When a plugin is enabled, disabled and then enabled again, on that second enable, that instance is being garbage collected causing the loading of the plugin to fail. Work around that until we can narrow down why is this is happening on the second enable.
Two more commands were added to setup.py:
* develop_plugins - Installs each of the plugins in development mode
* egg_info_plugins - Create the '.egg-info' distribution directories for each plugin. This will make the plugin discoverable by deluge
Both these commands are intended to be used while developing deluge.
Translations are now setup on `deluge.common`. Where they used to be setup, a call to `setup_translations(setup_pygtk=False)` is now made.
Every call to `pkg_resources.resource_filename()` is now made through `deluge.common.resource_filename` to make sure that we're loading data from the right deluge install.
While developing, if there's a second deluge package, installed globally and another in develop mode somewhere else, while pkg_resources.require("Deluge") returns the proper deluge instance, pkg_resources.resource_filename does not, it returns the first found on the python path, which is not good enough. Work around this issue.
Since the plugins namespace was merged into master, calling "logging.getLogger(__name__)" will result in a properly named logger for plugins which will allow logging filtering.
The previous workaround "getPluginLogger()", is now deprecated.
Implement a pre-filter for the search-box which will filter the currently visible torrents while waiting for the filter request to sent to the demon. This will make the searches seem way faster :)
Now, instead of permanently having a search box to filter the visible torrents by name, we now, mimic a toolbar just for that, mapped to CTRL-F. There's also a menu item in the "View" menu and a toolbar icon to toggle it. Implemented "Match Case" for the search.
