Apache Abdera: Changes in 1.1.3 Release --------------------------------------- ABDERA-290: Apply patch from Andreas Veithen and Olve Sæther Hansen to fix using Axiom 1.2.12 breaks Abdera-parser - compile-errors r1137038: further ISO 639-2 language codes added r1136911: refactoring;added some new ISO 639-2 language codes r1136868: added some new ISO 639-2 language codes ABDERA-285: deps.properties javamail.dir is not valid, updated directory in description r1136817: harmonized geronimo-activation version r1132940: removed broken test ABDERA-281: Add axiom-api dependency to parser ABDERA-282: Maven build problem r1065346: added quaility factors to request accept header ABDERA-259: NullPointerException happens when trying to decode HEAD response r1060056: finally removed outdated couchdb adapter ABDERA-279: patch from Antoine Toulme to fix Faulty classloading - unreachable code r1058008: updated readme and build script r1057926: moved doco to wiki r1210094 ABDERA-284: Updated the Axiom version in the ant build (to match the one used in the Maven build). r1210126 Only add default namespace declarations on "service" elements that are created programmatically, but not on elments that are created by the parser. Otherwise this may result in inconsistent namespaces. This change also makes Abdera 1.x compatible with Axiom 1.2.13 (which detects this situation and triggers an exception). r1371874 Let the root POM inherit from the Apache super-POM. r1371875 Prevent the build from creating a derby.log file in the source tree (instead, create it in the target directory). r1371876 Fixed svn:ignore. r1371882 Avoid usage of deprecated UUIDGenerator. r1371885 Removed unused attribute. r1371887 Removed unnecessary null checks. r1371956 Removed method that duplicates code from the base class. r1371962 Realigned the control flow in FOMBuilder with the control flow in the standard Axiom builders. r1371984 Removed unnecessary repository declaration. The Apache snapshot repository is configured in the Apache super-POM. r1372029 Don't let the build depend on external resources (other than artifacts from Maven repositories). Added the test cases from feedvalidator.org to the sources. They are MIT licensed, which is compatible with the Apache license. r1372057 Added a unit test to test various parser options for which there was no test coverage yet. r1372155 Started to decouple Abdera from the internals of the Axiom builder by moving the filtering logic (for whitespace, comments, etc.) out of FOMBuilder into an XMLStreamReader wrapper. r1372161 No need to replace the entire StAXOMBuilder#next() method. At this point we only need to override the method to transform exceptions. r1372165 A processing instruction is always complete. r1372174 Use meaningful parameter names. r1372497 Since we filter out DTDs, we don't need to handle them in FOMDocument#clone(). r1374550 Fixed the qnameAliasMap stuff. It was not exactly working as expected because Axiom resets the namespace information after an element has been created. Moved the logic to FOMStAXFilter so that QName aliasing is entirely transparent to Axiom. Also added a test case. r1374699 Forgot to give the unit test a meaningful name. r1374700 Added a unit test for attribute filtering. r1374702 Also move attribute filtering to FOMStAXFilter, so that FOMBuilder no longer needs to override StAXOMBuilder#processAttributes. Now, all filtering and transformation is done in FOMStAXFilter. r1374707 Removed constructors that would allow to create OMElement instances that are not linked to an OMFactory. r1374724 Some code simplifications. r1374732 When creating an element from the builder, only initialize the local name. The namespace will be set by StAXOMBuilder#processNamespaceData. This should avoid unnecessary creation of OMNamespace objects. r1374760 Avoid calling OMElement#removeAttribute(OMAttribute) with a null value. r1374766 Eliminated a couple of unnecessary/incorrect calls to setParentElement. r1374815 Upgraded Axiom to 1.2.14-SNAPSHOT. r1374816 Removed empty svn:mergeinfo properties. r1407291 added patch from Willem Jiang for ABDERA-293 r1407379 fix for ABDERA-309 r1409289 Set svn:ignore. r1409331 Adapt to latest changes in Axiom. r1412329 fix for the Java7 build problems from ABDERA-311 r1422616 Switched to Axiom release version. Apache Abdera: Changes in 1.1.2 Release --------------------------------------- ABDERA-278: Fix to relax the version requirement for commons-codec in OSGi manifests ABDERA-277: Apply patch from Olve Sæther Hansen to fix Excessive calls to flush() on the Writer responsible for writing JSONStream data. added unit test for ABDERA-251 and and ABDERA-243 applied patch for ABDERA-251 Charset issue in FOMDiv.getInternalValue() leads to corrupt return value on non-ASCII platforms removed couchdb adapter from build files to get rid of the additional maven repo; updated some dependencies fix for ABDERA-275 Error in Turkish Locale when using GZIP compression fix for ABDERA-258 Abdera needs to support IPv6 fixed null pointers; added unit tests - preparing a refactoring Apache Abdera: Changes in 1.1.1 Release --------------------------------------- r1039944: ABDERA-262: patch from Nathan Reese to fix EntryRequestProcessor and MediaRequestProcessor do not support POST" r1039664: ABDERA-270: Apply patch from Guillaume Nodet for OSGi meta-data is broken r1039210: Update to use the mow released 1.2.10 version of Axiom instead of the snapshot r1006114: ABDERA-267: Apply patch from Andreas Veithen to use latest axiom to get fix for performance issue r956475: Add an explicit cast to fix a compile error when building with JDK 1.5 r955891: Add some notes on releasing the artifacts r955868: Update changes file with latest fix from Josh Gordineer for URI templates Apache Abdera: Changes in 1.1 Release ------------------------------------- r955502: Fix negative behaviour on uri templates r941259: Add a changes file listing the changes since 1.0 r944143: Remove old KEYS files, the actual one thats used is at: http://www.apache.org/dist/abdera/KEYS r944731: minor refactoring of unit tests; adopt test structure to src structure r944740: removing all occurences of 'extend Testcase' r944745 and others: replaced all occurences of 'extends Assert' with static imports r944780: update to junit 4.8.1 r944792: easymock version update r944803: fix for ABDERA-102 - Unit tests equality assertions, params in wrong order r944810: removed unnecessary try/catch from Unit Test r944813 and others: fix for ABDERA-102 r944883: updated jetty version r945376: fix for ABDERA-242 - Failing to convert some big capital letters from HTML escape characters to UTF-8 r946755: added FOMParserOptionsTest; fixed OElig r946759: added license header r946769: added Testcase for HTML special characters r947102: minor version updates r947450: extended and refactored unit tests in core r948204: patch for ABDERA-250 (the original patch couldn't be applied as there was a refactoring in the meantime) r951666: Reformat codebase using the template at https://svn.apache.org/repos/asf/abdera/etc/ r940380: ABDERA-102: Apply patch from Christine Koppelt for Unit tests equality assertions, params in wrong order r940218: Fix GeoRSS extension from Nathan Reese on ML - http://www.mail-archive.com/dev@abdera.apache.org/msg00259.html r940215: ABDERA-233: initBaseUri() in the class ServletRequestContext will append the port 443 to https requests, fixed as suggested in JIRA by J G r940195: Update jaxen license text r936036: Allowing to use a custom XMLReader to parse documents. Thanks to Dan Beaulieu. r936030: reverting ABDERA-210, since had other side effects. Thanks to Ck r936027: fixing a bug initializing providers r916287: fixes redundant assignation of mimeType. Thanks to Pascal Christoph r914998: fixing providers discovering errors. Now we just need to fix the failing tests r782732: Don't use the discovery mechanism for cache impls. for now, this means the AbderaClient is hard coded to use the LRUCache in-memory cache implementation. The original design meant for the cache implementation to be pluggable. Ideally, however, the HttpClient would incorporate some kind of cache so we wouldn't have to do this at all. I'm wondering if there's some way we could separate all the http client related stuff out to a distinct subproject? r782570: Applied patch from ABDERA-236 - Create OSGi Bundles for Abdera using the Maven Felix plugin r780662: LRUCacheFactory needs a constructor that takes an Abdera parameter r780661: Added dependency on commons-codec for i18n build r774027: avoid potential NPE r773610: Fix compiler error r773298: remove unnecessary compiler warning r773297: Fix for https://issues.apache.org/jira/browse/ABDERA-231 r772446: Refactoring of some of the base level class discovery code. The moves things to a more logical location, improves the reliability of the implementation and simplifies the code structure a bit. r772439: Fix NPE waiting to happen r772417: Refactoring of some of the base level class discovery code. The moves things to a more logical location, improves the reliability of the implementation and simplifies the code structure a bit. r772409: remove unnecessary compiler warnings r772343: Some refactoring. the EncodingUtil class isn't strictly required, was mostly deprecated... moved the still relevant code to other areas and removed the EncodingUtil class. r772324: Fix for https://issues.apache.org/jira/browse/ABDERA-225 r772316: Fixing the ant build r771562: Ignore eclipse artifacts r771487: Fix for https://issues.apache.org/jira/browse/ABDERA-222 The PeekAheadInputStream was writing too much back to the buffer, which introduces 0x0 chars to the stream r771450: For https://issues.apache.org/jira/browse/ABDERA-210 ResponseContext needs to be able to support non-Atom binary resources so we cannot assume that AbderaServlet should always use a Writer. This adds a new method to ResponseContext called isBinary, which defaults to false. If true, the AbderaServlet will pass the OutputStream to ResponseContext, if false, the servlet will pass the Writer to ResponseContext. This also fixes a bug in ResponseContextWrapper. r771446: One fix for two issues: ABDERA-213, ABDERA-211: No way to explicitly declare namespaces using StreamWriter interface StreamWriter interface not outputting correct namespace for atom:link element in service docs r771444: IRI.toString() does not preserve empty fragment identifiers and queries Per https://issues.apache.org/jira/browse/ABDERA-218 r771442: Creating AbderaClient using preconfigured HttpClient r749155: Remove mis-spelled element. r737883: htmlParser notice added to notice-supplement