Prior to first release (on the way to v0.3)

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
   #. New FeedFetcher Implementation
   HttpClientFeedFetcher uses the Apache Commons HTTP Client
2. Abstract test classes excluded in project.xml
   Tests now run correctly under Maven
3. 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
2. Support for rfc3229 delta encoding
   The Fetcher now supports rfc3229 delta encoding. See http://www.ietf.org/rfc/rfc3229.txt and http://bobwyman.pubsub.com/main/2004/09/using_rfc3229_w.html. Note that this is support is experimental and disabled by default

Prior to 0.6

1. Feed passed to FetcherEvents
   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.rometools.rome.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
3. Support for clearing cache
   See http://java.net/jira/browse/ROME-119 for details

Prior to 1.0

1. Support for preserving wire feed data.
   The fetcher now has a setPreserveWireFeed() method which will setup ROME to preserve WireFeed data. See PreservingWireFeeds for further information.