#beets

thomas_sch: It looks like beets doesn't have a choice -- there's no album that contains just those tracks.

You probably want to keep those songs as singletons.

howdy sampsyo

got all the -bb stuff passing now

Yay!

yo yo sampsyo

what's up?

Hi! Just checking in briefly while I get some work done

work ?

I was thinking of releasing 1.4.0 with "experimental" Python 3 support soon—do you see any potential problems with that?

yep, it never ends

there's no python 3 support until we do soemthign about the stdio

resolve everything in my test PR (which you never commented on)

Well yeah, it's not that anyone should *actually* use it; mainly I'd like to get the recent changes out so we can start finding bugs from the Python 2 side.

One thing at a time. :)

i don't see how it's gonna be much different than what they got in 1.3.19

so i think we should try to tackle the stdio first. and it really isn't hard

at least on linux/mac

and i think i know how to do it for windows too. we just need to talk about it a bit first

and try win-unicode-console

We did change a huge amount of stuff, and we introduced the six dependency, so there's a lot of surface area already

if the six even has a chance of breaking anything.. i'd be like.. waaay annoyed

i'm ok doing a release , but I'd prefer not to talk about py3 at all

well we can talk about py 3, but let's not call it py3 support.. more like.. changes on the road to py3

Yeah, that's what I was getting at

I just raise issue with ""experimental" Python 3" that you said

Sure; I worded that wrong

quick question ..

tem.path = '/ipfs/{0}/{1}'.format(album.ipfs,os.path.basename(item.path)) so what do i use there around os.path.basename when is bytestring?

just text_string ?

If you know it's a bytestring, then text_string is unnecessary—you can just .decode() it

But for the encoding, it's probably wisest to use _fsencoding()

And since, in this case, it's more important to be a usable name than to be exactly correct, the appropriate error more is probably `ignore`

*error mode

uggh.. that's all so very confusing

can we just rage about this nonsense for awhile? a long while? break some glass maybe

i don't know

i wish i had a punching bag

encoding stuff, especially filename encoding stuff, is horrible nonsense, it's true

it's 2016 and we still have to deal with Latin-1, etc., etc.

and OSes that think ASCII is a reasonable encoding

do you ever play any video games?

not often, to be honest

saying what i said reminded me of a line from Bastion that the narrator says.. "The kid just rages for awhile" when you go around breaking breakable items

ha—I have actually played Bastion, and I remember that line

i usually buy all the humble indie bundle games.. and that's about all i play.. i still have a huge backlog of stuff i never get to

before a few years ago.. i was a coowner of a bookstore/cafe .. not much time for games then

sampsyo, check your pm ..

you gave up the cafe thing?

i left town

twas in baltimore

aha

it wasn't a normal coowner arrangement. as per the pm

i think i'd be slightly less annoyed by all this if it all wasn't so verbose and make the code less clear

Agreed.

maybe once it's all there, we can rethink our methods

Yeah—a more uniform way of dealing with paths specifically would be worth pursuing. One can dream that pathlib would be that way…

It can be.. in 3.6.. I wonder if we can trigger __fspath__ like functionality ourselves in preparation

i know they've adopted it for all sorts of methods like tempdir and open() and whatnot, but no idea if it has or will percolate to sqlite3 and friends

It certainly should!

i wish they'd take it so seriously that they'd backport it to 3.5 or make it available via __future__

well 3.4 too.. but i doubt we'd get that lucky

alternatively.. distributions could make it easier on us by providing more python versions :(

it sucks that windows and mac are actually easier to deal with for this stuff

becauwe it's no big deal for windows programs to require an updated python

and in the end, we'd use pyinstaller or something anyways

yeah, but a huge hassle if your Debian doesn't have the right Python, for example

it is a bit tragic

yeah.. it's linux that's the problem

in the end.. it's really ubuntu and debian (and their derivatives)

because even rhel 7 has SCL.. which allows you to get newer pythons

fedora will be behind for about 6 months.

they don't even have an LTS. so the current release -2 is always EOled

oh.. it's a 6 month cycle

well flatpack/snaps might save us eventually

depending on if they offer one with updated pythons..

and there's the -bb fixes

it wasn't that bad

now that i'm getting a chance to look at the codebase more on a whole, i'm noticing some perhaps easy wins for refactoring

sampsyo: yeah I probably will keep those as singletons, but it forces this on some albums that are nearly complete. My problem is that my collectio is neither complete nor in order :/

any tips to identify freezing in beets? I just started to import my collection, and it's copied 2 albums and hasn't made any progress for the past 20 mins or so and doesn't seem to be doing anything

fryguy: http://beets.readthedocs.io/en/latest/faq.html#...

DjSlash: it doesn't seem to be copying or doing anything, and no real cpu usage going on