* Remove labelSeparator and manually add ':' so text matches gtk translations.
* Use consistent quotes around strings. This can affect gettext script picking up
marked strings.
* Added the equivalent deffered translation as gtkui for Filters and Progressbar
The status strings were incorrectly marked for translation which when combined with
some translations using 'connected' and 'online' as the same word resulted in
users being unabe to connect to running daemon.
* Removed translation markup from json_api but left as original capitalised word in
case other third-party scripts do comparison on these status strings.
* Added translation markup prior to displaying ConnectionManager using template.
* Reworded password prompt and added translation markup.
* Update gettext.js
Display settings for the WebUI are persisted using cookies created by
Ex.state.CookieProvider. When no expiration date is provided, a default
value of (now + 7 days) is used. This causes display settings to be
lost frequently.
This fix adds an 'expires' parameter with a value of (now + 10 years).
This change does not affect the lifespan of the session cookie, which
is created by a separate system.
With xgettext set to python it will parse the comments in javascript files, so
single apostrophes or quotes are flagged as 'warning: untermined string'.
This change just rewrites js comments to not use apostrophes.
Removing multiple torrents with remove_torrents will be faster
than calling remove_torrent for each torrent, as the state file
will be written only once, after all the torrents are deleted.
Use a context manager to wrap the common steps:
1) disconnect the treestore from the listview
2) disable treestore sorting
3) add rows (different in add dialog vs files tab)
4) enable treestore sorting
5) connect model to listview
* Added custom trial reporter for TODO with test example in test_torrentmanager.py
* Set Stats plugin tests as todo
* Disable new_release_check when running unit tests
* Added pytest.mark.slow to test_core.test_test_listen_port
* Get rid of unit test warnings (Caused by bad names in test classes)
* Removed warnings.filterwarnings in test files.
* Added separate tox target for generating test coverage HTML report.
* component registry shutdown() now cleans up the component list
this ensures that no old components are left when running unit
tests.
* Added class BaseTestCase that all tests that create components
should inherit from. It verifies the compoent list before and
after the tests are run.
This issue was caused by an unclean shutdown of Deluge, usually on system shutdown, and upon rebooting
the PID stored in the lockfile is in used by another process thus the lockfile is never removed. It
affects users with Deluge set in startup applications as the PIDs are more likely to be reused.
* Lockfile is removed if Deluge is restarted in IPC.
* Renamed the old_tempfile variable to make it clearer as to it's role.
Targets:
* Runs the unit-tests for python 2.7
* Tests unit-test coverage
* Try to build docs
* Code style checks:
* flake8
* isort
Codes changes:
* Fixed tests for httpdownloader (using tmp dir)
* Implemented a couple of tests for Stats plugin but they fail to run on travis
Issues:
* Can't get py26 to work because of installing libtorrent through apt and
the option system_site_packages fails for 2.6.
* Tabs can now be scrolled if too many for height.
* Change window_pane_position default to 235, based on default 480 window height.
* Set the vpaned size before show call.
* Reduce top padding and set bottom padding to 2, to account for decender letters.
