1. Updated to handle removal of IO methods using byte streams
Byte Stream IO was removed from Rome itself. The Rome Fetcher is now updated to support this
2. Add FeedFetcherI interface and FeedFetcherFactory class
There is now a FeedFetcherI interface, which FeedFetcher implements. Use FeedFetcherFactory to create instances of FeedFetcher (as suggested by Joseph Ottinger) (FeedFetcherFactory was later removed)
3. Event Support Added to FeedFetcherI
The FeedFetcherI interface now supports feed polled, feed retrieved and feed unchanged events
4. Samples added
Samples are now included with the Rome Fetcher
5. Unit Tests Added
JUnit based tests which invoke the Rome Fetcher against an embedded Jetty webserver are now included
6. Bug fixes in the FeedFetcher event model
The JUnit test suite uncovered some bugs in the event model used by the FeedFetcher. These bugs are now fixed.
7. Refactored the SyndFeedInfo class
SyndFeedInfo now extends ObjectBean
8. Removed FeedFetcherFactory
The benefit of the FeedFetcherFactory was arguable. Now the client code will need to manage the creation of specific implementations of the FeedFetcher
Prior to second release (on the way to v0.4)
1. Refectored to match Rome naming standards
FeedFetcherI renamed to FeedFetcher
2. New FeedFetcher Implementation
HttpClientFeedFetcher uses the Apache Commons HTTP Client
3. Abstract test classes excluded in project.xml
Tests now run correctly under Maven
4. Added GZip support to HttpClientFeedFetcher
HttpClientFeedFetcher now supports GZip compression. Tests have been added.
Prior to third release (on the way to v0.5)
1. SyndFeedInfo implements Serializable
SyndFeedInfo implements Serializable to make it easier to store
When a feed is retrieved it is now passed to the Fetcher Event. This makes it easier to code applications using an event oriented style.
Prior to 0.7
1. Fix for URL Connection leak
In some circumstances URLConnection objects were not closed. This could cause problems in long-running application.
0.8 was never released
Prior to 0.9
1. Fix for potential synchronization issue
There was the possibility of synchronization issues in the FeedFetcher. Fixed, thanks to suggestions from Javier Kohen.
2. New LinkedHashMapFeedInfoCache FeedFetcherCache implementation
The new LinkedHashMapFeedInfoCache has the advantage that it will not grow unbound
Prior to 1.0RC2
1. BeanInfo class added for AbstractFeedFetcher
com.sun.syndication.fetcher.impl.AbstractFeedFetcherBeanInfo was created to allow introspection to correctly find the events
2. Callback to allow access to HttpClient HttpMethod object
Add a HttpClientMethodCallbackIntf to allow the calling code to modify the HttpClient HttpMethod used to make the request (eg, add additinal headers, etc.) Also fixes a reported bug where the user agent wasn't being set properly