logs archiveBotHelp.net / Freenode / #mopidy / 2015 / September / 16 / 1
jodal
adamcik: I'm tempted to port mopidy_spotify.images to requests and fix proxy support there too, and then release m-s 2.1
adamcik
np given that mopidy itself now depends on it
jodal
but not tonight :-) *zzz*
adamcik
(Action) removed a bunch of skips / expected failures and started fixing them
I'm getting close to most of the remaining issues likely being bugs
henkkeli
hey, anyone have ideas how to organise split albums?
like, 2 artists on 1 album, and i'd like to see the album under both artists
apparently symlinks wont work
jodal
adamcik: we need to get moving on the pygi transition... mopidy and all extensions will be autorm-ed from debian testing oct 16
adamcik: so if you suspect the remaining issues are bugs, maybe we should base a pygi branch on top of what you have now
bjorn`
wait, what?
Freso
I'm with bjorn`; Debian is removing *old* software now?? D:
jodal
debian is removing everything that depends on gstreamer 0.10, to get rid of it for stretch
if we do nothing, mopidy will still be in stable and unstable, just not testing
bjorn`
ah. makes sense.
jodal
it's a nice way of making people take release-critical bugs seriously some 1.5-2y before the release, making sure the RC bug count is kept down and the release will be possible to do on time
adamcik
jodal: bad timing though, I'll be gone all of next week without a computer :-) and then the week of the deadline is a work summit
jodal: but yeah it's getting closer
we can probably consider what I have gapless-part-1 soon
part 2 being doing all track changes that isn't a seek via about-to-finish
after that I think only stop and pause can break streams
or II hope at least
Freso
adamcik jodal: Typos have been fixed (though probably not live yet); the #metabrainz fix *is* live now though. Thanks for your prodding. ;)
jodal
adamcik: would you strangle me if I release an 1.2 with pygi while you're gone? how hard will be to rebase gapless on that?
adamcik
no idea
state right now is that most of the failing test are the backend routing tests
which I haven't done anything with yet
but those should not be effected by the gapless testing, other than play etc becoming async
jodal: 5 failed, 2055 passed, 1 xfailed in 11.00 seconds
almost there now
jodal
nice!
adamcik
though the current change is 8 files changed, 623 insertions(+), 629 deletions(-)
as I've had to rewrite almost all of tests/core/test_playback.py
s/rewrite/modify/
the xfail is that we don't jump to the next playable track in about to finish
and the failures are all event emission related and likely bugs and/or broken tests
jodal: but as for 1.0 / pygi, I would say you should just continue getting that ready. better to have it as a backup and worst come to worst we've gotten exceedingly good at postponing gapless :P
jodal
mhm, and everything done in a pygi branch is learning we can reapply if redoing that branch
I have it in a state where mopidy can start and I can push spotify audio at it without it crashing, but no audio output. for other backends, it crashes on playback, iirc.
crash-driven development so far. I know there's some unported corners of the code, so the next step is simply to read through it looking for code that obviously needs porting.
but seems I broke pyspotify last night... so another round of releasing there
adamcik
1 failed, 2059 passed, 1 xfailed in 11.82 seconds
jodal
:D
adamcik
after that I need to either give up and just commit the blob, or see about splitting things more up in nicer commits
2060 passed, 1 xfailed in 11.81 seconds
jodal
:-)
adamcik
though I also have a growing todo list of other loose ends and improvements
but will be needing to transfer all that to new bugs
jodal
sounds like a good plan. divide and conquer
adamcik
part of it is gapless part 2 and the rest is just other things that certain cleanups enable
like getting rid of the endless loop for unplayable case should be trivial now
(Action) has no idea what he would do without his precommit hook
jodal
break travis, like I do
adamcik
I have it setup to create an env on the side with exactly what is being commited
and then it does flake8 on the diff and py.test on the env
so I know that commits in isolation don't break anything
2 files changed, 464 insertions(+), 550 deletions(-) left now
« prev next »