All times shown according to UTC.

Time Nick Message
00:05 JasonWoof joined #miro-hackers
00:05 JasonWoof left #miro-hackers
00:05 JasonWoof joined #miro-hackers
00:11 bendk left #miro-hackers
00:23 arpu left #miro-hackers
00:52 ajonas joined #miro-hackers
01:02 DGMurdockIII left #miro-hackers
01:31 Odysimus joined #miro-hackers
02:03 Odysimus left #miro-hackers
02:17 Odysimus joined #miro-hackers
02:18 ajonas left #miro-hackers
02:25 ajonas joined #miro-hackers
02:50 ajonas left #miro-hackers
03:19 mattjohnson left #miro-hackers
03:22 mattjohnson joined #miro-hackers
03:23 Odysimus left #miro-hackers
09:23 maggie_s left #miro-hackers
10:00 paulproteus deardiary: Diagnosed the issue with http://bugzilla.pculture.org/s[…]_bug.cgi?id=12324
10:19 deardiary: ...and hopefully just fixed it.
13:02 janetPCF joined #miro-hackers
13:20 willkg joined #miro-hackers
13:27 willkg windows build box hung again last night.  guess i'll be spending all day running ram tests.
13:37 janetPCF: i'm pushing out a new nightly now.
13:38 janetPCF ugh - stupid windows build box
13:41 z3p good morning
13:52 z3p left #miro-hackers
13:53 willkg deardiary: i updated cygwin things on the windows build box and now i'm running ram tests.  it's passed 3 sets of tests so far, but paulproteus says that can happen with bad ram.
13:53 afk for coffee.
13:54 z3p joined #miro-hackers
14:01 Jarrhead joined #miro-hackers
14:04 zanoi morning
14:17 willkg kcw: i'm secretly looking forward to starting miro and it opening the tab i was last in rather than the miro guide.
14:17 kcw: jwz is looking forward to it, too.
14:20 janetPCF: so i updated to libtorrent 0.15.5 and that doesn't fix any of the issues i was having with torrents.  thus i don't think the issues i was having are related to bug 13549.
14:22 janetPCF: i'm going to spend the rest of today poking around with libtorrent download issues.
14:22 and shoveling snow.  it's snowing again today.
14:22 janetPCF 50 and sunny here
14:23 I still get torrent stuck in starting up too
14:25 willkg:  i do like the stored crash reports - I never saw the bug  d/t the spam factor - all my bugzilla mail was going tothe spam folder
14:26 willkg janetPCF: that "torrent stuck in starting up" behavior is the behavior i'm seeing with and without libtorrent 0.15.5.  my current theory is that the cause of that is different than bug 13549.  so after i investigate 13549 a bit more, i'll do something with that bug.  then i'll move on to the problem we're having.
14:27 one of the things i did was add a "torrent details" column to the downloading tab.  it's a shorthand set of details.  i doubt the ui people will like it, but it's hard as hell debugging libtorrent problems without seeing the data.
14:27 janetPCF i agree - I think people had the libtorrent 0.15.x issue before we had the starting up problem, and we had the starting up problem with earlier libtorrent versions
14:28 0.14.whatever
14:28 willkg i don't recall having this specific starting up problem with other libtorrent versions.
14:28 i want to be really careful that we're not lumping all the starting up problems together.  i think there are different issues here.
14:31 so right now i'm looking at 13549 which is about starting up problems with libtorrent 0.15.4 on Arch.
14:31 i'm making sure that issue is different from 15778 which is the starting up problems i'm having with some torrents which seems to be a regression from miro 3.5 or 3.5.1.
14:49 zanoi willkg: do you mind if i add some text about which shuffle/repeat policies I went with to the wiki?
14:49 willkg: or is that only for bigger things
14:50 willkg: i think it could save someone a lot of time in the future though
14:50 willkg zanoi: please feel free to document any kind of design stuff in a design doc in the wiki.
14:50 you're exactly right--it'll save someone gobs of time in the future.
14:51 further, it makes my job much easier when i'm trying to write the bits in the miro user manual because i have a document that specifies the behavior.
14:51 zanoi: there's a category and a rough template here: http://develop.participatorycu[…]tegory:MiroDesign
14:51 zanoi: plus, you're my new best friend for asking.
14:52 zanoi heh
14:53 willkg that goes for anyone else implementing new things--design docs are really helpful!
14:55 zanoi willkg: should there be absolutely no implementation stuff on the wiki?
14:56 willkg zanoi: what do you mean by "implementation stuff"?
14:59 zanoi willkg: not sure how to explain it...i boiled down the next track progression policies to a few if statements
15:00 willkg zanoi: so are you asking whether you can put code in the design doc?
15:00 zanoi willkg: well very pseudo-ish pseudo code
15:01 willkg i think a design doc should list the problem statement, the requirements involved, related bugs, links to discussions, ...
15:01 then it should walk through the behavior implemented.  if pseudo-code helps with that, that's fine.
15:01 zanoi though since you've started calling it design doc it sounds wrong to add pseudocode :)
15:01 ah ok
15:01 willkg and also outstanding issues, future directions, etc.
15:02 but generally, anything you document is super and better than no documentation at all.
15:02 mmm....  i think actual code and pseudocode are totally appropriate for a design document.
15:03 zanoi: so here's the one i wrote on extensions: http://develop.participatorycu[…]p/ExtensionSystem
15:04 zanoi: and here's one glee wrote on sharing: http://develop.participatorycu[…]index.php/Sharing
15:07 zanoi willkg: thx
15:12 zanoi left #miro-hackers
15:16 ajonas joined #miro-hackers
15:19 willkg afk a smidge.
15:25 zanoi joined #miro-hackers
15:32 bendk joined #miro-hackers
15:32 bendk hey everyone
15:34 zanoi heya
15:34 z3p hey ben
15:51 Jarrhead left #miro-hackers
16:00 jamessan bendk: willkg asked me to prod you about seeing high CPU usage (~40% per concurrent download) when downloading files.  looks similar to 13433.  I'm using 3.5.1 with libcurl 7.21.3 (on Debian/sid if it matters)
16:00 bendk okay, i'll take a look at that in a little while
16:01 jamessan it's not really noticeable with audio files since they download so fast (for me), but I definitely see it when downloading videos
16:13 janetPCF willkg:  here's the original fast resume bug: http://bugzilla.pculture.org/s[…]_bug.cgi?id=13052
16:16 arpu joined #miro-hackers
16:21 paulproteus z3p: I really need to pick your brain about Thumbnailable in the Miro Community source code.
16:21 So I'm hoping you're there, and willing to talk.
16:21 z3p paulproteus: sure
16:21 willkg janetPCF: thanks!
16:21 paulproteus So... first of all, on some dev sites, we spend >1 minute just looping where the S3 backend keeps incrementing the filename looking for a valid one.
16:21 I guess because delete() doesn't do that we thought it does, or something.
16:22 The django_storage.save() function doesn't guarantee that it saves with the filename you suggested.
16:22 It returns the filename it actually used in the end.
16:22 willkg afk a bit to put on pants and get some food.
16:23 paulproteus There's this temporary_file_path thing I don't understand.
16:24 z3p paulproteus: let me pull up the code
16:25 paulproteus I made a bit of a mess of it last night, so you might want to also look at a version from ca. 1d ago.
16:25 z3p paulproteus: the temporary file path thing is to work around a Django issue
16:25 paulproteus And then I don't have a systematic understanding of what parts of the rest of the code use the various methods provided by the Thumbnailable mixin.
16:26 I don't know why the S3 storage module is so willing to increment forever looking for a different name to use.
16:26 That's where the time is going. I'll read the code for that module right now.
16:26 z3p paulproteus: I didn't write most of that code, so I'm not sure how helpful I'll be
16:26 paulproteus Heh (hi paroneayea ...?)
16:27 z3p paulproteus: he didn't either
16:27 paroneayea hrm?
16:27 yup, I sure didn't
16:27 z3p paulproteus: I forked it to fix some bugs I ran into, but it's not code PCF wrote
16:33 willkg dev call notes: http://bluesock.org/~willg/blo[…]all_20110202.html
16:35 ajonas paulproteus: can you take a crack at http://bugzilla.pculture.org/s[…]_bug.cgi?id=16096 next, after the thumbnail thing?
16:37 paulproteus Breakfast for me for now.
16:40 kcw deardiary: implementing #16075 - separate available columns from default columns
16:40 and updating metadata design doc
16:44 z3p deardiary: working on unbreaking feed/playlists folders
16:53 bendk jamessan: you still here?
16:59 wow, upgrade 134 is slow
17:05 jamessan bendk: yeah
17:11 bendk jamessan: okay, so you're using 3.5.1 and if you're downloading just 1 item miro uses 40% of the CPU?
17:15 also, are we talking about torrents or http downloads?
17:17 jamessan http downloads.  e.g., I added the feed http://linuxconfau.blip.tv/rss and was downloading videos from there
17:19 arpu left #miro-hackers
17:20 bendk does it happen for all feeds, or just that one?
17:26 jamessan all feeds that I've tried.  I had only download podcasts before, so I hadn't noticed but they spike too.  the download is just so quick that I hadn't noticed the side effects
17:31 bendk is the CPU usage for the downloader or for the regular miro process?
17:32 jamessan downloader, but the weird part is that I don't see anything when I strace it, so it seems to not be making system calls -- just python stuff
17:32 (at least weird from an outsider's perspective)
17:34 willkg zanoi: the shuffle/repeat stuff works great.
17:34 zanoi yay
17:40 bendk jamessan: that's not surprising, python is definitely the slowest part of the system
17:40 so I don't have any suggestions at this point, can you write all the info you just gave me into a bug report?
17:41 things seem to be working well on ubuntu lucid, so maybe it's an issue with the libcurl version
17:41 maybe it's something else, I don't know at this point
17:41 jamessan do you want me to reopen 13433 or just create a new one?
17:41 willkg jamessan: no--that bug was for when the http downloads were entirely in python.  this is a different issue.
17:41 jamessan ok
17:42 willkg though the symptoms are similar.
17:49 zanoi afk, errands
18:00 willkg this sucks...  i don't want to shut miro down so i can work on bugs because i'm listening to things i want to listen to.
18:01 can someone check in some crashy stuff so that miro isn't really usable again?  that'd be better for my productivity.
18:01 jamessan but miro remembers where you were :)
18:02 willkg i'm listening to irish reels on shuffle/repeat.  it's really nice.
18:02 jamessan: oh!  i bet you're having a snow day!
18:02 jamessan yeah. been able to work from home yesterday and today ;)
18:02 willkg nice!
18:03 jamessan and taking a lunch break right now after a 2 hour concall
18:03 was only supposed to be one hour :/
18:03 * willkg nods.
18:03 willkg we had a call this morning that was really to the point, fast and furious.  i really liked it.  some calls drag on forever.
18:04 now that i think about it, i probably did half the talking.
18:05 janetPCF1 joined #miro-hackers
18:05 willkg i think i'm going to have to get miro and my music working on another computer so i can listen and do work at the same time.
18:06 janetPCF left #miro-hackers
18:10 willkg deardiary: 9 ram test passes suggests windows build box has working ram.  i ran windows updates, cygwin updates, and checked to see if the drives needed to be defragged.
18:25 z3p deardiary: trying to wrap my head around how drag&drop works
18:26 willkg z3p: might be worth pinging bendk--pretty sure he implemented what's there.
18:27 z3p willkg: the new logic I need for the sidebar is more complicated, which is tough
18:27 * willkg nods.
18:36 kcw z3p: if what you're working on re:sidebar happens to be able to be done well by making a unified list of all sidebar tabs.... that would also take care of the difficult part of the "save tab list selection" bug (part of #12861)
18:38 z3p kcw: that would make everything significantly more complicated, I think
18:38 kcw: in what way is it complicated now?
18:39 kcw z3p: I've just looked over it briefly but I don't see any easy way to get the current selected tab, or reselect it when restoring
18:39 *current selected tab(s)
18:42 z3p: on second look it's probably easier than I'd thought, anyway
18:42 z3p kcw: app.tab_list_manager.get_selection()
18:53 willkg afk a smidge.
18:54 cool-RR_ joined #miro-hackers
18:55 cool-RR_ Hello!
18:55 zanoi hi
18:55 cool-RR_ I wanted to ask: Are Miro tests visual? i.e. Do they use Sikuli or something?
18:55 jamessan bendk: ok. #16117 filed
18:56 bendk jamessan: great, thanks
18:56 cool-RR_: The tests that are checked in are using unittest.py
18:56 zanoi cool-RR_: yes
18:56 bendk we have sikuli tests, but I'm not sure if they are publicly available
18:56 zanoi oh
18:57 bendk deardiary: worked on a pretty weird HTTP corner case with resume + auth info, hopefully it's fixed
18:57 cool-RR_ Ah, so you are experimenting with using Sikuli for your tests, but it's still not finished enough to release?
18:59 zanoi janetPCF1: ^
18:59 cool-RR_ And the existing tests are not visual? They interact directly with PyGTK?
19:00 zanoi cool-RR_: the unittests are for the non gtk stuff
19:01 cool-RR_ So I'm asking regarding the GUI tests
19:01 Are the GUI tests visual?
19:02 zanoi sorry, don't know much about that. why are you asking?
19:02 ajonas paulproteus: once tiers launches, we're keeping the manual implementation of custom domains for the time being, is that right?
19:02 paulproteus: trying to figure out the easiest way to tell people how to set that up
19:05 cool-RR_ left #miro-hackers
19:26 willkg oh, he's gone now.
19:26 oh, well.  i know everything there is to know about tests.  :)
19:27 the eggplant tests are stored in the miro-testing repository.  looks like janetPCF1 hasn't added the sikuli tests there, yet.
19:38 bendk: you on?  the httpdownloadertest is now failing for me after your last checkin.
19:38 bendk: three tests fail all with this sort of thing which forces a HadToStopEvent error: http://dpaste.com/378887/
19:39 bendk: i can paste the whole output if that's useful to you.
19:43 bendk I'll check it out
19:45 willkg bendk: also, the platform name for osx is "osx"--not darwin.  so those tests that are marked to be skipped on darwin get skipped on all three platforms.
19:45 bendk: i'll add some checking for that.
19:57 bendk: also, i'm working on 15840 and i'm not really sure what to do when decode_auth_header kicks up a ValueError or HTTPAuthPassword kicks up an AssertionError (which is in the __init__ which is icky).
19:58 zanoi back
20:00 bendk I think eventually we want to call the errback with AuthorizationFailed
20:01 I think with decode_auth_header, it's just a matter of catching the ValueError
20:02 I don't understand the HTTPAuthPassword issue too well, I don't have a suggestion about that one
20:02 paulproteus Yeah, manual custom domains for now.
20:02 There's a script that makes it fairly easy, so it's not too bad.
20:03 bendk willkg: unittests should be fixed now
20:03 paulproteus ajonas: ^^ also, I was feeling kind of generically stressed-out for a lot of today and so took a longer break than usual, but I'll be back to work shortly.
20:03 Also I'm installing Debian on my new laptop.
20:03 ajonas k
20:03 willkg bendk: ok, thanks.
20:11 mmm...  so i'm going to rework how decode_auth_header interacts with HTTPAuthPassword because it's icky.  then i'll go through the various places where they get used and figure out what to do with the errors.
20:12 but first, a snack.
20:44 janetPCF1 willkg:  - it looks like I missed someone asking about the testing.  I was doing the kids school/dinner/bed thing.
20:44 but for future reference - I did start a wiki page here: http://develop.participatorycu[…]iroTestAutomation
20:45 willkg janetPCF1: no worries.
20:45 janetPCF1 the tests are actually stored on github - with the other selenium tests - seemed to be a more community friendly place
20:45 willkg oh!  i totally forgot you moved it to github.  we should probably nix the miro-testing repository, then.
20:45 janetPCF1 not yet -
20:45 willkg ok.  well, when you're ready, let me know.
20:45 janetPCF1 I using them to base the other tests off - since it's mostly the same sequence of steps
20:46 actually having that repo local, is what saved me when I was internet-less
20:46 otherwise I would have been weeding, or planting the tulips that i've negelcted for about 4 months
20:46 * willkg nods.
21:15 willkg hrm.  every time i look at the roadmap, there are more bugs.
21:16 paulproteus willkg: I know the feeling.
21:17 ajonas paulproteus: i've downgraded some, too!
21:18 paulproteus I know! You're the only thing keeping me sane. (-:
21:19 ajonas paulproteus: i'm trying to figure out if i can manage the links on the getting started dashboard now. we'll get through this!
21:28 willkg afk to shovel for a while....
21:56 * paroneayea logs into miroguide.com, puts Al Jazeera in the featured channel list ;o
22:00 bendk deardiary: fixed a bunch of downolader bugs and some more fallout from the playback changes
22:00 afk: eating dinner
22:08 zanoi paroneayea: al jazeera is awesome
22:25 maggie_s joined #miro-hackers
23:14 willkg back.
23:23 ajonas night all
23:23 ajonas left #miro-hackers

← Previous day | Index | Server Index | Channel Index | Today | Next day → | Atom Feed | Search | Google Search | Plain-Text | plain, newest first