[00:27] DBO: Pong [00:28] so uhhh [00:28] sqlite3 + banshee + initial music import == massive fail [00:28] I am told you know something about this [00:28] Insomuch as I've touched Banshee code in the past, I guess. [00:29] What part is massive fail? [00:29] the sqlite queries tend towards taking infinite time [00:29] so at first they are fast [00:29] but pretty soon they take 5 seconds a pop [00:30] until they are taking 30/40 seconds a pop [00:30] at which point I usually give up and stop trying to see how bad it will get [00:30] missing index perhaps [00:30] that or contention [00:31] well I am testing this by deleting the banshee db [00:31] so I doubt an index issue (seeing as I seem to be the only one effected) [00:33] Hm. I'll delete my library and give it a whirl. [00:33] Actually, first I'll relocate so I can bask in the verandah sun. [00:34] basking [00:34] fabio [00:46] DBO: How big is your library, exactly? Mine's 2K to the nearest order of magnitude, and while I noticed that the import paused a couple of times, it completed in not much more than IOwait. [00:47] 11k [00:47] RAOF, I get through about 3k before any slowdown is evident [00:47] Yeah, that'd be right. [00:47] I only ever get bugs that I can't reproduce :P [00:48] i literally let it run 8 hours [00:48] it never finished [00:48] Meep! [00:48] its like every 2 minutes I get half way there [00:48] its frustrating as hell [00:48] "okay only 20 more..." 4 hours later... "only 10 more" [00:49] 100% CPU the whole time [00:49] Oh, so the problem you see isn't periodic pauses, but the actual per-file import time increasing? [00:50] more precisely [00:50] its when it seems do the recalculation of the smart playlists [00:50] those SQL query calls take longer and longer each time they happen [00:52] RAOF, this one in particular: [2 Debug 17:02:38.432] Executed in 6177ms SELECT COUNT(*), SUM(CoreTracks.FileSize), SUM(CoreTracks.Duration) FROM CoreTracks CROSS JOIN CoreArtists,CoreAlbums, CoreSmartPlaylistEntries WHERE CoreArtists.ArtistID = CoreTracks.ArtistID AND CoreAlbums.AlbumID = CoreTracks.AlbumID AND CoreSmartPlaylistEntries.TrackID = CoreTracks.TrackID AND CoreSmartPlaylistEntries.SmartPlaylistID = 6 [00:52] (note how long that took, that was after 2.5k) [00:53] Right. [00:54] eventually that tends towards insane numbers [00:54] wait...isn't cross join n^2? [00:55] Aha! Someone with actual SQL knowledge! [00:55] i did web dev in a past life :-P [00:55] its array multiplication [00:55] err, fine - so it's O(nm) :-P [00:55] the two statements that had that issue were both cross joins [00:55] I play a web dev in my office [00:56] that is one thing I noticed [00:57] RAOF: fyi http://en.wikipedia.org/wiki/Join_(SQL)#Cross_join [00:57] its an odd syntax to use here [00:57] because from foo,bar,quux is a cross join anyway [00:57] i can't see any reason you'd want a cross join at all [00:58] it seems like you'd want to start at the CoreSmartPlaylistEntries and left join your way over to CoreTracks [00:58] CoreArtists.ArtistID = CoreTracks.ArtistID constraints it to be an inner join anyway [00:58] CoreAlbums.AlbumID = CoreTracks.AlbumID ditto [00:58] err, right [00:58] CoreSmartPlaylistEntries.TrackID = CoreTracks.TrackID ditto [00:58] although who knows if sqlite's planner is smart enough to adjust joins based on where clauses [00:58] so possibly a sqlite bogon [00:59] the explicit CROSS JOIN *might* be confusing it [00:59] shouldn't those be on/using clauses? [00:59] broder: equivalent as far as results are concerned [00:59] lifeless: results, yes. planning, possibly not [00:59] using a cross join and constraining in the where is often better for prostgresql's planner [00:59] (implicit cross vs explicit cross) [01:00] broder: yeah, agreeing with you [01:01] oh wow [01:01] its stupid [01:01] AND CoreSmartPlaylistEntries.SmartPlaylistID = 6 [01:01] ^ I *bet* you it runs that for each playlist [01:01] rather than group by [01:01] its nearly guaranteed to be table scanning [01:01] anyhow [01:02] you all need bigger music libraries it seems [01:02] DBO: I don't use banshee ;) [01:02] lifeless, dogfooding is an artform [01:02] DBO: do you have the sqlite file handy? [01:02] DBO: I'm a cat person [01:02] oh i forgot to mention on other REALLY weird thing [01:03] if you do it manually [01:03] its fast [01:03] DBO: besides which, i got tired of fixing upstream evolution performance bugs. [01:03] if you restart banshee and import more files [01:03] its fast [01:03] DBO: I was going to ask if you had the sqlite file handy [01:03] i do [01:03] DBO: but that kindof points at concurrency / contention [01:03] anyhow [01:04] i just figured you were going to try the query by hand [01:04] how long do these take to run: [01:04] SELECT COUNT(*), SUM(CoreTracks.FileSize), SUM(CoreTracks.Duration) FROM CoreTracks CROSS JOIN CoreArtists,CoreAlbums, CoreSmartPlaylistEntries WHERE [01:04] CoreArtists.ArtistID = CoreTracks.ArtistID AND CoreAlbums.AlbumID = CoreTracks.AlbumID AND CoreSmartPlaylistEntries.TrackID = CoreTracks.TrackID AND CoreSmartPlaylistEntries.SmartPlaylistID = 6 [01:04] SELECT COUNT(*), SUM(CoreTracks.FileSize), SUM(CoreTracks.Duration) FROM CoreTracks , CoreArtists,CoreAlbums, CoreSmartPlaylistEntries WHERE [01:04] CoreArtists.ArtistID = CoreTracks.ArtistID AND CoreAlbums.AlbumID = CoreTracks.AlbumID AND CoreSmartPlaylistEntries.TrackID = CoreTracks.TrackID AND CoreSmartPlaylistEntries.SmartPlaylistID = 6 [01:05] to start with - the same query by hand to vet the test environment, and a simpler equivalent to see if th eplanner is at fault [01:05] first one is very quick [01:06] second one is just as quick [01:06] SELECT COUNT(*), SUM(CoreTracks.FileSize), SUM(CoreTracks.Duration) FROM coreartists inner join coretracks on [01:06] CoreArtists.ArtistID = CoreTracks.ArtistID inner join corealbums on CoreAlbums.AlbumID = CoreTracks.AlbumID inner join coresmartplaylistentries on CoreSmartPlaylistEntries.TrackID = CoreTracks.TrackID AND CoreSmartPlaylistEntries.SmartPlaylistID = 6 [01:06] just for kicks and giggles [01:06] are there potentially any library/sqlite/etc options you could tweak to make it do weird things? [01:06] lifeless, again very quick [01:07] broder: use it from two threads/processes at once [01:07] you can also set various pragmas, but the only go-slow knobs I know of are: [01:07] - missing indices [01:07] - concurrent use [01:07] - [very rare] planner issues [01:08] missing indices would effect it even when queried by hand [01:08] right [01:08] and its atable scan anyway, so unlikely. [01:09] which is why I'm coming back to concurrency [01:09] the stop-and-start-and-its-good hints at that too [01:09] why is that? [01:09] I do believe that the smark playlist scan is on a timer. [01:09] DBO: why is what? [01:09] DBO: its a table scan because its doing a count of the world. [01:09] lifeless, why does the stop/start thing hit at it [01:09] hint* [01:09] ah [01:10] RAOF, if its a timer... and it was getting slower [01:10] because after each commit to sqlite [01:10] if it ever got slower than the timer... [01:10] oh my... [01:10] the db is in a stable state [01:10] looking at the sqlite docs, replacing most of those where clauses with on clauses would change how sqlite plans the query [01:10] compaction isn't automatic and would probably be nicable [01:10] sqlite processes on/using when it loads data for the join, and it doesn't process where clauses until later [01:11] broder: its a table scan though, almost certainly) [01:11] you guys are wizards [01:11] DBO: sqlite only lets one process write to the db at aonce [01:11] all I wanted to do was whine about banshee sucking the big one [01:12] and unless you use sqlite3 and make sure your db format is set correctly won't let readers in either [01:12] * RAOF wanted a distraction from grepping through 23GiB of dmesg [01:12] RAOF: how about a 23GiB sqlite file? [01:12] lifeless: Nothing in the X stack is likely to generate that, so I don't have to care about it :P [01:13] RAOF, whats going on with X? [01:13] RAOF: I can totally make a patch for you. [01:13] still working on that hardlock with intel? [01:13] RAOF: also, doom-power-manager-memory-leaks. [01:13] lifeless: They're still happening for you? And I still can't seem to reproduce them. [01:14] RAOF: yes, but not every day. [01:14] DBO: Yeah. It's actually not a hardlock. When dpms state gets toggled it's possible for the vblank code to go all doolally, and so the vblank event never gets sent. [01:14] RAOF: I suspect its when I am switching batteries around, or when it falls out. [01:14] which reminds me to ring lenovo and whinge about that [01:14] :) [01:14] Well, I'll do more battery switching today then. [01:14] RAOF: no panic [01:15] RAOF, some days, you really earn the title of "Fabio" :D [01:15] RAOF: I have a workaround [01:19] That's a funny sql statement? Why does it qualify with WHERE 1=1? [01:20] RAOF: context ? [01:37] lifeless: http://pastebin.ca/2050675 [01:37] (Sorry for the delay, IRC went mad) [01:46] RAOF: probably a bad ORM [01:46] RAOF: we do similar whackiness in lp [01:46] Heh. [01:48] its used as a noop to force default where clauses to be removed with sqlobject, for instance === asac_ is now known as asac [03:12] hi all === rickspencer3_ is now known as bertiewooster [03:15] bertiewooster, you having a gin&tonic or something? [03:16] mterry, yes, a bit of the g and t, for bracing of the brain ... [03:16] ah screw it, I can't pull it off [03:16] :) [03:16] could probably do a better jeeves ;) [03:20] Now I'm thirsty for a g&t [03:21] hehe === skaet_ is now known as skaet [06:57] Good morning [07:26] good morning [07:29] good morning [07:35] good morning everyone [07:46] good morning didrocks [07:46] hey duanedesign, chrisccoulson, alex3f [07:47] didrocks: do you know if rodrigo will be in togay? [07:47] today* [07:47] duanedesign: I think he should be around later, yeah [07:47] :) [08:14] * pitti reminds himself about the weekly reminder [08:14] hey didrocks, morning chrisccoulson [08:14] did you guys have some nice holidays? [08:14] hey pitti! how are you? [08:14] didrocks: I got very little sleep over the weekend, but recovering :) good otherwise [08:14] pitti: 3 wonderful and sunny days there, yeah! :-) and you? [08:14] oh? [08:14] too many packages? [08:15] didrocks: same here; we met a lot of friends, and had our traditional family hike on Sunday [08:15] didrocks: no, went out late every night to meet friends, and got up early [08:15] and had to squeeze in some packing, too :) [08:15] pitti: heh :-) hope the week will be quiet so that you can recover! [08:16] as quiet as a release week can be [08:16] right :-) [08:21] pitti, RAOF: seem on bug #728745 that people are experimenting freezes with some nvidia cards. Installing 173 instead of currrent fixes this for them (right now, I blacklist unity/compiz for them) [08:21] Launchpad bug 728745 in nux "[nvidia, 7300, 7400] display freeze when using unity desktop" [High,Fix released] https://launchpad.net/bugs/728745 [08:22] Hurray :/ [08:22] didrocks: interesting; but seems there's not much we can do about this at this point? [08:22] except perhaps for release-noting it, that it might be worth trying the older version [08:22] pitti: well, as they are balcklisted, they need to add the env variable as well [08:23] I'm blacklisting more and more cards meanwhile, the choice of card is done on nvidia side? [08:23] didrocks: how do you mean, the "choice of card"? [08:24] pitti: like, this card uses the nvidia current driver, or the 173, or … [08:25] IIUC jockey will recommend the most recent driver that claims to handle the card. [08:25] So, that'd be nvidia-current. [08:26] didrocks: right [08:26] the drivers themselves have a list of supported vendor/product IDs [08:26] didrocks: ^ you can see these with apt-cache show nvidia-current [08:27] or nvidia-173, etc. [08:27] the "Modaliases:" line [08:27] pitti: you mean, the insane long line? :-) [08:27] yes [08:27] thanks, I'll know now how to check that list :-) [08:27] didrocks: it's for the purpose of things like jockey, not really for human consumption :) [08:28] didrocks: just filtering out the product IDs should make it a lot shorter [08:28] RAOF: pitti: do you think we can have a look in a SRU to change it for those detected card? (at least, for those who reported that installing 173 worked) [08:29] That would make it impossible to install nvidia-current on those cards though, right? [08:30] didrocks: possible in theory, but we would break the driver for everyone else [08:30] RAOF: right, but seems that unity and compiz session are not in a good state (looking at the bug report) with those card [08:30] (Kubuntu, GNOME classic, gamers, etc.) [08:30] so not sure if it's unity specific or a more 3D handling issue with the driver + card [08:30] well, let's keep it on track [08:43] pitti: I'm looking at bug #664920 as a part of patch-piolting. There's an obviously-safe patch available, and it's upstream. Is that SRUable? [08:43] Launchpad bug 664920 in pygtk "SRU: 100% CPU usage when calling a child process from a python script" [Undecided,Confirmed] https://launchpad.net/bugs/664920 [08:44] RAOF: yes, that looks fine [08:44] RAOF: not sure how much it hurts and whether it should also go to lucid/maverick [08:45] Not sure whether it *hits* Lucid. [08:46] It seems like it hits relatively high-profile apps like Deluge, though. [08:59] chrisccoulson: good morning [09:00] links like apt:abiword now open in Software Center, does this break what you intended with ubufox? [09:06] seb128: bonjour [09:06] seb128: did you have some nice holidays? [09:07] hey pitti, hey desktopers [09:07] pitti, excellent, thanks! splendid weather and quite some outside and relaxing time [09:07] what about you? [09:07] seb128: similar; I met a lot of friends, and we had our traditional family hike on Sunday [09:07] very little sleep :) [09:08] salut seb128 [09:08] lut didrocks [09:09] didrocks, how are you? had a nice 3 days off? [09:09] seb128: yeah, very sunny, a lot of walking in forest and enjoying the city :-) very relaxing thanks! [09:09] seems you had some nice time as well :) [09:09] great! ;-) [09:09] Howdie seb128! [09:10] hey RAOF, how are you? [09:10] Pretty good :) [09:10] how is natty looking btw? [09:10] Less lazy than the other .au-ers, who have today off, too ;) [09:11] RAOF: you're in perth now right? [09:11] lifeless: No, Hobart [09:11] dpm, didrocks , bug 758621 is not nice, is it a problem with translation not being exported or not integrated with this dialog ? [09:11] Launchpad bug 758621 in gnome-session "Need translated fallback message in gnome-session" [High,Triaged] https://launchpad.net/bugs/758621 [09:12] RAOF: ah, returned home :> [09:12] Natty's looking pretty good. There are a couple of known bugs; I've got a handle on chrisccoulson's ‘unity hang’ bug, and will test a fix tomorrow. [09:12] lifeless: Yup :) [09:12] lifeless: Also, Perth's too flat. And hot! [09:12] good morning RAOF, pitti, didrocks, seb128 [09:12] hey chrisccoulson, how are you? [09:13] seb128 - i'm good thanks, although not very rested from the long weekend [09:13] how are you? [09:13] jibel: it's exported normally, let me check [09:13] hey again chrisccoulson ;) [09:13] chrisccoulson, I'm fine thanks, got a relaxing weekend there ;-) [09:13] hi jibel, good morning. I'm not sure about that one, I was about to ask didrocks about it myself [09:13] chrisccoulson: did you see my question above? [09:14] hello again didrocks :) [09:14] seb128 - heh, lucky you ;) [09:14] dpm: I use the gnome-session translation tool, let me see [09:14] i spent sunday tidying my garden and yesterday cleaning my car [09:14] my back feels ruined today ;) [09:14] jbicha, no, i didn't see your question [09:14] links like apt:abiword now open in Software Center, does this break what you intended with ubufox? [09:15] not sure. in any case, there's not much we can do about that [09:15] actually [09:15] no, it shouldn't ubufox calls apturl directly [09:16] so, it's ok ;) [09:16] and it worked the last time i tried it (and software-center has been the default handler for apt: URI's since maverick anyway) [09:17] so it's nothing new [09:18] chrisccoulson: apt: links open in SC for me in Firefox [09:18] yes, but ubufox calls apturl directly for the plugin installer [09:19] oh, so plugins are different? [09:19] yes [09:19] chrisccoulson: cool, thanks [09:22] hmmm, i'm wondering whether to upgrade jo's desktop to natty today [09:22] if she leaves the house ;) [09:22] see if she notices :) [09:25] I'm guessing she'd notice unity! [09:27] she might do [09:52] bryceh: hey, when you have people reporting blacklisted card, please head them to bug #728745 [09:52] Launchpad bug 728745 in nux "[nvidia, 7300, 7400] display freeze when using unity desktop" [High,Fix released] https://launchpad.net/bugs/728745 [09:52] chrisccoulson: I subscribed you to bug 769759, seems some XPI files still have a too strict maxVersion? [09:52] Launchpad bug 769759 in language-pack-en-base "English and a few other Firefox langpacks won't work with 4.0.x" [High,Triaged] https://launchpad.net/bugs/769759 [09:53] pitti - that's fixed with the latest language packs already [09:53] oh, nice [09:53] (or it should be. i didn't check yet) [09:54] chrisccoulson: ah, indeed; I checked now [09:54] thanks, closing [09:54] thanks [09:58] didrocks, no, I'll reassign the bug to unity and let you guys sort out what it should be duped to. I don't expect to get many of these bug reports, and don't want to have to track bugs already closed as fix released. [09:59] let's see how many of them we will get… [10:01] didrocks, even if we get a lot filed against X, I'll stop triaging the natty X bugs in a couple days and turn attention to oneiric, so you needn't worry about getting flooded with dupes [10:01] bryceh: ok, thanks then :-) [10:40] chrisccoulson, did you investigate on bug #765736 [10:40] ? [10:40] Launchpad bug 765736 in bamf "Thunderbird won't stay in launcher and no quicklist" [Medium,Fix released] https://launchpad.net/bugs/765736 [10:49] seb128 - yeah, i think i found another issue in bamf. i need to look again just to be sure though [10:49] ok [11:03] morning [11:03] hey rodrigo_ [11:03] how are you? [11:05] hi seb128, I'm fine and you? [11:06] how was the easter break? :) [11:06] rodrigo_, I'm great thanks [11:06] it was excellent ;-) nice weather, relaxing time [11:07] cool :) [11:07] hey rodrigo_ [11:08] hi pitti [11:19] hmm, https://code.launchpad.net/~rodrigo-moya/ubuntu/maverick/json-glib/fix-756426/+merge/57474 in LP says it's merged, but lp:ubuntu/maverick/json-glib doesn't have it [11:19] I guess it's still in the queue? [11:20] yes, right, it's in https://edge.launchpad.net/ubuntu/maverick/+queue?queue_state=1 [11:21] rodrigo_, right, that's confusing because we don't really commit to those vcs but let the autoimporter do it once the upload is accepted [11:22] ok [11:22] so it's not really merged but it has been handled === MacSlow is now known as MacSlow|lunch [11:34] dpm, didrocks any new on bug 758621 ? [11:34] Launchpad bug 758621 in gnome-session "Need translated fallback message in gnome-session" [High,Triaged] https://launchpad.net/bugs/758621 [11:39] jibel: we discussed that with dpm, I need someone to confirm it in French (can't reboot right now as I'm syncing unity bugs + bug triage) [11:40] jibel: there are French translation, did you try it with them? [11:40] g_key_file_get_string (isn't patch to take translations btw?) [11:41] didrocks, I confirm [11:41] I tried french and german [11:41] and noneof them is translated. [11:42] jibel: and you confirm having the langpack with the translation installed? [11:43] didrocks, and before you ask, I tried today's iso :-) [11:44] german is on the iso [11:44] jibel: ok, there are two solutions: I'll wait for the sync to finish and try it myself. Or I can give you a deb which will hopefully fix it if you prefer [11:45] (the sync before rebooting the session will still takes a couple of hours) [11:45] didrocks, well it's on i386 but not amd64 :/ I'm trying with spanish just to be sure. [11:46] jibel: I'm on i386, I can give you a package for that hardware shortly [11:46] jibel: anyway, seems a nice candidate for a 0-day SRU [11:46] didrocks, indeed [11:48] is anyone able to use language-selector in natty? [11:48] i can't install anything with it here because it tries to install the missing gnome-user-guide-xx packages [11:58] urgh [11:59] pitti - is that in response to my last comment? ;) [11:59] I tried a while ago, but that might have been before the recent ubuntu-docs update [11:59] yeah. we need to drop the gnome-user-guide- bit from pkg_depends don't we? [11:59] i can't use language-selector at all here :( [12:00] pitti - this broke it: https://launchpad.net/ubuntu/natty/+source/gnome-user-docs/2.91.90+git20110306ubuntu1 [12:01] seems a pretty big change so close to release :/ [12:01] chrisccoulson: do you mind filing a bug for it, to track it for an early SRU? [12:01] pitti - sure, can do [12:01] I wonder if that actually breaks installation [12:01] as parts of it are also used in the installer [12:30] jibel: fixed FYI, uploading to -proposed [12:44] hi all, am I in time to propose a fix for bug #769520? [12:44] Launchpad bug 769520 in ubuntu-sso-client "missing dependency on gnome-keyring" [Medium,Triaged] https://launchpad.net/bugs/769520 [12:47] where should firefox be looking for dictionaries? it's currently looking in /usr/share/myspell/dicts, but it seems that not all of our dictionaries install files in to there [12:49] ah, pants [12:49] that's wrong [12:52] nessita, hey you ;-) [12:52] nessita, you should get a sru for it [12:52] seb128: hey there! ok, np [12:56] nessita, how are you? [12:57] seb128: good! I just moved to another house, so I'm a bit overwhelmed with boxes and messes. You? [12:57] oh, nice ;-) still in the same city ? [12:58] hey nessita [12:58] I'm fine thanks, got a nice weekend, relaxing with nice weather [12:58] yeap, and same neighborhood [12:58] rodrigo_: hey there! how are you? [12:58] nessita, fine, still in the same house :-) [12:58] hehe [12:58] moving houses is a pita indeed, I remember when I moved last time [12:59] rodrigo_: yes, is pretty stressing. Not to mention I don't have internet connection there, yet. The provider is taking longer than expected. [13:00] well for a girl as energetic than nessita it's probably nice, she has something to spend energy on after sitting for a day on the computer ;-) [13:00] :) [13:01] nessita, right, that sucks, had the same problem myself, I had to use the local cybercafe for a month or so [13:01] seb128: speaking of that, my partner is mad at me for trying to organize our boxes until late last night :-/ [13:01] nessita, :-( [13:01] well soon you will be done with the moving and will enjoy your new house ;-) [13:01] but yes, I enjoy using my energy (also) in that :-) [13:01] yeah [13:02] do you come to UDS btw? [13:02] seb128: yes! and I'm also going to the summit, so I'll be leaving Argentina next Monday [13:02] seb128, rodrigo_: how about you? === zyga is now known as zyga-afk [13:02] ok, lot of moving and traveling for you then ;-) [13:02] nessita, going to UDS, not to the summit [13:03] I guess you mean the summit before UDS? [13:03] rodrigo_: yes [13:03] nessita, I'm going to UDS, I will not be at the summit but I will there with dx at the same time [13:04] seb128: yes, I still need to find my clothes between the boxes to pack my suitcase. *That* will be fun :-D [13:04] ;-) [13:04] nessita, you are lucky it's almost summer, just get some t-shirts and you will fine ;-) [13:05] seb128: duly noted. Another idea for someone in online services was to buy all the clothe there [13:05] which was very apealing! [13:05] appealing* [13:05] ;-) [13:05] "shopping" [13:05] ;-) [13:05] nessita, yes, because bringing all the boxes to Budapest doesn't work, right? :-D [13:06] rodrigo_: right :-D [13:06] yeah, I thought so :-D [13:10] didrocks, thanks === MacSlow|lunch is now known as MacSlow [13:43] * rodrigo_ lunch === Pici` is now known as Pici === zyga-afk is now known as zyga [14:43] mterry: hey, dejadup has built-in code to poke ubuntu one, upstream, right? [14:44] dobey, sort of. It has code in deja-dup itself, but the command line tool it relies on doesn't have the bits to actually push/get files yet [14:44] the bluetooth icon still isn't showing up in system setting under EN_GB translations by the look of it :( [14:44] mterry: ah, you mean u1sdtool? [14:45] dobey, no, I meant duplicity [14:45] dobey, because the REST API wasn't ready for natty [14:45] oh [14:46] mterry: ok. i'm mainly just wanting to use it as an example of 3rd parties developing against u1 [14:47] dobey, code is there, but not visible.... Probably not a good example since no one can see it yet [14:48] mterry: ok, i added a future tense classifier to my statement then :) [14:49] dobey, :) will do it fine for 11.10 [15:04] cjwatson, skaet: hello there [15:05] so yeah from a desktop team perspective reverting the scrollbars in the classic session would be fine if that's sabdfl or rickspencer3's acked [15:05] hi seb128 hi [15:05] hi, so we'd like to discuss turning overlay-scrollbar off for classic, on the grounds that even though it isn't part of unity we're setting expectations that the desktop UI will largely be "old stuff" with the "Ubuntu Classic" name [15:05] how this ties into GNOME 3 next cycle I frankly have no idea [15:05] I would not feel comfortable dropping those without a sabdfl ack since he forced them on us to start [15:06] cjwatson, right, I dropped appmenu from the classic session around beta2 on the same rational [15:06] I had an unsubstantiated guess that sabdfl might care primarily about unity [15:06] yes, that would be my guess as well [15:06] sabdfl: are you around? [15:06] I would prefer rickspencer3 or sabdfl to ack they are fine with it though just in case [15:06] seb128: understood [15:06] hey rickspencer3 [15:07] I see no reason to worry [15:07] I don't understand what has changed [15:08] part of the problem is that the Classic session is AFAIK not defined anywhere [15:08] seb128, has the desktop team received any feedback from users about this [15:08] some people (e.g. skaet) appear to understand it as "keep the major outlines of the UI roughly as they were before" [15:08] does this appear to be an urgent problem? [15:08] I can't find anywhere where we describe it to users [15:08] it should clearly be in the release notes, and that's probably a tomorrow problem [15:09] rickspencer3, we didn't get a lot of feedback about the scrollbars either way no [15:09] so then why is this an issue? [15:09] http://www.ubuntu.com/testing/natty/alpha2 mentioned it [15:09] but as cjwatson it's somewhat a matter of defining what classic is supposed to be [15:09] why are we even having this conversation? [15:09] "There are now three kinds of sessions in gdm: Ubuntu Desktop will run Unity by default and the Ubuntu Classic Session will run gnome-panel. Ubuntu Classic supports all video hardware and video drivers. Ubuntu Desktop requires 3D driver support. Finally, you can force a "2D mode only" with Ubuntu Classic Session (no effect) which has the same interface than the Ubuntu Classic session." [15:09] which is not very explicit and rather jargon [15:09] I think o-scrollbars checks an env var at some point, so an SRU to set that in the classic or unity session seems feasible [15:10] pitti: technically it's certainly feasible (LIBOVERLAY_SCROLLBAR=0) [15:10] rickspencer3: I think skaet is bringing this up as a user, at least in part [15:11] cjwatson, I would prefer that we focus on an awesome release [15:11] and let this one slide [15:11] I'm not really seeing it as a bona fide issue at this point [15:11] users can fairly easily turn this off [15:11] we also need to check with UX/DX what the intent was [15:11] and pgraner suggests that a release note about how to do it would be feasible [15:11] right now the package doesn't make any effort to set its behavior according to the session type [15:12] i. e. whether it is meant to also change the classic gnome session or not [15:12] so, may I suggest we focus on testing what we have and ensure that users will get a solid experience, for example look for upgrade bugs and crashers, instead? [15:12] rickspencer3: to be clear, I at least don't see this as a respin issue, but many of the things we typically discuss in these few days end up as early SRUs [15:12] I don't see a conflict between discussing this kind of thing and focusing on an awesome release [15:12] and an awesome release requires awesome release notes, at least [15:13] cjwatson, that's fine [15:14] nobody seems to have a clear vision of what Classic means, and that seems essential to have in the release notes [15:15] so far I considered it as "gnome-panel instead of unity" [15:15] i. e. not as far as stracciatella-session [15:15] +1 on pitti (short and concise description) [15:15] the latter should have standard scrollbars, metacity, etc. [15:15] OK, but I think we need something less jargon for the release notes [15:16] agreed === bjf[afk] is now known as bjf [15:44] anyone know where robert ancell is today? [15:45] it's over his work hours but holidays as well [15:46] desrt: asleep? :) [15:46] i suppose that makes sense! [15:46] * desrt hasn't seen him all weekend [15:47] desrt, I think there was a holiday for them on Mon and Tues, and now it's early early morning on Wednesday for them [15:48] desrt, he's off this week it seems [15:48] so next week [15:48] seb128: ah. that's closer to what i was expecting [15:48] or drop him an email [15:48] ya. i dropped him a couple [15:48] he appears to be ignoring his mails [15:48] good for him :) [16:02] desrt, it's what time off work is for it seems ;-) [16:02] mterry, hey [16:02] seb128, hello! [16:02] mterry, how are you? [16:03] seb128, good. Did you hear I got core-dev yesterday? [16:03] mterry, is deja-dup supposed to add a "go back to the previous version" to nautilus context menu on any file and location? [16:03] mterry, no I didn't, congrats! [16:04] seb128, it does add the context menu to everything, though ideally it would restrict it to only things in your backup. I just never got around to restricting it. there's a bug for it [16:04] mterry: oh really! nice ;) Hey! [16:04] didrocks, :) [16:05] mterry, ok, I just had a "wth" moment trying to figure what was context menu entry was doing when right clicking on an icon on my desktop until I tried and got a deja-dup dialog ;-) [16:05] (I installed it recently) [16:05] hi all! Is it ok to set bug 746375 to incomplete as per last comment? [16:05] Launchpad bug 746375 in libreoffice "soffice.bin crashed with SIGSEGV in uno_type_sequence_construct()" [Medium,Confirmed] https://launchpad.net/bugs/746375 [16:05] mterry, btw do you get subscription to bugs filtered or do they land in a noisy launchpad box? [16:06] seb128, I filter [16:06] Sweetshark, hey, you are the maintainer for that package so feel free to deal with bugs the way it works for you, but it's usually fine to use incomplete if it lacks details or infos you need [16:06] or if it might be fixed and need testing [16:07] mterry, ok, I Cc-ed on an unity bug in case you were interested [16:07] seb128, oh I saw that, the drag and drop to startup lists [16:07] seb128, if you subscribe me especially, it will land in a high-notice filter [16:08] skaet: as I said before, I think the new scrollbars should be in classic, but not in stracciatella-session; having it in the latter is a bug [16:08] seb128: thanks. I will set to incomplete to urge reporter to comment on the issue. [16:08] skaet: I'll add a task to bug 766660 to fix it in stracciatella [16:08] Launchpad bug 766660 in overlay-scrollbar "[FFE] Switch the ayatana-scrollbar on by default" [Undecided,Confirmed] https://launchpad.net/bugs/766660 [16:09] pitti, thanks. When you say "stracciatella" what should I be translating that to? ;) [16:09] pitti, did you see bug #747796 btw? [16:09] Launchpad bug 747796 in pygobject "gnome-language-selector crashes after a fresh install" [Undecided,Confirmed] https://launchpad.net/bugs/747796 [16:14] skaet: apt-cache show gnome-stracciatella-session :) [16:14] pitti, will do. [16:15] seb128: I didn't; I'll have a look after I'm done with my current bug fix (writing the commit message, so that'll be "soon") [16:15] pitti, will you take a pass at composing the release note for the scroll bars, and what uses should be expecting [16:15] skaet: can do, yes [16:16] meh [16:17] * Sweetshark uses tmux to get more stability for his irc session, and not less by tmux itself crashing .... [16:17] Sweetshark: ever tried bip? [16:18] * pitti has znc running on his server as an IRC proxy [16:18] pitti: you left bip? [16:18] no, I have used znc for years, too lazy to switch [16:18] it never failed me [16:18] pitti, draft is up on https://wiki.ubuntu.com/NattyNarwhal/ReleaseNotes/ [16:19] but I heard that bip is a bit more robust, so might be better for first-timers [16:19] bip was working fine when I used it :) [16:19] skaet: oh, not to https://wiki.ubuntu.com/NattyNarwhal/TechnicalOverview any more? [16:19] just need the intrepid version for some bug fixes to backport to hardy [16:20] pitti, yup we're starting the transition to the expected location now, and reworking it to follow the precedents a bit. [16:20] Lots of wordsmithing and editing still needed, so feel free to change beyond. [16:20] I'd like this change in, so I can take the tone from your comments for the rest of the collateral (announce material, etc.) [16:22] pitti: thanks for the tip. I might have a look at bip later. [16:24] skaet: I updated the bug report now [16:28] pitti, thanks [16:29] Sweetshark, bryceh, chrisccoulson, didrocks, tremolux, Riddell, kenvandine, cyphermox, mterry, rodrigo_, seb128, tkamppeter: meeting in 2 mins [16:29] ok [16:29] zug zug :) [16:29] ready ;) [16:29] hey [16:30] * Sweetshark whistles "the final countdown" by europe [16:30] cyphermox, was that a warcraft reference? [16:30] skaet: https://wiki.ubuntu.com/NattyNarwhal/ReleaseNotes?action=diff&rev2=2&rev1=1 -> not sure what level of detail you want? [16:30] rickspencer3, it was [16:31] we're under attack! [16:31] Sweetshark, bryceh, chrisccoulson, didrocks, tremolux, Riddell, kenvandine, cyphermox, mterry, rodrigo_, seb128, tkamppeter: meeting time [16:31] hi [16:31] hello everyone! [16:31] forgive the partial insanity, I'm reading packet captures for dhcpv6 stuff [16:31] https://wiki.ubuntu.com/DesktopTeam/Meeting/2011-04-26 [16:31] hey! [16:31] \o/ [16:31] hey [16:31] hi! [16:31] topic list: [16:31] 1. AWTY? [16:31] 2. ... [16:31] 3. PARTY! [16:32] hi [16:32] can we go to 3 directly? :-) [16:32] hi [16:32] didrocks, :) [16:32] WI status and RC bug status look good from my POV [16:32] is anyone aware of a dealbreaker? if so, please speak up NOW [16:32] but natty is pretty much what it is now [16:33] kenvandine, didrocks, and tremolux already sent their partner/unity/software-center reports to the wiki (thanks!), do we need to discuss anything for them? [16:34] not for me [16:34] we are in a good shape (crossing fingers as well ;)) [16:34] nothing really more to say for me also [16:35] Riddell: what's the latest word on the Kubuntu side? [16:35] pitti: we're about good to go [16:35] sweet [16:35] although kubuntu-meta needed a -proposed update because a load of language packs disappeared [16:35] ah, I saw that in #release [16:35] but no killer problems [16:36] Riddell: we should be able to get that to -updates quickly, and rebuild the DVD against that? [16:36] it's just a seed update and a -meta rebuild, right? [16:36] yes [16:36] well no seed update needed, just -meta [16:37] ah [16:37] ok, cool [16:37] hi [16:38] last thing I have is a general call to help with ISO testing, as we needed to do a couple of respins [16:39] as nobody else has something, let's get ready for the final mile, and get natty out of the door! [16:39] thanks everyone [16:39] thanks pitti [16:39] pitti, so, where are the isos to test? I have a vm where I can do some testing [16:40] and for the record, I'm such a dork [16:40] oh, i just saw bug 760131. that might explain why my laptop barely lasts 1 hour in natty :) [16:40] Launchpad bug 760131 in linux "Power consumption raised significantly in natty" [High,Confirmed] https://launchpad.net/bugs/760131 [16:40] I booked my Budapest train tickets for next week.. [16:40] heh [16:40] rodrigo_: http://iso.qa.ubuntu.com/qatracker/ is the starting point, it links to images and test reports [16:41] thanks pitti ;) [16:41] ok [16:41] * Sweetshark is slowly tapping his feet. [16:41] Are we at top 3 already? [16:41] pitti, you can join the dx sprint this way ;-) [16:41] Sweetshark: Thursday :) [16:42] seb128: nah, I think I'll spend the week with setting up our new flat [16:42] pitti, seems a better option indeed ;-) [16:43] pitti: Early train tickets are still better than missing an intercontinental flight. I managed to do that once ;) [16:54] Sweetshark: heh, absolutely [17:00] didrocks, skaet: ugh, stracciatella-session currently starts unity instead of gnome; ugh [17:01] pitti: argh, it should unset the COMPIZ variable I guess [17:01] pitti: I can have a look tomorrow into that if you prefer, stilll working on the 0 day SRU [17:01] didrocks: which one in particular? [17:01] didrocks: I can work on it, I want to fix it for LIBOVERLAY_SCROLLBAR anyway [17:01] pitti: COMPIZ_CONFIG_PROFILE [17:01] ah [17:01] didrocks: merci [17:01] pitti: de rien :) [17:02] didrocks, pitti: unsetting the profile will not start gnome-panel though? [17:02] seb128: unsetting the profile will start the default profile [17:03] which is the one in ubuntu classic [17:03] and by default, no unity, until the user set it himself of course in ccsm for instance [17:03] didrocks, well that will not start gnome-panel still? [17:04] oh sorry, misread [17:04] dunno how the stracciatella-session works but it seems it should use another session [17:04] rather than unsetting the environment [17:04] straciatelly should use the classic-gnome.session I guess [17:04] right [17:04] pitti, ^ [17:05] how? [17:05] Exec=env -u COMPIZ_CONFIG_PROFILE LIBOVERLAY_SCROLLBAR=0 gnome-session [17:05] I tried this now [17:05] disabling the scrollbars works [17:05] pitti, ack. [17:05] and $COMPIZ_CONFIG_PROFILE isn't set any more either [17:05] pitti: --session=classic-gnome [17:06] ad an arg to gnome-session [17:06] as* [17:06] didrocks: ah, cool; should I still unset $COMPIZ_CONFIG_PROFILE? [17:06] see /usr/share/xsessions/gnome-classic.desktop [17:06] I'm not sure what this really does [17:06] pitti: right, you need to unset it [17:07] didrocks, why? it should not be set if you run gnome-session --session=classic-gnome? [17:07] pitti: and in oneiric, we can see so that all this use the session system to avoid this kind of quirks (in the display manager even!) [17:07] desktop* [17:07] hmm [17:07] now I have both gnome-panel and unity running [17:08] COMPIZ_CONFIG_PROFILE= LIBOVERLAY_SCROLLBAR=0 gnome-session --session=classic-gnome should work [17:08] I guess [17:08] Exec=env -u COMPIZ_CONFIG_PROFILE LIBOVERLAY_SCROLLBAR=0 gnome-session --session=classic-gnome [17:08] in /usr/share/xsessions/gnome-stracciatella.desktop [17:08] COMPIZ_CONFIG_PROFILE will not unset it, will it? [17:08] COMPIZ_CONFIG_PROFILE= rather? [17:08] seb128: it's not set [17:09] seb128: "-u" does that [17:09] hum [17:09] env | grep COMPIZ ? [17:09] in the session [17:09] nothign [17:09] but actually I think I want to run 2d-gnome anyway [17:09] is that a fresh user? [17:09] as compiz isn't vanilla gnome either [17:09] or an user where you enabled unity via ccsm in the standard profile [17:09] seb128: no, existing one; I'll try with a fresh one [17:11] awesome [17:11] now it works for both existing and fresh user [17:11] pitti, works? [17:12] indicators are still there, though [17:12] not easy way around that [17:12] but that's our panel configuration I guess, hard to avoid thiose [17:12] yes [17:12] seb128, didrocks: Merci pour votre aide! [17:12] de rien! ;-) [17:12] pitti: mais de rien :-) [17:25] bah, why does indicator-weather depends on desktopcouch? [17:49] seb128: due to bug 738762 (see also the changelog entry for 11.03.20+repack-0ubuntu2) [17:49] Launchpad bug 738762 in weather-indicator "indicator-weather crashed with ImportError in /usr/lib/pymodules/python2.7/desktopcouch/records/server.py: No module named application.server" [High,Fix released] https://launchpad.net/bugs/738762 [17:50] geser, right, I've seen that, I wonder why they need desktopcouch at all [17:53] the stupid desktopcouch,erlang eats cpu still it seems, I didn't have it installed for a while and I didn't miss it [17:57] is usb-creator listed in the compiz alt-tab list for others? [17:58] seb128 - yeah, it is here [17:58] ok, weird [17:58] well the initial dialog is there [17:58] not the one you get when writing an iso [17:59] good night everyone! still need to do some packing tonight [17:59] seb128: oh, yes, I saw that also a couple of days ago when I used it [17:59] 'night pitti [18:00] mdeslaur, thanks ;-) [18:17] * kenvandine waves [18:29] pedro_, hey, around at all? [18:32] rickspencer3, hello, yeah [18:32] hi pedro_ [18:32] rickspencer3, saw the questions on the gstreamer bug? [18:33] pedro_, I did [18:33] thanks [18:33] I'll work on it later, there is some owrk for me in that question :) [18:33] pedro_, https://bugs.launchpad.net/ubuntu/+source/gnucash/+bug/771369 [18:33] Launchpad bug 771369 in gnucash "Display corruption/empty ledger" [Undecided,New] [18:33] ok :-) [18:33] mdz ran into this today [18:33] would you please see if you can do some root cause analysis? [18:34] gnucash is an important problem [18:34] I'm sure mdz will help [18:34] rickspencer3, yes, i'll have a look to it [18:34] pedro_, of course, if you have other priorities that are pressing, please talk to your manager before you change those priorities around :) [18:34] thanks pedro_ === Pici is now known as Guest67539 [18:36] no problem [18:39] DBO - i added a comment to bug 765736, because people are still having matching issues in thunderbird (and other apps) [18:39] Launchpad bug 765736 in bamf "Thunderbird won't stay in launcher and no quicklist" [Medium,Fix released] https://launchpad.net/bugs/765736 [18:39] i think i know what is happening, but you understand this better than me ;) [18:40] im glad one of us does [18:40] rickspencer3, pedro_: the gnucash problem might be overlay-scrollbar [18:41] chrisccoulson, I thought I made it fixed so it does fall back [18:41] let me double check that === Pici` is now known as Pici [18:41] Hello Ubuntu land, does anyone know any Shell commands that I could type to get an application to start automatically every time I login? [18:42] kenvandine, you saw similar issues due to those? [18:42] no [18:42] but i think that is what i had heard [18:43] LIBOVERLAY_SCROLLBAR=0 gnucash [18:43] ok [18:43] to confirm [18:43] mdz, ^ [18:43] seb128, thanks, will try right now [18:44] ap hi, this channel is really for developers who are making Ubuntu, we aren't too good at tech support here, unfortunately [18:44] ap, have you tried #ubuntu? [18:45] ap: run gnome-session-properties and use the add button [18:45] but what rickspencer3 said [18:45] usually better to use #ubuntu for such questions [18:45] ap you're welcome to hand out though [18:46] and if you want to help with Ubuntu, this is a good place to come, too [18:46] mdz, we are preparing an SRU for overlay-scrollbar which includes adding gnucash to the blacklist [18:47] kenvandine, seb128, that fixes it [18:47] * rickspencer3 shakes fist at scrollbars [18:47] * kenvandine does too! [18:47] have a good night everyone! [18:47] good night didrocks [18:47] 'night didrocks [18:47] mdz, cool, it'll be fixed soon in an SRU :) [18:47] didrocks, splits before the scrollbar ranting starts [18:47] he's very smart [18:47] * ogra_ wonders if thats the natty dance you guys probe here [18:47] thanks, have a nice day kenvandine and nice evening seb128 :) [18:48] kenvandine, can you define "fixed"? [18:48] rickspencer3: exactly :-) [18:48] I will not comment on scrollbars [18:48] ap - you could start by having a look at the autostart files in /etc/xdg/autostart for other applications installed on your system, and then create your own and stick it in ~/.config/autostart [18:48] blacklisted [18:48] kenvandine, next you're going to tell me that list is hardcoded in the source so I can't change it locally without rebuilding the thing :-) [18:48] SRU for it [18:48] yeah, so that's not really "fixed", but I know wha tyou mean [18:48] mdz, you are a smart man [18:48] :-D [18:48] * mdz beats his head on the desk [18:48] mdz, not our decision if that makes any difference [18:48] kenvandine, really? I thought there was a list somewhere [18:48] ;-) [18:48] in code! [18:49] *sigh* [18:49] rickspencer3, right, in a .c source [18:49] (yeah, in code) [18:49] in C code, needs to be compiled :p [18:49] rickspencer3, you know how excited we were about this :) [18:49] what if I uninstall liboverlay-scrollbar-0.1-0? [18:49] mdz, you could do that [18:49] mdz, you will get back to old boring scrollbars ;-) [18:49] but then we lose you as a tester [18:49] it's easier to just turn them off though [18:49] * mdz cheers [18:50] kenvandine, isn't there a file that you drop in somewhere that you can just set the scrollbars not to be active? [18:50] mdz, until last thursday, the plan was to enable them for a whitelist of apps [18:50] you can set a variable [18:50] kenvandine, the bug is currently filed on xserver-xorg-video-intel. should I move it to overlay-scrollbar or something? [18:50] LIBOVERLAY_SCROLLBAR=0 [18:50] in your .profile [18:50] we will get some other surprised, I just noticed today they don't work correctly in anjuta either [18:50] mdz, yes please [18:51] we will probably do quite some sru rounds [18:51] kenvandine, this is what happens when I take the bank holiday weekend off, eh? ;-) [18:51] seb128, daily SRUs for the first 2 weeks :) [18:51] * kenvandine hides [18:51] mdz, yup! [18:51] :-D [18:51] hmm [18:52] seb128, is it too late to advocate for a whitelist? [18:53] mdz, you can try but we had that and sabdfl forced on us to drop it [18:53] I doubt they will respin iso for it [18:53] mdz seb128 I think we should consider an SRU that brings bakc the whitelist [18:53] no, no respin for this [18:53] not even close [18:53] sabdfl decision [18:53] you can try to argue with him [18:54] we try, we loose [18:54] tried [18:56] rickspencer3, I would support doing that [18:57] the argument for enabling it globally is to help put pressure on fixing apps as well as help identify all those apps that need fixing [18:57] but it would have been useful if we had done that from the beginning [19:00] hmmm. i like how *every* release of ubuntu we have, *i* seem to be the only one who sees all kinds of memory leaks. [19:00] well every recent release anyway [19:01] hyperair, i was just noticing one... just now [19:01] i think [19:01] indicator-datetime [19:01] kenvandine: in maverick, the entire indicator stack was leaking [19:01] indicator-applet especially [19:01] and indicator-messages-service [19:01] i see that in natty, the situation is no better [19:01] appindicators where, i knew that [19:02] i don't see that [19:02] * ogra_ sees massive CPU hogs with some weird erlang stuff pulled in by indicator-weather [19:02] but, indicator-datetime does bad things [19:02] indicator-messages-service is using 78 MB right now [19:02] 78 freaking megabytes. wtf is it doing that needs so much memory? [19:02] indicator-datetime-service is using 489M of RSS right now [19:02] and then there's e-calendar-factory from evolution occupying 125M. [19:03] and pegging a cpu [19:03] but that's evolution, so it's excusable, because we all know that evolution is a bloody hog. [19:03] kenvandine: ++ [19:03] that is the trigger, evolution --force-shutdown [19:03] causes indicator-datetime-service RSS to grow [19:03] u oh [19:03] i just did that [19:03] then restarting evo makes it grow more [19:03] hrm [19:04] and here i was thinking that indicator-datetime seemed nice and lean and undependent on evolution's stupid network hangs [19:04] can you imagine, when i booted ubuntu up, i stared at the used memory value and it said 800M. [19:05] now it's 2.18G. [19:05] to be fair, let's kill firefox and thunderbird.. [19:05] 1.65G. [19:05] hyperair, on my desktop box that hasn't be up since thursday... it is using 1.3G of mem [19:05] looky, the memory usage more or less doubled since booting up. === czajkows1i is now known as czajkowski [19:06] of which 1.1G is firefox [19:06] and all the memory is used by, *dun dun duuun* the useless system daemons. [19:06] kenvandine: lolwut. [19:06] 200M for the rest? [19:06] that's impossible [19:06] yeah, doesn't add up :) [19:06] well RSS for most other things is small [19:06] compiz is using 215M of memory. [19:06] whee. [19:06] compiz is only 53M [19:07] you know what? [19:07] compiz minus unity was seriously lean [19:07] and evo isn't running :) [19:07] it never took more than 30M of memory [19:07] then unity came along [19:07] and bam [19:07] 215M of memory. [19:07] seriously. [19:07] you know what? i'm reluctant to say this, but there's a common denominator [19:07] compiz is using 91MB here [19:07] anything that comes out of canonical leaks memory. [19:08] only the stuff that went through the drilling dept. [19:08] which is pretty much everything [19:09] the only thing from canonical that doesn't leak here seems to be ubuntuone [19:09] but back in the day, desktopcouch, which ubuntuone is so fond of, leaked like hell anyway [19:09] i recall killing beam.smp at least once a day [19:10] in maverick, the frequently killed stuff were: indicator-applet, gnome-power-manager, indicator-messages-service, everything ibus [19:10] at least that wasn't as disruptive as killing compiz. [19:10] thats the thing i had to kill frequently for the weather indicator too [19:11] kenvandine, you don't keep twitter open in a tab in firefox by any chance do you? [19:11] no way [19:11] (beam.smp) [19:11] ah, ok [19:11] i try my best not to use a web browser :) [19:11] kenvandine, i did that, and then firefox started climbing to over 1GB [19:11] seriously, why can't we have *leak-free* applications? [19:11] firefox and thunderbird are lost causes of course [19:12] running firefox + thunderbird simultaneously was the reason i got 2 more G of RAM. [19:12] chrisccoulson, i have only 7 tabs open on that box too [19:12] almost all of them are wiki.ubuntu.com pages [19:12] nothing heavy [19:12] and 1 launchpad [19:12] hmmmm, that doesn't sound normal :/ [19:12] oh.... and google calendar! [19:14] right, i'm going to do a build of firefox with all the debugging turned on [19:15] kenvandine, it has some pretty useful stuff for leak detection. would you mind running it once i've built it (and figured out how to use it)? [19:15] sure [19:16] chrisccoulson, this might just from having that google calendar page open for 5 days :) [19:17] that might be a useful reproducer. i know that twitter makes eat it RAM like crazy too [19:19] a great excuse to use gwibber :) [19:24] beam.smp used to eat RAM like crazy too [19:24] which was used by gwibber [19:24] thankfully, that stopped. [20:38] ogra_: killing beam.smp for the weather indicator makes no sense. is it written in erlang or something? :P [20:39] dobey, well, it depends on erlang stuff and i dont have ubuntuone installed on that machine [20:40] ogra_: really? that seems weird [20:40] why does weather indicator use desktopcouch? [20:40] crazy [20:40] no idea [20:41] but removing it gave me 20-30% of one of my CPUs back [20:42] dobey, i think it syncs your preferences or something [20:43] ugh [20:43] i removed it recently too for the same reason [20:43] right, I was pondering uninstalling it today due to that as well [20:44] it also became more crashy when they added desktopcouch support [20:44] i think some of the same problems gwibber had [20:45] doesn't always start fast enough, etc [20:45] * kenvandine was thinking about forking it from before the change and maintaining it in a ppa :) [20:45] why did they start using desktopcouch? [20:45] i assume to sync your settings [20:45] but i really don't know [20:45] we should just distro patch that out ;-) [20:45] hehe [20:45] because someone was probably like "oh, we can store stuff in desktopcouch and it'll sync. let's put EVERYTHING in there." [20:46] dobey, i've heard that before :) [20:46] you don't want to put anything in there [20:46] we should just make ubuntu conflicts on desktopcouch [20:46] haha [20:46] poor desktopcouch :) [20:46] no the problem is that people use it dumb and then it goes nuts [20:47] *cough*gwibber*cough* [20:47] dobey, just remember i didn't push for that... :) [20:47] that was all ryan [20:47] haha i know [20:47] it works great for stuff that doesn't change constantly [20:47] there was some nice things about it... i do like the records api [20:48] indicator-weather shouldn't change constantly... but it still gets unhappy [20:48] so i suspect that indicator-weather is probably storing some kind of state in it [20:48] they have 3 calls in one source file to it [20:48] it should be easy to make optional at least [20:48] maybe i'll look for some hybrid :) [20:48] sounds like a fun project for the evening :) [20:49] well it's not like contacts and notes have this problem [20:49] kenvandine, you will get paid in drinks at UDS if you drop the depends to a recommends ;-) [20:49] excellent reason :) [20:49] seb128: i don't think it's that simple [20:49] wtf... i can't branch it! [20:50] kenvandine, ? [20:50] dobey, it should be doable... [20:50] it looks like it imports it unconditionally [20:50] so would require a nice hefty patch [20:50] dobey, hum, we don't have the same definition of simple I think [20:50] bzr: ERROR: Not a branch: "bzr+ssh://bazaar.launchpad.net/%2Bbranch/indicator-weather/". [20:50] dobey, to me it seemes an hour hack, not a week hack [20:50] yeah [20:50] it should be easy [20:50] kenvandine: yeah, that breaks [20:51] oh it's easy [20:51] wtf [20:51] but it's not as simple as just changing the dep to a recommends [20:51] kenvandine: i think it's actually hosted in git or something [20:51] of course [20:51] humm [20:51] kenvandine: the package branch should work though [20:51] LP says lp:indicator-weather [20:51] kenvandine: yeah, there's an empty branch that was never pushed to [20:51] sigh [20:51] don't you love open source? :) [20:52] i'll make it conditional and propose the branch upstream :) [20:52] kenvandine, oh, I got bitten by that today as well [20:52] kenvandine, bzr get lp:weather-indicator [20:52] indicator-weather is another broken project [20:52] don't ask... [20:52] wtf! [20:52] haha [20:53] I spent 5 minutes trying to figure why indicator-weather was not translatable but had translations today [20:53] before figuring the project is weather-indicator [20:53] it's pretty confusing [20:53] especially that the indicator-weather one exist [20:53] yeah [20:53] their trunk just point to an user vcs [20:54] ugh, uses cmake for a python project? [20:54] this keeps getting more and more exciting [20:54] :-D [20:54] kenvandine, you might want to check out the 2.0 serie if you want to get an SRU for natty ;-) [20:54] lol [20:54] i wasn't thinking SRU for natty... [20:54] think we can do that since it is in universe? [20:54] it would be nice [20:55] we can drop the depends to a recommend for a sru [20:55] plus the patch... [20:55] oh, it is python and vala mixed [20:55] it's an universe source and the diff should be something the sru team can review [20:56] well I meant that's something we should be able to justify for a sru, I didn't imply it was changing the control only [20:56] but like doing a try: for the import and not enable some feature if the recommends is not installed should be sru-able [20:57] kenvandine: ugh, cmake. [20:58] i don't have to touch the client stuff, which is in vala [20:58] oh... 2.0 series is very different [20:59] did they start porting to vala? [20:59] kenvandine: LOL; https://twitter.com/#!/segphault/status/62967435854229504 [20:59] seb128, yes [20:59] service is python client is vala [21:00] dobey, haha :) [21:00] seems like it should be the other way around [21:00] seems like it should be all vala [21:00] service is what is running with the session and should be low use [21:00] right [21:01] python is like the marijuana of programming languages. it's a great starter, but need to move on to the hard stuff. [21:02] trunk looks way more complex that 2.0 [21:04] kenvandine, let's just "fix" 2 since that's what is in ubuntu still and let the upstream guy figure what they do with the refactoring? [21:05] yeah :) [21:05] well, it might not be to hard to forward port it either [21:41] hey kenvandine, feeling better? [21:41] pitti, much! [21:41] th [21:41] thx [21:41] nice to hear [21:49] pitti, seemed pretty short lived, my daughter and I both ended up with a fever last night and we both had a horrible night [21:49] but after a few hours of sleep this morning i feel great ! :) [21:55] kenvandine: it's because of the socialist health care, obviously [21:58] a fever, in the summer, thats a double whammy. glad things are better! [22:09] it's not summer yet. just tree orgy season still right now [22:22] good night everyone [22:25] yeah my car is now yellow [22:25] used to be red [22:39] pitti: will you kill me if I'd propose a new release for the scrollbars with a nice bugfix? [22:54] Cimi_: i am pretty sure at this point it will have to be a 0-day SRU [22:55] dobey: I've fixed some bugs related to a wrong colorization that people were getting [22:55] dobey: but I've done the release today with other bugfixes, and just landed in natty-proposed [22:56] dobey: I don't want to make people losing time with another release, but the reality is that it'll be nice to have this fix in as well [22:57] Cimi_: all i can say is upload it to -proposed and bug people in the morning to get it in :) [22:57] Cimi_: i have to do a couple tomorrow myself [22:59] Cimi_, it's great that the scrollbars will get better [22:59] to be clear, it will be an SRU [22:59] I know [22:59] rickspencer3: maybe 0sru? [22:59] Cimi_, well, I don't know if we'll have zero day SRUs [23:00] ok [23:00] I haven't seen anything with that kind of urgency [23:00] but certaintly, as soon as it is ready [23:03] rickspencer3: that's because *it is* ready :) [23:04] jasoncwarner: ping? [23:04] rickspencer3: I've intentionally decided to split the releases: one with some fixes over stability (like this), later with your tip for disabling the thumb on selection [23:04] hey Sweetshark...otp...get back to you in a few? [23:05] rickspencer3: so to have as soon as possible the stability/compatibility fixes, later new features [23:05] Cimi_, "ready" means that it's been through the SRU process [23:05] ok [23:05] jasoncwarner: alright ;) [23:05] Cimi_, no worries, all is well [23:05] the desktop team is really good at this part :) [23:05] (well, they're good at all parts, of course) [23:18] RAOF, present? [23:18] actually bryceh, you are a better candidate if you're here === bjf is now known as bjf[afk] [23:34] DBO, yeah [23:35] DBO, in the midst of sorting out an arrandale gpu lockup issue [23:35] bryceh, so on resize of my screen, nvidia is setting a lot of textures to white... [23:35] is there something perhaps we is doing wrong? [23:37] how are you resizing it? [23:37] and what arr you resizing it too? [23:37] and does it occur only on compiz/unity or also classic/no-effects? [23:37] :-) so many questions [23:38] if you prefer, file a bug report via ubuntu-bug xorg and give me the bug #, that might be easier than playing 20-questions ;- [23:38] ) [23:38] sorry [23:38] uhm [23:38] onesecond [23:38] okay [23:39] bryceh, https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers/+bug/752445 [23:39] Launchpad bug 752445 in nvidia-graphics-drivers "Intermittent white window contents when maximizing/switching windows" [Undecided,Confirmed] [23:39] its this bug [23:39] but triggered specifically for me when I change my screen relution [23:39] resolution [23:42] DBO: Yo yo! [23:42] DBO, generally white window issues are compositing related [23:43] bryceh, yeah [23:43] initially I figured it was a damage event issue [23:43] so I told compiz to do a full screen damage every time (essentially ignoring damage) [23:43] this changed nothing [23:43] DBO, wonder if it might be related to the overlay scrollbar stuff? [23:43] it looks to me like the pixmaps themselves used to back these windows are being wiped to white [23:44] (when the windows redraw potions of themselves, that pops up but the white elsewhere remains) [23:44] RAOF, I dont want to steal time from both of the graphics gurus [23:45] I pinged you before I noticed bryce had already commented on the bug :) [23:45] thus I pinged him instead :) sorry :) [23:45] DBO, I saw bug #770304 this morning which sounds vaguely similar [23:45] Launchpad bug 770304 in overlay-scrollbar "GnuCash Accounts don't redraw with overlay scrollbar" [Medium,Fix committed] https://launchpad.net/bugs/770304 [23:46] this is specific to changing screen resolution with nvidia + compiz + unity? [23:46] dont know if unity is required or not [23:46] seems to help [23:47] but I swear I see it without too [23:47] other people are reporting similar issues in the classic session [23:52] DBO: You're going to like this. I woke up today realising how to fix the WaitMSC hang :) [23:59] DBA I think that's because of the overlay scrollbars, which exist in classic [23:59] DBO, ^