Shows "Please Select an episode" when shownotes shown and refreshing
channels without new episodes.
Shows downloaded and new count after channel name
Shows download progress before episodes
WebKit1 for Gtk2 support is deprecated in Debian, we may want
to re-introduce WebKit(2) support once we have ported gPodder
to Gtk3 and its GObject Introspection-based Python bindings.
See also: https://bugs.debian.org/790218
Changed sync code to check needed pace against available space and raise
error if there is not enough room for the track. This causes the track
to be listed in the failed_sync list. I could not get it to be in the
failed_sync list without raising an exception. Added a new sync failed
exception.
Also changed the main gtkui code to not include the failed_sync tracks
in the list of tracks to perform post-sync processing on. This prevents
tracks that were not copied from being marked as played.
This adds auto-discovery of the channel ID and new-style feed for old-style
(username-based) feed URLs when the V3 API key is available, and also adds an
extra menu item for migrating subscriptions.
The 'Episode details', or shownotes, could be displayed but not
hidden. Further, the menu item and the context sensitive menu gave
inconsistent behavior when multiple episodes were selected. And
I've removed the 'Episode details' context sensitive menu from the
'Progress' notebook page since this displays the shownotes on the
hidden 'Podcasts' notebook page.
steps to reproduce:
0) have update setting on "download immediately"
1) when a new episode starts downloading
2) pause and then resume the download
3) manually update the podcast
result: a duplicate download task gets added/queued, can lead to corrupt
downloads.
there are other circumstances that can trigger the bug, but this one is
simplest/most reliable to reproduce.
while i can't claim to understand all of the download-handling code,
this seems (more) correct, and fixes the problem.
note: if the code looks confusing, the git diff patch just came out
weird -- compare the "before" and "after" files manually..
When all downloads/syncs are finished, gPodder displays a summary of
downloaded episodes, having cut the titles if they are too long.
However, Russian descriptions are regular, non-unicode python strings,
and gPodder may cut only a part of a multi-byte UTF-8 sequence. It
causes an exception like this:
UnicodeDecodeError: 'utf8' codec can't decode byte 0xbe in position 51:
invalid start byte
This patch fixes that by converting the title to a unicode string if
it's not unicode.
This bug is similar to bug 1825, commit
e1ce9b0551.
In src/gpodder/gtkui/main.py: Use "callback is None"
instead of "callback==None", and also make the function
signature nicer (spacing, newline).
Remove trailing whitespace in all files affected by the
latest device sync patch:
git show HEAD | diffstat -l | xargs sed -i 's/\s\+$//g'
