2011-12-13

gPodder 3.0.2 "Agent 99" released

A new release of gPodder is out, bringing some refinements and bug fixes to your podcatching experience:
  • Updated Mac OS X icon for gPodder 3
  • Updated Portuguese translation
  • Remove downloading of favicon as cover art
  • Custom feeds: Delete episodes when removed from feed
  • Improvements to the fake D-Bus module for systems without D-Bus
  • Write the default configuration on startup (bug 1511)
  • YouTube: Fix pubDate via monkey-patching (bug 1492)
  • Common prefix elimination for episodes
  • Gtk UI: Avoid errors when querying tooltips of separators
  • QML UI: Scrubbing (in the media player) and other tweaks
  • Updated translation templates from source
  • Updated German translation
  • QML UI: Add subscribe info for empty podcast list
Detailed ChangeLog: gpodder.org/changelog/3.0.2.

Download: gpodder-3.0.2.tar.gz
md5sum: b87755c75c6bcf754b73aa5ad028d514
sha1sum: ba547e5472466fe1b18afb72c1e2a5c25561b000

For packages in distributions, see the downloads page.

Please note that the Desktop version of gPodder 3 does not contain the MP3 player syncing feature. See this blog post for more information and add your use cases and offer for volunteering to this wiki page. Thanks for your help!

2011-12-05

Common prefix elimination for episode lists

Something that I've been wanting to get into gPodder for a long time has finally landed in our Git repository: Common prefix elimination. We apply this to all episode titles that we show in GUI lists (e.g. in the Gtk UI and the QML UI), and it really helps to show you more information that is useful to you, especially on mobile devices:


On the first screenshot, you can see the Harmattan QML UI of gPodder 3.0.1 showing the Linux Outlaws episode list. As you can see, about half of the screen is "wasted" with information that is global across all episodes. On the second screenshot, you can see the development version of gPodder, showing the exact same screen, but this time with common prefix elimination enabled.

Right now, we have a very conservative algorithm for detecting common prefixes (i.e. it has to be a real common prefix across all episodes) - one could think about improving this with some heuristics, e.g. removing the podcast title in front of episodes, calculating the common prefix of a certain majority of episodes, etc. - I'm interested to see how the user base likes this feature, and if there are situations where this does more harm than good.