logs archiveBotHelp.net / Freenode / #mopidy / 2015 / September / 2 / 1
adamcik
it was missing the https version or something like that
MattFisher
adamcik: thanks for clearing up my misunderstanding. very fundamentally wrong :)
Thanks for the help, I got everything working with icecast streaming :)
jodal
adamcik: I started a pygi/gst1 branch yesterday. remains to get it fully working. I assume it will be easy enough to rebase/redo on top of gapless if it was working before gapless.
Ulrar
Is there a way to add some file in a playlist if it's not in the database ? I have playlists with spotify songs, and I'd like to add a local mp3 simply
Tried to find a way to add a file by URI but can't find out how
txomon|sauron
jodal / adamcik, I am leaving my job at fon :( Really need to be with my gf so... I will be going london...
adamcik
congrats, sounds like good news imo :-)
never fun being split up like that
(Action) did something similar for about 14 months
txomon|sauron
indeed although the circumstances are not natural :( but anyway, I think we will sort them out :)
adamcik
hope it all gets sorted for both of you :-)
txomon|sauron
yeah, =) I am really nervous, but I think it will all get sorted =)
adamcik
jodal: did you have time to take over the playlist one I started?
jodal
I can have a second look
adamcik
jodal: great, I'm taking an other stab at cleaning up the enough of gapless to start PR
*a PR
marginal
I am having trouble playing an airplay stream. I'm using airfoil and airfoil seems to think that it's working but no sound's coming out of the pi. It plays spotify songs and other streams OK though
Is there something special I need to do to get airplay to work?
jodal
marginal: sorry, I've never tried shairport/airfoil/airplay
adamcik: your playlist expanding stuff... it is working, just missing tests and proper use of config?
adamcik
the dpaste implents the core algorithm for it at least
so yes I guess
jodal
mhm
adamcik
though I would probably rename it to resovle(uri)
jodal
I've migrated the http failure mode tests from stream to internal/test_http.py now
so the dpaste is next up
resolve() reminds me of dns... what about unwrap_stream()?
adamcik
the "current" name is dereference(uri)
we might also want to make the scanner.scan(uri, timeout=...) - that is depending on how we want to limit things
my thinking was that we have an overall time limit, so the scanner gets whatever is left for each level deeper it ends up going
and same for the download, but I've not tried this in practice properly
and before I forget, we will probably need to add some special casing for HLS
but I would probably postpone that for now
jodal
mhm, it's a bug about it in the 1.3 milestone
to retest it after pygi/gst1 transition
adamcik
the other loose end is file:// uris
which requests doesn't do out of the box
but that also sounds like a followup bug
or I can hack up something similar to http://cgit.freedesktop.org/gstreamer/gst-plugins-bad/tree/gst/hls/gsturidownloader.c?h=0.10 in the scanner code :P
(ab)using gstreamer elements as a generic file downloader :-)
hihi, I'm looking at https://github.com/mopidy/mopidy/issues/17 now
"This is a candidate for a 0.1.1 bugfix release." :P
(Action) might have broken the tests for that case...
jodal
hehe
old times
adamcik
hmm, so regarding gapless...
the holdup is currently that I've started this hugeass refactoring of playback internals
with the goal of making it clean enough to better handle the disconnect between playing track and pending track
jodal
is it possible to land that part first?
adamcik
we could in theory ship gapless without this, _but_ behaviour during the switchover will be rather supprising to end users :(
it might be ok for develop though
jodal
we could ship it if it was turned off by default, and you would have to set core/experimental_gapless=true to turn it on
adamcik
perhaps
jodal
when we could get gapless in, I could continue on pygi/gst1, and you could continue on playback cleanup
because I only block on the gstreamer part, not the playback part
adamcik
I'm just wondering about how to get progress other than just committing to finishing what I started in one huge PR
jodal
if we feature flag it away, we can have it in develop and not end up deferring the next release for months just because the playback cleanup is missing
if doing it all now, I guess it would be nice if it was possible to review the playback cleanup first, then add the gapless part second
adamcik
that might be possible, but not a 100% sure
thing that ties them together is events setting pending track -> current track
but could probably be split
« prev next »