Sonntag, 30. Dezember 2012

PublicTransport 0.11 alpha 2

There is another alpha version for PublicTransport (0.11 alpha 2). It includes many build, crash and other fixes, some small new features (eg. snap scrolling at the top of departure items in the applet), a polished TimetableMate UI and one bigger change: Provider plugins are now hosted on, you can watch them on or using the download dialog in the applet configuration.
This is great for a number of reasons: Provider plugins can be updated independently of the data engine (ie. users can get fixes on the plugins very quickly), better performance because the data engine does not need to track all available plugins (only downloaded ones), users can add new plugins very easily (TimetableMate can do it), use of social features like comments, ratings for provider plugins, etc.
This fixes the long broken GHNS feature of the applet (libpublictransporthelper). I moved from to an easy to create Synchrotron repository first (just add publictransport to synchrotron.git and add provider files to synchrotron-sources.git). This was a quick way to test it and it worked. But it misses most GHNS features like upload, ratings or descriptions. Therefore I moved to

All provider plugins are now available to download, also for some new GTFS providers. No provider plugins will be installed by default. Unused plugins from previous versions should be removed (eg. "make uninstall" or manually from /usr/share/kde4/apps/plasma_engine_publictransport/serviceProviders/ or a similiar path).

The engine was updated to never delete connected data sources, eg. a "ServiceProvider <id>" source was deleted when the provider was uninstalled. Instead now the sources get updated accordingly. This makes it possible to install a new provider plugin and directly use it. The cache is now also properly cleaned up from data for no longer installed providers.

All build problems (missing protobuf or pthread, not found "javascriptcompletiongeneric.h", etc.) should be fixed now. There were also some fixes for crashes that happened randomly in the data engine (bad synchronization when a job was aborted). In the runner these crashes happened much more often because of the many job aborts while typing. Getting route data for HAFAS plugins now works more reliably. The GTFS service/importer also saw some improvements and is now able to import more GTFS feeds.

Thanks for testing and a happy new year! :)

1 Kommentar: