDate' property value is set in both 'modified' and 'issued' properties of the Atom entry.<br />This Change/Fix is to be aligned with the Atom 0.3 spec.</li></ul></li>
|
|
<li>Fix. Trim enclosure length attribute<br />Fix from Trey Drake: At least 1 podcast site (ESPN) occasionally leaves trailing spaces in the enclosure content length attribute. This causes a NumberFormatException.</li>
|
|
<li>Fix. Conversion to RSS 1.0 if Channel URI is not specified<br />Fix for problem converting to RSS 1.0 if not URI is specified at the channel level (it will now attempt to use the Link element)</li>
|
|
<li>Changes to support Atom 1.0
|
|
<ul>
|
|
<li>In com.sun.syndication.synd, added SyndLink and SyndPerson.</li>
|
|
<li>In SyndEntry added. In SyndEntry, added summary, updatedDate, links collection and support for multiple authors.</li>
|
|
<li>In com.sun.syndication.synd.impl, added Atom10Parser.java, Atom10Generator.java and ConverterForAtom10.java.</li></ul></li></ol></div>
|
|
<div class="section">
|
|
<h3>Changes made from v0.6 to v0.7<a name="Changes_made_from_v0.6_to_v0.7"></a></h3>
|
|
<ol style="list-style-type: decimal">
|
|
<li>Fix. RFC-882 dates parsing and generation were using localized names for day and month names<br />The date parser and generator were using the JVM default Locale instead forcing an English Locale to use day and month names in English as specified by RFC-822. Now US Locale is used.</li>
|
|
<li>Fix. The 'ttl' element of RSS0.94 and RSS2.0 feeds was not being parsed<br />The parsers now parse the 'ttl' element and it is available in the resulting Channel bean. Note that 'ttl' info is not available in the SyndFeed bean, thus it's lost when converting from WireFeed to SyndFeed.</li>
|
|
<li>Change. RSS enclosures with empty 'length' attributes are accepted<br />Parsing an RSS feed with an enclosure where the length attribute was an empty String were failing. Now they are parsed and the length is set to 0.</li>
|
|
<li>Change. RSS 1.0 feeds use URI/Link for unique ID (rdf:about).<br />RSS 1.0 specification recommends that the rdf:about attribute URI use the value of the item's link element, though this could be different if the user chooses to override it by specifying their own URI. RSS 1.0 feeds now use the URI if specified, otherwise the link for the item.</li>
|
|
<li>Fix. toString() was reporting NullPointerException with List properties<br />When a List (or Map) property had a NULL element the toString() logic was failing partially due to a NullPointerException.</li>
|
|
<li>Fix. DC creator elements were being lost when converting to SyndFeed<br />DC creator elements were being lost when converting to SyndFeed. This was happening with RSS versions that have native author elements (0.94 and 2.0) and for the managingEditor element at channel level (available in 0.91 Userland and onwards).</li>
|
|
<li>Change. Date and enumeration elements are trimmed during parsing<br />There are some feeds that add whitespaces or carriage return characters before or after the proper date or enumeration value. This was causing ROME to fail processing those elements. This is taken care now as all dates elements in all feed types and Modules and the 'channel.skipHours.hour' and 'channel.skipDays.day' (RSS0.91 - RSS2.0) are trimmed before parsing and setting their values in the beans.</li>
|
|
<li>Fix. SyndFeed description now maps to atom:tagline<br />Previously, atom:info was being mapped to the feed's description. According to the Atom03 spec atom:info should be ignored by parsers, and the more appropriate element is atom:tagline.</li>
|
|
<li>Fix. RSS cloud is now generated/parsed correctly<br />The 'path' attribute from the cloud was not being generated/parsed. The parser now process all cloud attributes and set the cloud to the channel.</li>
|
|
<li>Fix. RFC-822 2 digit years<br />Previously RFC-822 dates did not work correctly with 2 digit years. This is now fixed.</li>
|
|
<li>Fix. No alternate link causes IndexOutOfBoundsException<br />Fix bug where no alternate link causes IndexOutOfBoundsException in ConverterForAtom03 (Thanks to Joseph Van Valen).</li>
|
|
<li>Change. Date parsing attemps RFC822 on W3C parsing on all feeds<br />All feed parsers (RSS and Atom) now attemp both RFC822 and W3C parsing on date values.</li>
|
|
<li>Fix. XmlFixerReader removes character from stream when parsing an entity that contains an invalid character<br />Fix bug in XmlFixerReader where an invalid entity such as "&ent=", gets put back on the stream without the last character (in this example, "&ent=" becomes "&ent"). This was most visible when the XmlFixerReader encountered an URL with a query string that has more than one parameter (e.g. <a class="externalLink" href="http://www.url.com/index.html?qp1=1&qp2=2">http://www.url.com/index.html?qp1=1&qp2=2</a>) -- all "=" after the first one would disappear.</li>
|
|
<li>Change. DateParser can use additional custom datetime masks<br />Besides attempting to parse datetime values in W3C and RFC822 formats additional datetime masks can be specified in the /rome.properties files using the 'datetime.extra.masks' property. To indicate multiple masks the '|' character must be used, all other characters are considered part of the mask. As with parser/generators/converter plugins the masks are read from all /rome.properties file in the classpath.</li></ol></div>
|
|
<div class="section">
|
|
<h3>Changes made from v0.5 to v0.6<a name="Changes_made_from_v0.5_to_v0.6"></a></h3>
|
|
<ol style="list-style-type: decimal">
|
|
<li>Fix. W3C date-time parsing now handles date-time with 'Z' modifier<br />The W3C date-time parser was not parsing times using the UTC modifier 'Z'.</li>
|
|
<li>Fix. XML prolog encoding parsing was failing when other attributes where present in the prolog<br />If there was an attribute following the encoding attribute the value of the encoding attribute was misinterpreted. For example, for the XML prolog the detected encoding was <tt>UTF-8" standalone="yes</tt> instead of <tt>UTF-8</tt>.</li>
|
|
<li>Change. XmlReader lenient behavior gives priority to XML prolog encoding over content-type charset<br />In ROME 0.5 the XmlReader first attempts to do a strict charset encoding detection. Only if the strict detection fails it attempts a lenient detection. When the HTTP Content-Type header is of type <tt>text/*xml</tt> and the header does not specify any charset, RFC 3023 mandates that the charset encoding must be <tt>US-ASCII</tt>. It's a common error for sites to use the <tt>text/*xml</tt> MIME type without charset information and indicate the charset encoding in the XML prolog of the document, being the charset encoding in the XML prolog different from <tt>US-ASCII</tt>. The XmlReader lenient behavior has been modified to give precedence to the XML prolog charset encoding, if present, over the HTTP Content-Type charset encoding.</li>
|
|
<li>Addition. XML Healer<br />ROME parsers, SyndFeedInput and WireFeedInput, have a new feature, XML healing.<br />The XML healing trims the beginning of the XML text document if there are whitespaces, enters or XML comments before the XML prolog or the root element. It also replaces all HTML literal entities occurrences with coded entities. These changes convert feeds technically invalid (from the XML specification perspective) into valid ones allowing the SAX XML parser to successfully parse the XML if there are not other errors in it.<br />This behavior is active by default. It can be turned on and off using the new 'xmlHealerOn' property in the SyndFeedInput and WireFeedInput classes.<br />The idea for this feature was taken from the FeedFilter from Jakarta's commons feedparser.</li>
|
|
<li>Addition. The XML prolog of generated feeds contains the feed encoding<br />ROME generators were creating feeds with the XML prolog encoding always set to 'UTF-8', if the given Writer had another charset encoding things would break for anybody consuming the feed (a mismatch between the char stream charset and what the XML doc says).<br />The SyndFeedOutput and WireFeedOutput generators now use the SyndFeed and WireFeed 'encoding' property to set the 'encoding' attribute in the XML prolog of the generated feeds. It is the responsibility of the developer to ensure that if the String is written to a character stream the stream charset is the same as the feed encoding property.</li>
|
|
<li>Change. SyndFeed to Atom convertion now uses 'escaped' mode for content elements<br />SyndFeed to Atom converter was using 'xml' mode for content elements. This was breaking feeds with content that was not propertly escaped as it was assumed to be XML fragments.</li>
|
|
<li>Change. RSS 2.0 parser and generator now handles DC Module<br />ROME configuration has been changed so RSS 2.0 parser and generator handle DC Module elements at channel level and item level.</li>
|
|
<li>Fix. RSS0.93, RSS0.94 and RSS2.0 'dc:date' element value was being lost under certain conditions<br />If a feed had 'dc:date' elements but not 'pubDate' elements, the 'dc:date' elements where lost when converting from Channel to SyndFeed.<br />This was happening for 'dc:date' elements at channel level and at item level.</li>
|
|
<li>Fix. RSS 1.0 'rdf:resource' and 'rdf:about' item linking attributes use a unique ID now<br />The value for the 'rdf:resource' and 'rdf:about' linking attributes was done using the value of the 'link' element. If there is more than one item with the same link the generated feed will be incorrect.<br />Instead using the link value now the index of the item is used for the linkage between 'rdf:resource' and 'rdf:about' for items.</li>
|
|
<li>Fix/Change. Parsing and setting of enumerated elements is case insentive now<br />Parsing and bean setting of enumerated values (such as RSS skipDay, Atom content mode, etc) are now case insentive, generation is strict (Postel Law).</li>
|
|
<li>Fix. Remove enumeration check on 'rel' attribute of Atom link elements<br />Because a misunderstanding of Atom 0.3 specification the Atom Link bean was checking the value of the 'rel' property against a set of valid values. The check has been removed.</li>
|
|
<li>Fix. DC subjects (in RSS versions with native categories) were being lost on conversion to SyndFeed<br />All RSS versions with native categories (at both channel and item level) now have the following behavior when converting to SyndFeed.<br />DC subjects are converted to SyndCategories. Native categories are converted to SyndCategories. They are both aggregated in a Set (to remove duplicates) then added to the SyndFeed.<br />When doing a SyndFeed to Channel conversion, if the RSS version has native categories and handles DC modules the categories will be duplicated as native and DC ones.</li>
|
|
<li>Fix/Change. RSS 1.0 rdf:about attribute in the channel.<br />RSS 1.0 uses the rdf:about attribute at the channel level as an identifier. This was not being parsed or generated (only supported at the item level). Support for this was added along with test cases.</li>
|
|
<li>Fix/Change/Addition. Multivalued Dublin Core element support.<br />Many feeds are using multiple DC elements to tag metadata, the interface for the DCModule has been changed to support Lists of elements, compatibility has been maintained with the existing interface though as the new methods are plural (creators vs. creator), the single value methods will remain as convenience methods. The implementation now uses the lists to represent all of the elements. The parser/generator modules for DC have been updated to reflect these changes along with a few other code cleanups in the DC* modules.</li>
|
|
<li>Fix. Removed length constraint checks from RSS1.0 generator<br />RSS1.0 specification does not require, only suggests, maximum length for some of the elements. ROME was enforcing those lenghts when generating RSS1.0 feeds. This enforcement has been removed becuase is not mandatory.</li></ol></div>
|
|
<div class="section">
|
|
<h3>Changes made from v0.4 to v0.5<a name="Changes_made_from_v0.4_to_v0.5"></a></h3>
|
|
<ol style="list-style-type: decimal">
|
|
<li>Change. Got rid of Enum class<br />All constants in the beans are Strings now, the corresponding property setters check that the value being set is one of the valid constants. Rome has not business defining an Enum class.</li>
|
|
<li>Change. Got rid of ToString interface<br />This is just an implementation convenience, it was polluting Rome API. Modified ToStringBean to work without requiring an interface and method to propagate the prefix to use with properties.</li>
|
|
<li>Change. ObjectBean, ToStringBean, EqualsBean are not part of the public API anymore<br />These are just an implementation convenience, they were polluting Rome API. Rome bean implementations don't extends ObjectBean anymore. Instead they use it in a delegation pattern. While these classes are not public anymore they are part of Rome implementation.</li>
|
|
<li>Change. CopyFrom interface moved to com.sun.syndication.feed package<br />The common package is gone now that the *Bean classes are not there anymore. No point keeping a package just for an interface.</li>
|
|
<li>Fix. PluginManager was not doing plugin lookup in the defined order<br />PluginManager (manages parsers, generators and convertors for feeds and modules) was not doing the lookup in order the plugins are defined in the rome.properties files. This is needed for parsers where the lookup involves detecting the feed type and the feed type detection goes needs to go from stronger to weaker.</li>
|
|
<li>Addition. Rome now recognizes RSS 2.0 feeds with 'http://backend.userland.com/rss2' namespace<br />These namespace was defined by an RSS 2.0 draft and later was dropped. There are feeds out there using this namespace and Rome was not parsing them.</li>
|
|
<li>Change. By default XmlReader does a lenient charset encoding detection.<br />If the charset encoding cannot be determined per HTTP MIME and XML specifications the following relaxed detection is performed: If the content type is 'text/html' it replaces it with 'text/xml' and tries the per specifications detection again. Else if the XML prolog had a charset encoding that encoding is used. Else if the content type had a charset encoding that encoding is used. Else 'UTF-8' is used.<br />There are 2 new constructors that take a lenient flag to allow strict charset encoding detection. If scrict charset encoding detection is performed and it fails an XmlReaderException is thrown by the constructor. The XmlReaderException contains all the charset encoding information gathered from the stream including the unconsumed stream.</li></ol></div>
|
|
<div class="section">
|
|
<h3>Changes made from v0.3 to v0.4<a name="Changes_made_from_v0.3_to_v0.4"></a></h3>
|
|
<ol style="list-style-type: decimal">
|
|
<li>Fix. Date elements on generated feeds use the right format now<br />There were some Date elements in Atom 0.3, DCModule, SyModule, RSS 0.91 and RSS 0.93 that were incorrectly formating the date [they were just doing a toString() ]. Now they use the RFC822 and W3C format as indicated in their specs.</li>
|
|
<li>Fix. SyndFeed and SyndEntry getModule(DCModule.URI) method always returns a DCModule, never null<br />This issue is related to Fix #19 in v0.3. The DCModule is 'special' for Synd*, it must always be there. If it is not, it is created implicitly when needed. This was not happening when asking explicitly for the DCModule through the Synd* interfaces.</li>
|
|
<li>Addition. Added ParseFeedException to the *Input classes<br />This new exception report the line and column number in the XML document where the parsing has failed.</li>
|
|
<li>Change. Renamed '*I' interfaces to just '*' and default implementations to '*Impl'<br />The Synd* and Module interfaces/classes were affected. For example the interface that used to be SyndFeedI is now SyndFeed and the class that used to be SyndFeed is now SyndFeedImpl.</li>
|
|
<li>Change. Ant 'build.xml' files have been improved<br />The build.xml were re-written instead just using the Maven generated ones.</li>
|
|
<li>Fix. DateParser now uses lenient parsing so as to work with JDK 1.5<br />DateParser currently setLenient to false. This does not work with JDK 1.5. Changing the DateParser to setLient to true.</li>
|
|
<li>Change. SyndCategoryListFacade is not a public class anymore<br />It's now a package private class (it should have been like that in the first place).</li>
|
|
<li>Addition. Added a protected constructor to the Synd*Impl classes<br />This constructor takes a Class parameter. It allows implementations extending SyndContentImpl to be able to use the ObjectBean functionality with extended interfaces (additional public properties). Use case: Hibernate beans that have an 'Id' Long property, a new interface HSynd* (extending Synd*) and a new implementation HSynd*Impl (extending Synd*Impl) where the clone(), equals(), hashCode() and toString() methods take the properties of the extension into account.</li>
|
|
<li>Project layout change. Moved samples project to subprojects dir<br />The 'samples' project was moved from 'rome/modules/samples' to 'rome/subprojects/samples'. The 'rome/modules' project is left for Module subprojects only.</li>
|
|
<li>Fix. Plugin manager bug didn't allow custom plugins to replace core plugins<br />All plugins are in a Map keyed off by its type (parsers, generators, converters) or URI (modules). There is also a helper List containing the plugins, this list is scanned when looking for a plugin (for example when selecting the right parser). Plugins were added to the List without checking if another element in the List was using the same key. Now the List is built using the Map values, thus the overwriting works fine.</li>
|
|
<li>Fix. RSS 2.0 Converter (Wire -> Synd - Wire) was not processing modules<br />The RSS 2.0 Converter now processes feed and entry modules both ways.</li>
|
|
<li>Fix. New properties introspection mechanism in common classes<br />The java.beans.Introspector does not work as expected on interface properties (it doesn't scan properties of the super interfaces). Now, a private implementation of a properties introspector is used by the common bean classes.</li>
|
|
<li>Change. Refactored private CopyFrom helper class<br />It was com.sun.syndication.feed.synd.impl.SyndCopyFrom now it is com.sun.syndication.common.impl.CopyFromHelper.</li>
|
|
<li>Fix. RSS2.0 Wire-Synd converter handles propertly RSS and DC categories<br />If the RSS2.0 feed had DC Module categories (subjects) this would override the RSS native categories. Now they are aggregated.</li>
|
|
<li>Change/Fix. CloneableBean can take an ignore-properties set<br />This change is useful for bean that have convenience properties mapping to other properties. SyndFeed and SyndEntry that map some of its properties (ie publishedDate, author, categories) to DC Module properties. There is not point cloning the convenience ones as they are just facades.<br />This fixes a SyndFeedImpl cloning problem with the categories. There is a package private list for the categories to DC subject mapping. The problem was related to accesibility of this package private list implementation by the CloneableBean. The change enables the blacklisting of certain properties (in this categories) when cloning.</li>
|
|
<li>Change. Refactored Parsers/Generator classes<br />Introduced a base parser and base generator to handle modules. For the feed types that define modules support, the modules have to be defined the rome.properties file. For RSS 1.0 and Atom 0.3 Dublin Core and Syndication modules are defined, the first one at feed and entry level the second one at feed level. Note this was already done but hardwired in the specific feed type parsers and generators, now it is done in the base parser and base generator. Some code clean up and removal of duplicated code was also done.</li>
|
|
<li>Change. Defined and implemented precedence order for native and module elements in feeds<br />For feeds supporting modules, some of the module defines elements collide with the feed native elements (this depends on the feed type, and it may affect the data such as publish-date, author, copyright, categories). The SyndFeed and SyndEntry properties documented as convenience properties are (can be) affected. The convenience properties map into DC Module properties.<br />Rome now defines precedence of feed native elements over module elements when converting from WireFeed to SyndFeed. This is, in the situation of a clash, the native element data prevails and the the module element data is lost.<br />When converting from SyndFeed to WireFeed, if a SyndFeed convenience property has a native mapping in the target feed type it will be in both the native element and the DC Module element if the feed type defines support for the DC module. The data will appear twice in the feed, in the native elements and in the DC module elements.</li>
|
|
<li>Change. Module namespaces are defined at root element<br />Module namespaces are always defined in the root element. The ModuleGenerator interface has a new method that returns all the namespaces used by the module, the generators use the namespaces returned by this method to inject them in the feed root element.</li>
|
|
<li>Change/Fix. Test cases refactoring<br />Some refactoring in the test cases for the Synd* entities. Some minor bugs (typos in constants) were found and fixed in the parsers.<br />The Ops and Synd tests are 100% complete. Not that they test 100% of Rome functionality but 100% of what they suppose to test.</li>
|
|
<li>Fix. Atom 0.3 content elements with XML mode were not being parsed and converted properly<br />When mode is XML, free text was not being processed, only sub-elements were being processed.<br />The Parser when parsing content with XML mode skips Atom namespace in the output of the processed fragment.<br />The Converter when going down pushes content data using XML mode, which is Atom's default (it was ESCAPED before).</li>
|
|
<li>Addition. Constraints in data length and number of items are observed by RSS generators<br />As part of the generators refactoring the generators now verify the data in the Rome beans does not generate invalid feeds.</li>
|
|
<li>Addition. New XmlReader that detects charset encoding<br />The XmlReader class handles the charset encoding of XML documents in Files, raw streams and HTTP streams. It following the rules defined by HTTP, MIME types and XML specifications. All this is nicely explained by Mark Pilgrim in his blog, <a class="externalLink" href="http://diveintomark.org/archives/2004/02/13/xml-media-types">http://diveintomark.org/archives/2004/02/13/xml-media-types</a></li>
|
|
<li>Rome now uses JDOM 1.0<br />Dependencies have been updated to use JDOM 1.0. No code changes were needed because of this.</li>
|
|
<li>Change. The length property in the RSS Enclosure bean is now a long<br />It used to be an int, but as it is meant to specify arbitrary lengths (in could be in the order of several megabytes) it was changed to be a long.</li>
|
|
<li>Addition. SyndFeed and SyndEntry beans have a 'uri' property<br />It is used with RSS0.94 & RSS2.0 guid elements and with Atom0.3 id elements. Refer to <a href="./RomeV0.4FeedAndEntryURIMapping.html">Feed and entry URI Mapping</a> for a full explanation.</li>
|
|
<li>Addition. New sample, FeedServlet<br />Added a new sample, FeedServlet. This servlet creates a feed and returns a feed.</li>
|
|
<li>Change. RSS0.91 Userland and RSS0.91 Netscape are handled as different feed types<br />Instead having a single set of parser/converter/generator implementations there is one set for each one of them. This allows to differenciate incoming RSS0.91 Userland feeds from RSS0.91 Netscape feeds as well as choosing which one to generate.</li></ol></div>
|
|
<div class="section">
|
|
<h3>Changes made from v0.2 to v0.3<a name="Changes_made_from_v0.2_to_v0.3"></a></h3>
|
|
<ol style="list-style-type: decimal">
|
|
<li>Changed loading mechanism for parsers, generators and converters<br />Previous mechanism was complicated and it wouldn't work in server environments where you cannot alter System properties at will.<br />Now there are no properties to set. Just including a rome.properties file in the root of a JAR file will make Rome to load any parser, generator or converter defined in the properties file and included in the JAR file (To be documented).</li>
|
|
<li>Added Modules support to RSS 2.0 parser and generator<br />We were only looking for modules in RSS 1.0 and Atom 0.3. Now we also look in RSS 2.0.</li>
|
|
<li>Modules parsers and generators are now per feed type<br />They were global, now each feed type parser and generator has it's own set of modules parsers and generators. They are configurable in the rome.properties file.</li>
|
|
<li>All parsers, generators and converters are loaded once per definition<br />There were some cases we were loading them on every WireFeedParser/WireFeedGenerator instantiation. As they are all thread-safe we just use one instance.</li>
|
|
<li>Changed some implementation (hidden) class names<br />Some typos corrections and adding consistency to the naming.</li>
|
|
<li>Added Fetcher module<br />Added a HTTP fetcher with conditional get and gzip support (See modules/fetcher)</li>
|
|
<li>Modified the samples module build to fix error<br />Previouly some people were having a maven error trying to run some of the samples. This should now be fixed</li>
|
|
<li>Removed methods with byte stream signatures from Rome IO classes<br />They have no way to know what the encoding is or has to be and they are using the platform default. That does not always work, for example when doing HTTP where the default is ISO-8859-1.<br />Leaving the char streams one and let the developer to do the right thing (Rome cannot do any magic here).</li>
|
|
<li>Added getModule(String uri) to SyndFeedI, Atom Feed and RSS Channel<br />There was not way to obtain a specific module using the module URI, you had to obtain the module list and iterate looking for the module. The getModule(string uri) is a convenience method in all the feed beans to do that.</li>
|
|
<li>New. Added 'encoding' property to WireFeed and SyndFeedI/SyndFeed<br />Impact: It affects RSS, Atom and SyndFeed syndication beans.<br />It's not being set or use by parsers and generators as they always deal with a char strean where the charset is already set.<br />The converters, going from Channel/Feed 2 SyndFeed and vice versa are wired to pass the encoding if set.</li>
|
|
<li>Fix. CloneableBean array cloning bug<br />Impact: It affects all Rome beans as they all extend ObjectBean that uses CloneableBean<br />Arrays were not being cloned but copy by reference. This was affecting all Rome beans as they all extend ObjectBean.</li>
|
|
<li>Change. CloneableBean, Basic (primitives & string) types not cloned anymore<br />Impact: It affects all Rome beans as they all extend ObjectBean that uses CloneableBean<br />Basic types are inmutable, no need to clone them. As things are done using reflection there were unnecessary objects creation.</li>
|
|
<li>Change. CloneableBean, added Date to list of Basic types<br />Impact: It affects all Rome beans as they all extend ObjectBean that uses CloneableBean<br />Same reasoning as #12</li>
|
|
<li>Change/Fix. EqualsBeans, works on defined class<br />Impact: It affects all Rome beans as they all extend ObjectBean, which uses EqualsBean<br />EqualsBean checks for equality by comparing all properties. Until now it was doing this using all properties of the class implementing the bean. This behavior is not correct as implementations may have other properties than the defined in the public interface (and example is: a SyndFeedI implementation for Hibernate that has to have an ID field, and this would apply to all bean interfaces).<br />Because of this change, both EqualsBean and ObjectBean take a Class parameter in the constructor. Equals will limit the comparison for equality to the properties of that class. If there is not interface for the bean, just the bean implementation (this is the case of Channel and Feed), the implementation Class is passed</li>
|
|
<li>Change/Fix. ToStringBean, works on defined class<br />Same as #14 but on ToStringBean instead EqualsBean</li>
|
|
<li>New. Added copyFrom functionality to synd.* and module.* beans<br />Impact: It affect all synd.* and module.* beans and other classes that extend ObjectBean.<br />A new interface 'CopyFrom' has been added to commons.<br />synd.* and module.* bean interfaces implement this interface and their implementations implement the copyFrom functionality.<br />The copyFrom functionality allows copying a complete bean feed from one implementation to another. The obvious use case is from the default bean implementation to a persistent aware (ie Hibernate) implementation.<br />The implementation uses the same pattern used by EqualsBean, ConeableBean, it's in the synd.impl.* package, the supporting class is call SyndCopyFrom.<br />Note that SyndCategories does not support the copyFrom functionality as it's just a convenience way of accessing DCModule's subjects (where DCModule supports copyFrom). The short explanation is that categories still work and are there after a copyFrom.</li>
|
|
<li>Fix. WireFeed constructor was passing the WireFeed.class to ObjectBean<br />rss.Channel, atom.Feed & WeatherChannel constructor implementations only, no change to their signatures.<br />The class is used for property scanning for toString, equals behavior of the ObjectBean. </li>
|
|
<li>Fix. copyFrom was failing with Enum properties<br />It was not possible to do a copyFrom on a feed with SyModule data as it had Enum properties.<br />Wrong comparison of classes when checking basic types, in the case of enumeration classes has to check it extends Enum not equality.</li>
|
|
<li>Fix. SyndFeed and SyndEntry where losing Modules<br />If they had modules other than DCModule they were being dropped when accessing convenience methods (getCategories, getLanguage, etc).<br />The Synd* convenience methods just map to the DCModule properties. If there is no DCModule a new one is created if needed. When this happened instead bean added to the list of current modules a new list with just the DCModule was being set.</li>
|
|
<li>New Sample and tutorial for creating a feed<br />A new sample and tutorial that creates and writes a feed has been added.</li>
|
|
<li>New Sample showing how to add a Module bean, parser and generator to Rome<br />This sample defines a dummy module for use with RSS 1.0 (it can work also with RSS 2.0 and Atom 0.3). Still have to write tutorial</li>
|
|
<li>Undoing change #13 as Date is not inmutable<br />The Date class is not inmutable, CloneableBean now clones Date properties.</li></ol></div>
|
|
<div class="section">
|
|
<h3>Changes made from v0.1 to v0.2<a name="Changes_made_from_v0.1_to_v0.2"></a></h3>
|
|
<ol style="list-style-type: decimal">
|
|
<li>FeedInput, added default constructor. Semantics is 'validation off'<br />We forgot to added it when it was added to SyndFeed.<br />NOTE that validation is not implemented yet. We need DTDs/XML-Schemas for the different feed syndication types.</li>
|
|
<li>FeedOutput and SyndOutput outputW3CDom() method typo correction<br />Fixed typo in outputW3CDom() method, it was ouptutW3CDom().</li>
|
|
<li>AbstractFeed, renamed to WireFeed<br />Never liked Abstract in the name. From Java inheritance it makes sense, but from the syndication feed perspective doesn't. It's the super class of the 2 wire feed beans Rome has, RSS Channel and Atom Feed.</li>
|
|
<li>Renamed SyndFeed createRealFeed(feedType) method to createWireFeed(feedType)<br />For consistency with change #3</li>
|
|
<li>SyndFeed, added feedType property<br />Read/write property. It indicates what was the feed type of the WireFeed the SyndFeed was created from. And the feed type a WireFeed created with createWireFeed() will have.</li>
|
|
<li>WireFeed, renamed type property to feedType<br />For consistency with #5. Also it's more clear what the type is about.</li>
|
|
<li>Overloaded SyndFeed createRealFeed() with a no parameter signature</li>
|
|
<li>FeedOutput, removed feed type from constructor<br />It now uses the WireFeed feedType property.</li>
|
|
<li>SyndOutput, removed feed type from constructor<br />It now uses the SyndFeed feedType property define in #5.</li>
|
|
<li>FeedOutput, removed getType() method<br />Now FeedOutput uses the WireFeed feedType property.</li>
|
|
<li>Removed dependency on Jakarta Commons Codec library<br />We were using the Codec component to do Base64 encoding/decoding. Based on feedback to reduce component depencies we've removed this one (yes, we implemented a Base64 encoder/decoder).</li>
|
|
<li>Removed dependency on Xerces library<br />This was an unnecessary dependency as Rome requeries JDK 1.4+ which includes JAXP implementation. JDOM can use that one.</li>
|
|
<li>Renamed syndication.io classes/interfaces<br />Renaming for naming consistency and to reflect on what type of feed they work on.
|
|
<div class="source">
|
|
<pre>
|
|
FeedInput --> WireFeedInput
|
|
FeedOutput --> WireFeedOutput
|
|
FeedParser --> WireFeedParser
|
|
FeedGenerator --> WireFeedGenerator
|
|
SyndInput --> SyndFeedInput
|
|
SyndOutput --> SyndFeedOutputt
|
|
</pre></div></li>
|
|
<li>Removed syndication.util package, PlugableClasses is now private<br />The PlugableClasses class has no business in Rome public API, it's implementation specific, it has been hidden (it should be replaced later with a micro-container).</li>
|
|
<li>Added samples to the Rome project directory structure<br />Rome samples are a sub-project located at rome/modules/sample.</li></ol></div></div>
|
|
</div>
|
|
</div>
|
|
<div class="clear">
|
|
<hr/>
|
|
</div>
|
|
<div id="footer">
|
|
<div class="xright">
|
|
Copyright © 2004-2013
|
|
<a href="http://www.rometools.org">ROME Project</a>.
|
|
All Rights Reserved.
|
|
|
|
</div>
|
|
<div class="clear">
|
|
<hr/>
|
|
</div>
|
|
</div>
|
|
</body>
|
|
</html> |