about the package on pypi? I *think* pip by default doesn't install development packages, so the picard_dev one might be unnecessary
2017-05-09 12948, 2017
samj1912
i just kept it there to test :P
2017-05-09 12951, 2017
Mineo
but that can be investigated outside of this meeting
2017-05-09 12900, 2017
Freso
Is it possible to classify a package as "development" in PyPI? If so, then I'm fine with moving picard_dev packages to picard proper.
2017-05-09 12939, 2017
Freso
PyPI should definitely not claim that anything 2.x is stable or "production ready" as of yet at least. :)
2017-05-09 12955, 2017
Freso
So.
2017-05-09 12951, 2017
Freso
Resolution: Investigate whether PyPI can have "in-development" package version, and if so, add picard 2.0.0devX builds to PyPI - if not, leave PyPI be until 2.0 is released.
2017-05-09 12902, 2017
Freso
Si?
2017-05-09 12906, 2017
Mineo
+1
2017-05-09 12908, 2017
zas
+1
2017-05-09 12909, 2017
samj1912
+1
2017-05-09 12922, 2017
Freso
🙌
2017-05-09 12924, 2017
Freso
!next
2017-05-09 12925, 2017
CatQuest
+1
2017-05-09 12945, 2017
Sophist_UK
+1
2017-05-09 12947, 2017
Freso
7 minutes! One more thing! :)
2017-05-09 12952, 2017
Sophist_UK
Threading?
2017-05-09 12902, 2017
samj1912
I would just like to point out that we should investigate asyncio as an alternative to threading
2017-05-09 12902, 2017
Freso
I think Sophist_UK would like us to talk threading. :)
2017-05-09 12913, 2017
Sophist_UK
(My eyebrows are getting a little bushy.)
2017-05-09 12930, 2017
Freso knows next to nothing about threading or related subject, but would like to learn :)
2017-05-09 12939, 2017
CatQuest ditto
2017-05-09 12942, 2017
Sophist_UK
samj1912: Can asyncio do async file i/o?
2017-05-09 12949, 2017
samj1912
Sophist_UK: yup
2017-05-09 12951, 2017
zas
+1 for "investigate asyncio"
2017-05-09 12958, 2017
samj1912
I am reading about it
2017-05-09 12903, 2017
kepstin
Sophist_UK: asyncio's primary purpose is to do async file io
2017-05-09 12916, 2017
Sophist_UK
In which case, yes, let's investigate.
2017-05-09 12919, 2017
samj1912
and async network requests
2017-05-09 12929, 2017
kepstin
I've used it for some personal projects, it's fairly easy to use - the downside is that you really need to use a recent python 3.x version to make full use of it
2017-05-09 12934, 2017
Mineo
there's no stable library for integrating pyqts event loop with asyncio afaik
2017-05-09 12938, 2017
Freso
samj1912: In *nix, everything is a file. Incl. network requests. ;D
2017-05-09 12946, 2017
Sophist_UK
samj1912: PyQt already does async network for us.
2017-05-09 12959, 2017
Sophist_UK
Mineo: True.
2017-05-09 12928, 2017
kepstin
and mutagen, for example, still does syncronous IO, so that's a limiting factor...
2017-05-09 12928, 2017
Freso
kepstin: Well, Picard 2 will by Py3.6 (maybe 3.5), so that's fairly recent... :)
2017-05-09 12931, 2017
Sophist_UK
Also not sure what the impact would be on GUI responsiveness.
2017-05-09 12944, 2017
samj1912
well yes, nada on that front (pyqt + asyncio)
2017-05-09 12910, 2017
Sophist_UK
And mutagen as kepstin says.
2017-05-09 12916, 2017
Freso
So... resolution: investigate asyncio (more)? Or is there something else about threading?
2017-05-09 12920, 2017
Freso
(3 more minutes.)
2017-05-09 12926, 2017
kepstin
so.. since mutagen doesn't use asyncio interfaces, and pyqt+asyncio seems to not be integrated...
2017-05-09 12948, 2017
Freso
Maybe lazka (hi :)) has plans of switching to asyncio? :p
But apart from mutagen and PyQt incompatiblity with asyncio, everything's perfect.
2017-05-09 12947, 2017
samj1912
well, from what I have read about threading, the only way to truly use python + threading is either to shift to something like pypy or jython
2017-05-09 12948, 2017
UmkaDK_ has quit
2017-05-09 12948, 2017
Freso
Well, those are two pretty big things.
2017-05-09 12908, 2017
samj1912
which either dont have GIL or are smart about using it
2017-05-09 12923, 2017
Freso
("GIL"?)
2017-05-09 12928, 2017
samj1912
or we can get done to gritty c and manage it ourselves
2017-05-09 12947, 2017
samj1912
Freso: basically since cpython isnt thread safe, it only allows 1 thread to be run at a time
2017-05-09 12948, 2017
kepstin
Freso: "global interpreter lock" - only one thread in python can be executing python code at any given time
2017-05-09 12900, 2017
Sophist_UK
Freso: GIL = Global Interpreter Lock - essentially makes Python execute on a single core.
2017-05-09 12904, 2017
samj1912
yup
2017-05-09 12922, 2017
Freso
So...
2017-05-09 12927, 2017
Mineo
jython is basically dead and pyqt doesn't work on pypy
2017-05-09 12929, 2017
kepstin
using threads is a way to handle blocking IO mostly, since if a thread is waiting on IO, that means it's *not* currently running python code, so some other thread can run
2017-05-09 12932, 2017
Freso
What's the plan? :)
2017-05-09 12934, 2017
Sophist_UK
However, that does not mean that threads don't have a use - just that you have to be careful about how you use them/
2017-05-09 12951, 2017
Sophist_UK
Because they are not preemptable.
2017-05-09 12902, 2017
kepstin
Freso: in the near term, I'd suggest looking at running mutagen operations which do blocking IO inside threads
2017-05-09 12918, 2017
Sophist_UK
kepstin - we already do that
2017-05-09 12938, 2017
samj1912
saving yes
2017-05-09 12944, 2017
samj1912
loading (not so much)
2017-05-09 12910, 2017
kepstin
yeah, there's some severe ui responsiveness issues when loading lots of files.
2017-05-09 12923, 2017
kepstin
particularly over slowish network filesystems like I use :)
2017-05-09 12959, 2017
Sophist_UK
PICARD-936 gh:P604 was an attempt at improving this.
We should cherry pick the first half of this and see if we can make it stable on v2.
2017-05-09 12927, 2017
samj1912
I didnt expect it to be solved withing minutes :P
2017-05-09 12938, 2017
Freso
So.
2017-05-09 12947, 2017
CatQuest
Sophist_UK: re this: not only devs want to run 1.4.x and 2.x at the same time, i for one would like this ability very much.
2017-05-09 12949, 2017
Freso
Are we going to try for same time next week, or... ?
2017-05-09 12958, 2017
CatQuest
+1111
2017-05-09 12903, 2017
samj1912
can we do it a bit earlier? :P
2017-05-09 12905, 2017
Sophist_UK
I am up for same time next week.
2017-05-09 12910, 2017
CatQuest
that works too
2017-05-09 12910, 2017
samj1912
its 2 oclock here
2017-05-09 12915, 2017
Sophist_UK
I can do earlier too.
2017-05-09 12915, 2017
CatQuest
+1 samj1912
2017-05-09 12922, 2017
Freso
samj1912: Yeah, silly India time. 🙄
2017-05-09 12922, 2017
Freso
:p
2017-05-09 12926, 2017
Sophist_UK
1 hr earlier?
2017-05-09 12931, 2017
CatQuest
stakars sam, it's way past bedtime :O
2017-05-09 12940, 2017
Sophist_UK
sam is a night owl.
2017-05-09 12949, 2017
CatQuest
well so am i, still at 2 am yo uare tired
2017-05-09 12956, 2017
Freso
18:00 UTC (20:00 CEST)?
2017-05-09 12906, 2017
samj1912
works
2017-05-09 12909, 2017
Sophist_UK
Good for me.
2017-05-09 12919, 2017
Sophist_UK
1.5hrs duration again?
2017-05-09 12920, 2017
CatQuest
Freso: i like your import settings plugin idea :D
2017-05-09 12920, 2017
zas
not great for me, but ok
2017-05-09 12921, 2017
Freso
Mineo, zas, CatQuest, kepstin, ... ?
2017-05-09 12930, 2017
CatQuest
20:00 cest is fab
2017-05-09 12934, 2017
Freso
Sophist_UK: I guess. :)
2017-05-09 12946, 2017
zas
1 hour, and clearer topics
2017-05-09 12956, 2017
kepstin is at work at this time, but will try to keep an eye out for meeting stuff
2017-05-09 12959, 2017
Freso
Yes, that would be great.
2017-05-09 12904, 2017
CatQuest
less going bakc and forth
2017-05-09 12923, 2017
Freso
samj1912: Can you make a more minutious agenda for next week? ;)
2017-05-09 12923, 2017
CatQuest
and maybe start with the long sophist paper whne we have time and peopel are not tired
2017-05-09 12932, 2017
samj1912
Freso: yup :P
2017-05-09 12943, 2017
samj1912
I didn't know this was going to be a weekly thing :P
2017-05-09 12945, 2017
CatQuest
and totlaly megaping me
2017-05-09 12947, 2017
samj1912
ill take care next time
2017-05-09 12951, 2017
CatQuest
:D
2017-05-09 12914, 2017
CatQuest
and i'm so, so sorry for showing up like in the last third and being distruptive and a nuicance
2017-05-09 12925, 2017
CatQuest
🙇 everyone
2017-05-09 12930, 2017
Freso
CatQuest: With all due respect, I don't think Sophist_UK's paper is the most critical thing to get resolved to get Picard on the path to 2.0. It's something that definitely *should* get taken care of before 2.0, but it's not as fundamental as the other things.
2017-05-09 12934, 2017
Sophist_UK
CatQuest: Perhaps ask questions by PM next time?
2017-05-09 12941, 2017
Freso
Sophist_UK: Naw.
2017-05-09 12943, 2017
samj1912
oh, I am fine with 19:00 UTC is we shift it to something like friday or weekends
2017-05-09 12944, 2017
Sophist_UK
Freso - I agree.
2017-05-09 12953, 2017
samj1912
zas: ^^
2017-05-09 12912, 2017
zas
19:00 utc on friday ?
2017-05-09 12914, 2017
Freso
Sophist_UK: :)
2017-05-09 12914, 2017
CatQuest
Freso: oh i just menat like "to get it out of the way2 it seemd a big thing :)
2017-05-09 12915, 2017
zas
ok
2017-05-09 12918, 2017
CatQuest
oh no not firday!
2017-05-09 12926, 2017
Freso
I don't like Friday meetings.
2017-05-09 12932, 2017
samj1912
lol :P
2017-05-09 12934, 2017
Freso
Esp. if it
2017-05-09 12934, 2017
CatQuest
yea :nogood:
2017-05-09 12938, 2017
Sophist_UK
Leave my paper - except cherry pick the bit about restruturing mbxml / format code.
