All times shown according to UTC.

Time Nick Message
00:04 kcw willkg: I couldn't reach ben by skype, but I have "fixed" the bugs - the bugs no longer happen but it feels like I'm patching something that shouldn't be happening in the first place.
00:05 So I've assigned them to bendk to review
00:05 willkg kcw: ok.
00:16 z3p deardiary: holy cow I figured out what the issue is
00:25 willkg z3p: did you?  what happened?
00:25 z3p willkg: on webkit, we were stripping the 'Safari/532+' part of the user agent
00:25 willkg: so amazon was refusing to serve us javascript
00:25 willkg ahhh...
00:25 screwy.
00:38 ajonas joined #miro-hackers
00:38 ajonas left #miro-hackers
01:10 maggie_s left #miro-hackers
02:50 willkg left #miro-hackers
03:21 paroneayea joined #miro-hackers
03:24 uwe_ left #miro-hackers
03:26 uwe_ joined #miro-hackers
03:38 maggie_s joined #miro-hackers
04:38 kcw deardiary: Save as Podcast crashes in feed folders - #17490. I've disabled it so it won't block the release; we can fix it in the backend for 4.0.1.
04:56 *disabled it for feed folders, not in general
05:06 mattjohnson left #miro-hackers
05:22 maggie_s left #miro-hackers
06:06 kcw deardiary: found a fixed some more crashes and freezes - #17492, #17493, #17495
09:21 maggie_s joined #miro-hackers
11:30 janetPCF left #miro-hackers
11:52 maggie_s left #miro-hackers
13:19 z3p left #miro-hackers
13:19 z3p joined #miro-hackers
13:20 janetPCF joined #miro-hackers
13:21 z3p morning
13:21 janetPCF hi
13:30 bendk good morning all
13:30 busy times last night, I see
13:30 janetPCF hi bendk
13:32 willkg joined #miro-hackers
13:33 willkg hi!
13:34 bendk willkg: so should I look at the ngrams bug first?
13:36 willkg bendk: they're all blockers, so i'm not sure whether ordering them makes much of a dfference.
13:37 bendk: we're doing a dev call at 11:am.  so it's probably good to look at all three a bit to figure out what your eta might be.
13:37 bendk sure
13:37 willkg bendk: then settle on fixing one at a time.
13:40 kcw: i think d0505e21 looks ok to me.
13:41 i need coffee right now.  afk a smidge.
13:45 bendk willkg: when you get some caffeine in you, can you check out the ben branch and run that
13:45 and reproduce 17489
13:46 I added some logging that might help
13:48 zanoi hi
13:54 willkg bendk: i'll do that now.
13:56 z3p kcw: checkins looked fine
13:58 willkg bendk: took a lot longer to reproduce, but i think that's a disk caching thing.
13:58 bendk can you send me the log file?
13:58 willkg bendk: i don't see any new logging, but it's hard because of all the soft crashes.
13:58 bendk: will do that now.
13:59 bendk: i posted the 5 log files here: http://pculture.org/wguaraldi/download/
14:00 bendk: miro.log through miro.log.4.
14:00 for miro 4000, we should write a dev utility that uploads all the logs somewhere and gives you the url to give someone else.
14:03 paroneayea left #miro-hackers
14:07 paroneayea joined #miro-hackers
14:24 willkg bendk: did the logs yelp?
14:25 er, help.
14:25 they're logs, not dogs.
14:25 janetPCF they were screaming
14:25 bendk still working on other ones
14:25 willkg ok.  no problem.
14:26 paroneayea left #miro-hackers
14:31 willkg janetPCF: if these logs were dogs, they'd be yelping something awful.
14:31 janetPCF: there are like 20 soft failures in there.  i had to kill the process.
14:32 janetPCF ugly
14:32 if you have a minute to push out some new nightlies, I can retest those last 3 fixes from Kaz this morning
14:36 paroneayea joined #miro-hackers
14:41 willkg janetPCF: can do.
14:44 janetPCF: new mac nightly posted.
14:49 janetPCF: new windows nightly posted.
14:50 janetPCF great thanks
14:51 fyi - did a 1st time install on a fresh fresh Windows 7 today
14:51 willkg and...  ?
14:51 janetPCF then just ran some basic tests
14:51 it was good.
14:51 I can see getting bonjour and adobe pop-ups that fast can detract from the experience.\
14:52 but probably most people don't install Miro as their 1st application, and likely already have flash by the time they get to miro - so it's just the bonjour popup
14:53 running the helpscripts on Natty now
14:53 so I can check that out a bit later
14:54 * willkg nods.
14:54 willkg i didn't do a source tarball nightly.  i can do that now if you want.
14:55 janetPCF I can wait - I really can't test all 3 at the same time.
14:55 2 yes.
14:55 since human have bilateral symmetry
14:56 bendk kcw: can you check out 0b22128
14:56 tell me if the code seems good and also if you can find a way to cause a crash with it
14:58 willkg: so for #17489
14:58 where are those items located in miro?
14:58 in a device tab?
14:58 willkg bendk: i'm clicking on hte Music tab of the device in question.
14:58 bendk ahh
14:59 and if you don't click on that, then the bugs don't happen?
14:59 willkg bah.  sorry about that.
14:59 bendk no problem
14:59 willkg i have no idea.  i never tested the "don't click on the Music tab" case.
14:59 i can do that now.
14:59 bendk please
14:59 I'm going to get some food/coffee before the call
14:59 willkg i've reported this sort of issue a bunch of times.  i think this is the first time i didn't include steps to reproduce.  probably bcause i was going to look into it.
15:01 bendk: looks like if i don't click on either the Video or Music tabs of the device, then things are fine.
15:02 so, what i did do was start a clean instance of Miro, then click on the Connect tab, then enable "Show all attached devices and drives" and then let it spin for a while.
15:02 and that seems not to kick up the dialog.
15:08 bendk z3p: #17489 might be a device issue
15:09 can you think of a way that DeviceItemSource could be sending items as changed, before they are added?
15:10 willkg bendk: z3p was thinking that setting file_type to None might do icky things for device stuff.  (just repeating what i think he said.)
15:10 z3p willkg: yes, that is what I said
15:10 bendk: let me look through the code again
15:10 willkg as a side note, kcw figured out why the crash dialog wasn't ignorring things.  fixing that makes it _so_ much less asstastic.
15:10 bendk that's very nice
15:10 willkg hrm.  xchat doesn't think asstastic is a word.
15:12 z3p bendk: so there's code in DeviceItemSource._on_device_changed() to figure out if an item has changed its filetype
15:15 kcw could calling signal_change in DeviceItem's setup_new be the problem?
15:15 we recently changed update_skipped from as_idle to running in the current thread, so for files that MDP would skip we signal_change during setup_new in request_update
15:17 z3p bendk: oh, maybe!
15:17 bendk: signal_changed() calls db.emit('item-changed')
15:20 bendk: willkg: I can't reproduce the issue, but does switching the last line of DeviceItem.signal_change() to 'if self.file_type not in ('other', None) or was_removed' help?
15:20 willkg z3p: i can test that now.
15:20 z3p: what file is that in?
15:20 z3p willkg: great
15:20 willkg oh, found it.  it's in devices.py.
15:21 z3p willkg: tv/lib/devices.py:897
15:25 willkg z3p: that doesn't seem to fix it for me.  i still hit the same issue.
15:25 z3p: also, it looks worse.  i'm seeing duplicates in the Music tab now.
15:30 z3p willkg: okay, I'll try the other fix
15:30 willkg z3p: sorry dude.  :(
15:31 z3p willkg: it's just frustrating for me that I'm not seeing this issue
15:31 willkg yeah, that's werid.  this is on my massive external drive with .ogg files and a gazillion other files.
15:31 so maybe it's so big that the timings work out to create mayhem or something.
15:31 z3p willkg: I have a flash drive with 20k ogg files, and it's not having this issue
15:32 bendk z3p: what if we just put some flag in DeviceItem that says if you're in the constructor or not and then look at it in signal_change()
15:32 z3p bendk: that's what I'm doing right now
15:32 bendk cool
15:33 z3p willkg: http://pastebin.ca/2065796
15:35 willkg testing that now....
15:36 same crash.
15:37 z3p: ^^^
15:37 z3p willkg: hrm
15:37 willkg i need to refill my coffee.  afk for 2 minutes.
15:41 back.
15:41 bah.  4 minutes.  that blows.
15:41 i should have doubled my estimate.
15:44 bendk kcw: I'm not getting the crash you talked about in #17483
15:44 can you give me exact steps?
15:44 z3p: so can I hand this bug over to you?
15:44 I think you're definitely the one to fix it
15:46 z3p bendk: yeah, assign it to me
15:47 janetPCF kcw:  so I put my DRM files in a directory and added it as a watched folder (on windows)
15:47 No DRM badge, and when I tried to play the file Miro crashed out, same as before
15:49 kcw:  I can make a new ticket or just reopen the original ticket
15:49 kcw janetPCF: hopefully that's a result of my changed mutagen DRM tag check, and not mutagen's check succeeding and MDP successfully process the file. if it's the latter that's a big problem. I'll check that out in a minute - looking at the filter issues now
15:49 janetPCF ok - do you have the data file?
15:49 kcw I haven't looked yet, but it's probably there somewhere
15:52 janetPCF 10 Thinkers.m4v, sent to you in an email 17 Feb. subject: drm test file.
15:53 z3p willkg: ah ha! I can reproduce it now
15:53 willkg yay!
15:53 z3p willkg: the problem was that all of my .ogg files are little sounds, not media files, so mutagen shortcircuits
15:53 ajonas joined #miro-hackers
15:53 willkg ahhh...
15:53 interesting.
15:53 z3p I wasn't hitting the code where file_type gets set to None
15:54 willkg i think mine are from ripped albums.
15:54 like, white zombie.
16:08 z3p see, I thought I had the issue fixed
16:08 but now MIro is segfaulting
16:08 aude joined #miro-hackers
16:25 z3p willkg: 80c82ba
16:26 willkg i'll test that.
16:27 z3p I'll rip a CD to ogg an throw it in with my other test stuff
16:27 janetPCF left #miro-hackers
16:28 willkg looking good so far....
16:28 i'll let it finish the import.  afk to refill my coffee.
16:28 gah!  it's noon!
16:28 :(
16:29 zanoi I ran out of coffee in my apartment!
16:31 willkg zanoi: that sucks.
16:32 z3p: it's still looking good.  it's not quite done the import, but it would have crashed a bunch of times by now.
16:32 z3p: so i'm +1 on that fix.
16:32 z3p willkg: my import is still going too, but I think it would have broken by now
16:33 willkg kcw, bendk: so there are two new blockers.  do those need to block an rc2?
16:33 bendk #17500 should
16:33 I'm almost done with it though
16:33 17504 too
16:33 both of them should be pretty quick though
16:34 willkg ok.
16:34 * willkg plays his white zombie.
16:42 aude left #miro-hackers
16:43 bendk zanoi: you around?
16:43 janetPCF joined #miro-hackers
16:51 janetPCF1 joined #miro-hackers
16:52 janetPCF1 left #miro-hackers
16:53 janetPCF left #miro-hackers
16:54 arpu joined #miro-hackers
16:56 bendk zanoi: kcw: can you check out 916a875
16:56 zanoi bendk, sure
17:01 bendk, this solves the crash, but it causes some weird shuffle behavior
17:01 bendk, but fixing the crash is the important part of course :)
17:01 bendk what changes?
17:02 I was hoping to keep the same behavior and just fix the crash
17:02 kcw bendk/zanoi: could you look at 0a5ed28/#17505?
17:03 zanoi bendk, if i understand it correctly, right now if you sort out lots of items, turn on shuffle, and then turn off all filters shuffle will still work on your old filtered playlist
17:05 bendk kcw: sure 1 sec
17:05 kcw isn't that what it did before? generate_upcoming_shuffle_items uses get_all_playable_items, which gets its items from the model
17:05 and I don't see any way that it was calculating a new list for filter changes before
17:06 zanoi kcw,  0a5ed2 looks great
17:07 bendk kcw: can you check out a0ad26b?
17:07 z3p time for more coffee and some breakfast; back in a few
17:07 willkg gah.  we fixed like 154 bugs since rc1.
17:07 these release notes will never be human-readable.
17:08 zanoi kcw, bendk, yes it's not so much that 916a875 causes the issue it's that not all problems of the bugs are fixed with it, but at least the crash is so that might be good enough for now
17:08 bendk kcw: changeset looks good to me
17:08 zanoi: can you explain more, what is happening wrong exactly?
17:09 zanoi: nevermind, I get the bug
17:10 yeah, let's just leave that for 4.0, since it doesn't block
17:10 file a ticket for 4.0.1
17:10 zanoi bendk, will do
17:10 bendk For all of these bugs, the root cause is #17498
17:10 I wanted to change that code, but I didn't realize that it would have such far-reaching impacts
17:13 okay, taking a break
17:15 willkg hrm.  janet's not on.  boo.
17:16 anyhow, i'm rolling a new set of nightly builds for windows and osx.
17:20 new mac build posted with everything up to now.
17:22 janetPCF joined #miro-hackers
17:23 arpu left #miro-hackers
17:23 kcw the new drm problem is bad... since mutagen can't reliably identify drm, the current approach is to verify it with mdp - but apparently, mdp on windows can succesfully examine drm'd files
17:24 so neither mutagen or mdp can tell that the file has drm; we don't know until we try to play it and crash
17:25 mattjohnson joined #miro-hackers
17:26 kcw janetPCF: does your thinkers test case show anything in the duration field on windows?
17:27 janetPCF it gets the correct duration for the files
17:27 can you be more strict with video files and more lenient with the music files?
17:28 willkg janetPCF: i just pushed new nightly builds for windows, mac and source.
17:29 janetPCF ok - this has everything bug the drm, now?
17:29 but
17:29 bugs on the brain.
17:30 kcw janetPCF: could you attach to #17442 the log you get when importing Thinkers on windows
17:30 willkg janetPCF: i think kcw is working on the drm issue.
17:30 janetPCF sure
17:30 kaz
17:30 if you want to be able to see some of this for yourself, I can also give you my itunes account info
17:31 I don't have any private info attached (cards…) since I"m not it the US anymore
17:31 and i have about 10 of the animal planet downlaods and that's about all
17:31 oh - maybe pirates of the carribean
17:32 kcw thanks, but I think all videos downloaded recently would have the same drm characteristics, so I think the individual test case should be fine
17:36 janetPCF added them
17:36 gluttons and jumpers
17:36 these are files that I probably downloaded about 4 yrs ago
17:37 kcw oh, it's possible that itunes has changed their drm tagging since then
17:37 so those two cases also are drm'd but don't show as drm'd?
17:37 janetPCF yes
17:38 kcw this doesn't look as bad as I'd thought. mdp is successfully failing for those files.
17:38 arpu joined #miro-hackers
17:41 willkg yay!
17:43 z3p is anyone else seeing segfaults?
17:44 willkg z3p: i'm not using it right now, but i haven't had it segfault in a long while.  i think zanoi saw a few earlier this week.
17:45 zanoi i saw one segfault in the past weeks i think
17:45 z3p I'll run it with gdb and see what happens
17:45 zanoi but then i immediately started miro again it worked fine
17:54 willkg afk to get breakfast or whatever meal it is now.
18:02 zanoi afk while I go hunting for coffee
18:08 z3p deardiary: I think it's funny that I got a segfault for g_object_ref while watching a video about memory use
18:12 willkg z3p: that is funny.
18:12 z3p: i see a bunch of gobject-related warnings in the logs.  do you see those, too, on maverick?
18:13 kcw 17243 is for the G_IS_OBJECT warnings
18:13 willkg z3p: ^^^
18:13 i pushed that bug off because i think we were thinking that they were just warnings and not real issues.
18:13 but, figured i'd mention it in case it's a real issue.
18:14 kcw: how're you doing with that drm issue?
18:14 z3p kcw: willkg: the last thing I see is Warning: g_object_ref: assertion `object->ref_count > 0' failed
18:16 kcw I can't see any way that this mdp log-line could possibly correspond to an item with drm-ish mutagen data that gets changed to not-drm'd - the log says duration is None, and the only place we're unsetting has_drm is when duration is not None... and my windows vm is borked right now. The mdp output looks essentially the same as on Linux, and the rest *should be* the same cross-platform.
18:17 janetPCF: could you attach the output of "mutagen-inspect filename_of_drmd_file" on the drm bug? maybe we're getting different mutagen outputs somehow
18:18 glee left #miro-hackers
18:18 janetPCF so you want me to run on linux, mutagen....
18:18 just give me a minute, ok?
18:18 kcw on windows
18:18 glee joined #miro-hackers
18:18 janetPCF ah
18:18 ok
18:20 glee left #miro-hackers
18:20 janetPCF kcw:  where would I find mutagen-inspect…..
18:22 kcw janetPCF: oh, I don't know. does the cygwin shell not find it?
18:22 willkg kcw: she doesn't have a dev environment.  she's got the windows nightlies.  pretty sure we don't include it in the nightlies.
18:23 kcw ah
18:23 willkg janetPCF: the files we're fiddling with are in the test-data drop box thingy, right?
18:24 janetPCF yes
18:24 can I just download mutagen?
18:24 willkg ok.  i've got a windows dev environment.  i'll see if i can do this.
18:29 i installed drop box on my windows machine...  waiting for it to sync.
18:30 so...  probably a few more minutes.  then i can do the mutagen-inspect thing.
18:32 CarlFK left #miro-hackers
18:44 arpu left #miro-hackers
18:44 willkg kcw: where can i find mutagen-inspect?  it's not in the binary kit as far as i can see.
18:46 kcw willkg: I don't know - it came with my mutagen package. I'd thought it would also come with mutagen for windows
18:46 willkg kcw: what package are you talking about?
18:47 kcw the gentoo mutagen package - it's probably in the tarball for linux
18:47 willkg kcw: i'm confused...  you wanted janetPCF to run this on windows, right?
18:48 kcw yes. it's possible that the mutagen-inspect tool is not included with the windows mutagen package - I only know that it exists on linux
18:48 willkg well, there's no "windows package" on their downloads page.  i'll grab a tarball and see what's there.
18:49 kcw it's available separately here - it's just a short python script: http://code.google.com/p/mutag[…]agen-inspect?r=98
18:51 willkg ok.  so i'm not entirely sure what we're testing here.  if we're testing with the mutagen that miro is shipping with, then clearly this test doesn't really help much.
18:52 http://pastebin.ca/2065941
18:53 janetPCF i pasted my output into the bug kcw
18:53 DGMurdockIII joined #miro-hackers
18:53 willkg janetPCF: how'd you get mutagen-inspect?  i had to do silly things to get it.
18:54 janetPCF I can do silly things too, that's what google is for
18:54 willkg fair enough.
18:54 janetPCF our output is different
18:55 willkg we did two different files.  i did the thinkers one and you did jumpers.
18:55 i'm not sure why your output doesn't have the xml stuff.
18:55 er, the plist stuff.
18:55 kcw that's the problem though
18:55 janetPCF right  - different seasons? different drm rules
18:56 it's funny how much my kids love that show - it's probably the only thing we miss from America
18:56 kcw they don't both have the asset-info stuff, but they both do have apID. apID should probably be the new drm indicator
19:03 janetPCF zanoi:  a blocker bug with no description beyond 'totally broken' is totally uncool at this point.
19:10 willkg janetPCF: i agree--that bug uber-sucks.
19:10 janetPCF: but i think what he's getting at is that on linux, when the interface is in german, it looks like it has no labels--there's no text in the dialog at all.
19:11 i'll go fix the bug.
19:12 kcw willkg, janetPCF: pushed a fix for the false negatives in drm detection in 766a8e1; haven't tested it with the problem test cases yet - syncing dropbox now
19:13 willkg kcw: awesome!
19:13 i'm looking into bug 17507.
19:23 who wrote the item edit dialog?
19:25 kcw willkg: I did
19:25 willkg amongst other things, there's a gross abuse of attribute lookup here.  using class attributes, but referring to them with instance attribute lookups.  total ick.
19:25 so...  this dialog doesn't seem to show any labels if the language isn't english.
19:26 hrm.  let me correct that.  this dialog doesn't seem to show any labels if the labels aren't english strings.
19:26 so, if miro is set to use Afrikaans (which is predominantly untranslated and thus english), it works ok.
19:26 but french and german definitely don't work.
19:27 i'm going to test this on osx and windows, though janetPCF says it works fine on osx.
19:30 it's not working on windows, too.  same symptoms as linux.
19:31 it works on osx, but the labels all look terrible because there's not enough space.
19:34 janetPCF: i can reproduce the issue on linux and windows, but not on osx.
19:35 janetPCF: i think this sucks--we should fix it.  but i think i'm going to push out rc2 anyways.
19:35 janetPCF: do you want to test kcw's drm fixes before i tag rc2?  if so, i can roll new nightlies.
19:37 janetPCF1 joined #miro-hackers
19:39 kcw so in itemedit, I think the labels are probably not drawing at all when they don't have enough space on gtk, as opposed to just being cut off on OS X
19:39 janetPCF1 left #miro-hackers
19:39 kcw so it would probably help a lot if I just force the space to be wider
19:40 willkg kcw: so, if you tweak some of the label text so that it definitely can't translate, then that won't show up either.
19:41 kcw willkg: it's not showing text that doesn't have a translation, where it should fall back to english?
19:41 janetPCF left #miro-hackers
19:41 willkg for example, if you change _("Rating") to _("Rting"), then it still doesn't show up.
19:42 janetPCF1 joined #miro-hackers
19:42 willkg the gettext side of things is working fine.
19:43 plus some of the fields are missing, too.
19:43 kcw so it isn't just a layout issue. I don't see how I'm using gettext any differently from other dialogs - it's all static strings; I'm importing gettext from gtcache as _...
19:43 willkg it kind of looks like it's throwing an exception somewhere that's not getting handled and the dialog isn't being rendered.
19:44 janetPCF1 willkg:  how about windows nightly for the drm test
19:44 willkg kcw: are you able to reproduce the issue?
19:44 janetPCF1: i'll do a new windows nightly right now.
19:45 janetPCF1 ok great
19:45 I have a whole collection of animal planets -
19:45 love them
19:46 kcw oh yeah, I see it. completely broken's actually a great description.
19:46 willkg ok.
19:46 can i pass this to you to tinker with and figure out what the cause is?
19:46 er, can you take this and tinker with it to figure out what's going on?
19:47 kcw yeah, I'm looking at it
19:47 willkg ok.  thanks!
19:48 Odysimus left #miro-hackers
19:52 aude joined #miro-hackers
20:05 willkg janetPCF1: bah.  i'm super sorry.  the windows nightly is posted.
20:06 janetPCF1 ok - just going to ask
20:06 willkg janetPCF1: it's been posted for about 15 minutes now.
20:06 janetPCF1 that's ok - I was putting kids to be, and thinking how sad it is that they have had tummy troubles the whole week and I have 3/4 of a great cake in the fridge going to waste
20:07 kcw huh
20:07 the rating dropdown breaks everything
20:07 janetPCF1 the rating??? I was about to put money the screen toggle
20:07 willkg kcw: that's fantastic news.
20:09 kcw: the RatingField dropdown?  if i comment that out, i still see the problem.
20:10 ahh...  but if i comment out the entire _pack_right call, then it works much better.
20:10 ok.  fun fun fun.
20:10 kcw are you sure? if I comment out right.append(RatingField ... it works for me, with the French version
20:10 willkg janetPCF1: can you test the drm stuff?
20:10 kcw: doesn't with german.  german is the only language that's completely translated.
20:11 kcw: so...  maybe that's the difference?
20:11 i'll double-check my findings, though.
20:11 kcw ah - so it could be a problem with the translated dropdowns in general
20:11 there are no dropdowns on the left
20:11 french has the file_type values all translated though
20:13 willkg yeah, i'm definitely seeing what i'm seeing.  if i comment out the the self._pack_right(middle) line, then the dialog renders.  if i comment out just the RatingField line, then it doesn't.
20:13 paroneayea left #miro-hackers
20:13 paroneayea joined #miro-hackers
20:14 paroneayea left #miro-hackers
20:15 paroneayea joined #miro-hackers
20:16 willkg kcw: and i see what you're seeing, too.  if i use french and just comment out the RatingsField line, then the dialog renders.
20:16 kcw so it's the rating field in french, and both rating and the thumbnail in german
20:17 willkg funky.  i agree.
20:18 janetPCF1 I was just looking at the german on os x -
20:18 and the world for thumbnail in german is really long
20:19 and actually makes the os x dialog look a but funky
20:19 willkg janetPCF1: not to be an ass, but how's the drm testing going?
20:19 janetPCF1 downloading the build - take me 10 mins to download
20:19 willkg janetPCF1: got it.  ok.  whew.
20:20 kcw oh yeah, when I change the label text for Rating and Art so that neither will translate, that also makes German work
20:21 and it's broken in english if I change one of those labels to something really long!
20:21 willkg so the totally "awesome" fix for rc2 is to tweak the strings so there aren't any valid translations for them.
20:22 i vote we remove all the e's.
20:22 we can say that we did it to save bits.
20:22 janetPCF1 Classement in frech for rating
20:22 classmnt
20:23 willkg also, i'm half joking and compltly srious about this.
20:23 janetPCF1 how about no vowels - a bit slavic
20:23 clssmnt
20:23 willkg there was an Onion article about how Clinton was going to send Serbia vowels or something like that and how it would help their economy.
20:23 kcw long dropdown values seem to be fine
20:23 so I think it's just a problem where we're using gtk.Labels
20:24 willkg kcw: do those labels have a fixed width?
20:24 bendk: you on?
20:24 bendk I'm here
20:24 willkg are you following this conversation?
20:24 bendk I'm reading now
20:24 willkg i just want a third set of eyes on it.  particularly regarding sizing issues.
20:26 kcw I don't think there's any way to give the Label class a pixel-based width, so I think the way I did layout is not allowing the fields as much room to render as they request... and then somehow that makes nothing render anywhere
20:27 willkg umm...  can't you call set_size_request on it?
20:27 i'm pretty sure i do that in a lot of other places.
20:27 kcw probably, but I haven't
20:27 bendk so are we giving the dialog a fixed width?
20:27 what if you increase that size?
20:28 kcw ooh I think I see it
20:28 willkg bendk: i'm not sure that fixes it unless we know the maximum size for all the languages.
20:28 bendk I agree, I'm just trying to diagnose exactly what's happening
20:28 * willkg nods.
20:29 kcw so I have a "TODO: calculate column widths automatically" that I've been planning to get to after 4.0...
20:30 what's happening is when a label is wider than LABEL_WIDTH, my size allocation formula is really stupid and tries to give things negative widths
20:30 *negative paddings
20:30 to make them fit
20:31 so "pad = max(pad, 0)" fixes this
20:34 so ba30600 clips padding to 0 - on Linux in English for long fields that fixes the problem, allowing fields to be as wide as necessary
20:34 I'll test other languages and then OS X
20:35 willkg ahh...  ok.
20:35 kcw so now when a field is wider than the column expects, it cuts into the field's input box on GTK
20:35 which doesn't look great, but it's totally usable
20:35 we should probably also increase LABEL_WIDTH a bit to accomodate wider languages better
20:36 willkg awesome.  that fixes it for me with german.
20:36 kcw: thank you!
20:36 i still like my "let's ditch the e's" "fix".
20:37 janetPCF1 willkg:  kcw - so the drm fix looks good on windows
20:37 willkg yay!
20:37 janetPCF1 my whole animal planet collection is flagged with DRM - but npr's music file is not and plays fine
20:37 willkg yay!
20:38 ok.  so...  i verified kcw's fix.
20:38 janetPCF1: is there anything else we want to test or can i tag rc2 now?
20:39 janetPCF1 I don't think there is anything else outstanding
20:39 willkg ok.
20:39 i'm going to double-check kcw's fix on osx.  just so i don't feel like a total jackass later if it turns out to break things for some karmic reason.
20:39 janetPCF1 man - don't you just want to reminisce  about how nice it was when we had 3.5 done and sat on it for 3 weeks waiting for Zugo
20:40 willkg no.  not really.
20:40 i had forgotten about them.
20:40 janetPCF1 well good thing we didn't have to make any of those changes for 4.0
20:41 willkg looks good on osx and windows.  w00t.
20:41 yeah, no kidding.
20:41 ok.  i'm tagging rc2.
20:45 kcw the Video pane of itemedit is cut off in German
20:45 the OK button is mostly missing
20:45 willkg^
20:45 willkg oh well.  write up a bug for it and we can fix it for rc3 or final.
20:45 kcw on OS X, that is
20:48 note to self: after implementing a new UI component, TEST THE LOCALIZATIONS
20:49 willkg well, so that's the thing...  you might not have seen anything until recently because translations are still bad-dish.
20:49 i think we need to have one non-english translation that we maintain so that we have something to test with that we know works.
20:53 Odysimus joined #miro-hackers
20:55 kcw hm, yeah. this wouldn't have come up until Art and Rating were translated. what would have caught this is if I had tested it with a faux localization: "_ = lambda t: 'long ' * 5 + 'string'"
20:55 willkg oh...  we could totally do that.
20:55 bah.  i should have thought of that.  i just read about it a few weeks ago.
20:56 it'd be so easy to build a faux translation when i sync strings.
20:56 i'll toss that in the todo list for 4.1.
20:57 it might have been ned batchelder.  i forget whose blog it was.  whoever it was suggested either removing vowels or adding vowels.
20:57 though i think it was to verify that everything was translated.  it wasn't to verify that things worked with strings of larger size.  we could just do a transform that increases the string sizes.
20:58 janetPCF1 it was a very good catch by zanoi, even if the bug sucked
20:59 zanoi janetPCF1, sorry, I forgot to attach the screenshot I made :(
20:59 janetPCF1 haha - that's funny
20:59 zanoi and then went away to buy coffee
21:00 I blame my too low caffeine intake
21:00 janetPCF1 ok - so willkg, can I assume about 1 hr before the rc2 builds are done?
21:01 willkg kcw: 17509.  i put it in the 4.1 target.
21:02 janetPCF1: i have source and osx builds done.  doing windows builds now.
21:02 aude left #miro-hackers
21:02 willkg janetPCF1: do you want to check them out?  they're at http://pculture.org/wguaraldi/download/ .
21:02 janetPCF1 sure
21:02 you are fast
21:02 willkg janetPCF1: the windows ones are uploading, so they may show up, but they're not there yet.
21:03 janetPCF1 got it
21:04 willkg if this conversation lasts another 50 seconds, then we'll have to start over because everything will be different.
21:04 janetPCF1 yeah
21:04 I don't want this whole day to start over again from yesterday
21:05 so I think we need to make sure all the parts and pieces are together, and send these build forth into the world and see if they can hold their own against the masses
21:05 * willkg nods.
21:05 willkg plus keep looking for issues ourselves and do an rc3 tomorrow if we need to.
21:05 ok.  windows builds posted, too.
21:05 janetPCF1 that's a very fancy downloads page
21:06 willkg janetPCF1: give me a wink when you feel good about them.  then i'll move them to the nightlies page, remove the rc1 builds, do the .htaccess thing, blog, dent, and then take a short break.
21:06 janetPCF1: it's a php script some guy wrote in like 2001.
21:06 i like it because it allows me to add descriptions.
21:07 janetPCF1 something like that on nightlies would be cool - to see the time stamp when we are doing several builds a day
21:08 willkg janetPCF1: totally.  some day i'm going to re-re-rewrite the nighlites page into something less asstastic.
21:09 also, some day i'm going to rewrite bogon deflector, the downloader, the windows platform code, the unit test system, the logging code, some of the wiki pages, ...
21:09 janetPCF1 OK - so I have to wait a bit for all the builds to download, and go figure out which instance of miro is using all my bandwidth.
21:09 willkg ok.
21:10 janetPCF1 then enjoy the last 57 minutes before for the world ends and I fail rapture class
21:10 willkg oh yeah!  rapture is tomorrow!
21:10 we'll probably have fewer testers, then.
21:11 janetPCF1 and we won't be able to trust them because all the virtuous ones will be gone
21:11 crap
21:11 willkg oh no!
21:11 i never even thought of that!  we're so scr3w3d!
21:11 hrm.  i want to write the faux language translations right now.
21:12 janetPCF1 if you do, you'll notice that the on-off button on the Connect page is less than attractive
21:13 holy crap - time zones
21:13 willkg la la la la la i can't hear you there are no issues la la la la la!
21:13 janetPCF1 do you think it already started????
21:13 has anyone heard from glee or is he gone already?
21:13 willkg he's not on irc.  i have no idea.
21:13 i'm totally doing faux language stuff right now.
21:14 janetPCF1 so what's your plan?
21:14 maybe add 'shit' after every 'o'
21:14 willkg heh.  that's funny, too.
21:15 janetPCF1 to be true to yourself, you should add 'sstastic' after every 'a'
21:17 willkg heh.
21:37 ajonas left #miro-hackers
21:51 z3p how are we looking?
21:52 I see a re-tag of rc2
21:52 so that's good
21:54 willkg z3p: it's going ok.  i re-tagged rc2 and janetPCF1 is testing the builds now.
21:59 paroneayea left #miro-hackers
22:06 paroneayea joined #miro-hackers
22:18 willkg wow.  miro looks totally awesome in swedish chef.
22:19 plus i found a bug in gtcache.gettext.  boo.
22:19 and i found some other bugs, too.
22:27 janetPCF1: i implemented swedish chef.  it's swch.po.  i think i need to tweak iso_639.py so it shows up properly, but if i don't, it just shows up at the bottom of the languages list.
22:27 janetPCF1: i think i'm going to leave it at the bottom.
22:28 i just got too excited to implement it.  so i did.  now it's a new feature in 4.0: the swedish chef ui.
22:28 on that note, going to go get dinner.
22:30 z3p a friend of mine's last show is happening, so I'm going to see that; I'll be on e-mail in case anything terrible happens
22:44 willkg z3p: ok!
22:45 janetPCF1: how're things going?
22:48 janetPCF1 ok - testing on XP - update from 3.5.1
22:50 I just got a crash (one of those stupid key errors) searching in Videos tab
22:51 willkg janetPCF1: write up a bug report and put it in the 4.0/blocker list.  i want to push rc2 out unless we hit something catastrophic.
22:51 janetPCF1: but that'll be good to fix for rc3.
22:51 janetPCF1 I think there already is a bug for it
22:51 I know there is
22:52 willkg it's not in the P1 4.0 list.
22:52 kcw: are you on?  are you done with bugs 17442 and 17507?  can we mark them as FIXED now?
22:53 kcw willkg: yeah, I was actually just doing tha
22:53 willkg kcw: ok, cool.  thanks!
22:53 janetPCF1 bug #17436 - it's the p1 that's still in the list
22:54 willkg janetPCF1: that's "exception during file import".  are you importing files?  i thought you said you got it while searching in the videos tab.
22:54 janetPCF1 you get the same result for 2 different actions
22:54 willkg ahh...  funky.
22:54 ok.
22:54 janetPCF1 I added the same crash report before
22:55 zanoi yay, finally I can have miro in a language I can understand!
22:55 janetPCF1 anyway - I running the quicktest+, but in Croatian on XP after an upgrade.
22:55 willkg zanoi: are you talking about swedish chef?
22:55 janetPCF1 trying to see how many different things I can cover at once
22:55 zanoi willkg, yes, it's amazing
22:55 willkg zanoi: that's hysterical.  :)
22:56 zanoi I can't stop laughing
22:56 kcw finally something supports my LC_ALL setting
22:56 willkg i kind of like the transform.  it's readable (if you sound it out), but it's also a transform for which translated strings are longer than their original ones.
22:57 zanoi willkg, yup
22:59 willkg, I think we should release miro with swedish chef as the default language and lots of swedish chef videos included
23:00 maybe we should at some point figure out what the most verbose language is and then test with that before a release
23:01 although I'm pretty sure German is a good contender
23:01 also, german is still way longer than german chef
23:01 errr
23:01 swedish chef!
23:02 willkg heh.  :)
23:02 yes, german strings look longer than swedish chef strings.  i can probably fix that if we want.
23:04 zanoi well, it's fine for now I think
23:12 is there a name for the menu bar that is at the top of each window?
23:13 willkg hrm.  there's already a tool that does the same kind of thing as what i wrote: http://translate.sourceforge.n[…]i/toolkit/podebug
23:14 zanoi but do they support swedish chef?
23:15 willkg yes.  there's a chef output.
23:15 they also do some other outputs that are interesting, too.
23:21 maggie_s joined #miro-hackers
23:24 zanoi willkg, how's RC2 looking?
23:24 willkg zanoi: janetPCF1 is testing.
23:25 zanoi uh-oh
23:25 janetPCF1 what?
23:25 zanoi please don't find any more bugs?
23:25 willkg zanoi: that bug you just created about menus in miro in german....  that's a unity problem.
23:25 zanoi: there's nothing we can do short of nixing menus.
23:25 janetPCF1 I'm just making sure all the parts and pieces are reasonably assembled
23:26 oh - i had my first look at Natty today
23:26 willkg zanoi: it works fine if you're not using unity.
23:26 janetPCF1 is unity what makes the menus appear and disappear?
23:26 it's so annoying
23:26 zanoi willkg, ok, I'll look if there is a unity bug for it when I have more time
23:27 willkg, I could also make the german translation shorter as a workaround
23:27 janetPCF1, yep, that's unity :)
23:28 willkg i wrote up a comment for the bug.
23:28 janetPCF1 you know - I went to a liberal arts college, but it also had decent science and engineering.
23:28 and my last year or two there, they decided to dumb down the science requirement for all the english majors.
23:28 the result was:
23:28 Chemisty of Things.
23:29 willkg nice!
23:29 janetPCF1 Poetry of Physics
23:29 and I can't remember what they did to Geology - but...
23:29 it's like that's what they did to Ubuntu
23:29 like why why why does it take 5 mins to locate the TERMINAL!
23:30 zanoi janetPCF1,  hehehe
23:30 janetPCF1 yeah ha ha ha
23:30 Math for Morans
23:35 zanoi sounds like a great college!
23:35 is there any way to browse translations by file in launchpad?
23:35 DGMurdockIII left #miro-hackers
23:42 janetPCF1 !@#!!
23:42 zanoi odd, I can't find the string I'm trying to fix in launchpad
23:42 oh no!
23:42 janetPCF1 In German on osx
23:43 zanoi what? :(
23:43 janetPCF1 the top level tabs for Sources and Playlists are missing their pretty content
23:43 * zanoi sighs
23:43 janetPCF1 so it's just Playlists in ugly red
23:43 when I switch back to English it's fine
23:43 zanoi huh, I don't understand
23:43 kcw oh no
23:45 janetPCF1 the top level tab - Sources and Playlists
23:45 no pretty styling
23:46 Croatian is fine
23:46 (but it's all in English)
23:46 zanoi odd
23:47 janetPCF1 french is ok
23:47 kcw in displays we're using translated strings to identify display types!
23:47 janetPCF1 oh - bizarre
23:47 now German is OK too
23:47 maybe I'm tired and was hallucinating those ugly red letters.
23:48 but I think no
23:48 I did a first time startup, and chose German in the install dialog
23:49 kcw I don't see why it would inconsistently cause problems, but identifying displays by their translated names seems very wrong and is probably the reason for that
23:49 zanoi willkg, when was the last time you synced translations?
23:50 janetPCF1 6 hrs ago
23:51 zanoi damn :(
23:52 didn't see that
23:52 I think I'll give up on German translations for now then
23:54 willkg what's up?
23:55 zanoi nevermind
23:55 kcw same problem with a fresh install in Swedish Chef on Linux - no displays for any of the root nodes or Connect
23:55 but then it's fine upon reopening
23:56 willkg mmm...  looks like that's because the strings are getting translated at import and not when the app gets launched.
23:58 so then when the user switches languages in the first time dialog, those bits of the app are using the wrong language and then things don't jive.
23:59 mmm....  kcw can you write up a bug for that?
23:59 kcw sure

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