[01:36] RAOF, is there a tool that lists all the X windows in a tree and their properties? Ideally graphically and dynamic [01:37] I believe you might be looking for “xwininfo -tree -root”? [01:37] think i can answer that with a no, regarding dynamically and graphically [01:37] that would be a nice tool [01:38] Sarvatt, cheers [01:39] RAOF, so I want properties, and in this case the RESOURCE_MANAGER property [01:41] xprop (obviously) dumps those properties, but that doesn't get you your tree. [01:42] But a quick shell iteration over the grep of the output of “xwininfo -tree -root” should build you a tree of properties. [01:42] ta [02:15] jasoncwarner_: ping :) [06:26] Good morning [06:27] good morning [06:27] Morning pitti! [06:28] hey RAOF, how are you? [06:28] I'm ok. [06:28] Somewhat in a twisty maze of nux. [06:30] RAOF: uh, what led you there? [06:31] Helping DX write some tests for unity. [06:32] yay, current glib/g-i/etc. landing in unstable [06:39] Oh! When is precise going to be debootstrapable? [06:40] RAOF: how do you mean? it has been for a long time [06:40] RAOF: we have had daily images since UDS [06:40] Hm. So why is mk-sbuild failing with perl-modules? [06:41] oh, still? that should only have been temporarily last Wed/Thu when Colin did the perl 5.14 migration [06:43] * RAOF tries again. [06:51] Still a winner: "perl : Depends: perl-modules (>= 5.14.2-5) but 5.12.4-6 is to be installed" [06:54] well, it seems the old binary is also present :-/ [06:54] perl-modules | 5.12.4-6 | precise | all [06:54] perl-modules | 5.14.2-5 | precise | all [06:56] o_O that shouldn't even be possible [06:56] right [06:56] * micahg wonders if the binary domination is broke, there was some work on that [06:59] * micahg pokes LP devs [07:00] Hah. And chrooting into that bootstrap and running “apt-get -f install” *does* fix things, so it seems only debootstrap is confused. [07:02] pitti: intended behavior, so debootstrap bug [07:03] micahg: because we still have 14 rdepends on libperl5.12, and they require the 5.12 version? [07:03] No, we don't go that far. [07:03] It's because there are still arch-specific binaries from the old version. [07:03] Because a build from one of the archs hasn't finished, or has failed. [07:04] ah, I just saw the "all" in above paste [07:04] Right, we keep arch-indep publications when there are still arch-specific publications for that source version. [07:04] * micahg isn't sure what's still needed [07:04] Because armel doesn't want to be uninstallable when they're lagging. [07:04] ok, so it's just debootstrap; thanks for the heads-up! [07:04] so it seems live-builder somehow doesn't have that problem [07:05] wgrant: no perl build failures...so is this really supposed to be? [07:05] It's also possible that there's NBS. [07:06] But there's some old binary from that source still published. [07:08] yeah, it's due to libperl5.12 still being around [07:58] morning pitti rodrigo_ [07:59] hey glatzor [07:59] morning rodrigo_ [07:59] how are you? had a nice weekend? [08:00] pitti, weekend shift :) [08:00] but nevertheless it was nice. yourself? [08:01] we went to Dresden and baked some cookies and a fortress (instead of the classical christmas house) [08:01] pitti, nice :) [08:01] and met some friends and family again [08:02] pitti, when I returned from work at Saturday I had the feeling just entering a backery: my girl firend baked 9 pieces of stollen [08:02] wow [08:02] who's going to eat all that? [08:03] hopefully the family and friends of us :) they are really huge [08:04] but it is weired to look at those stolen and having to wait 4 weeks before you can actually eat them [08:04] we have one sitting here as well, can't wait for next weekend [08:10] https://picasaweb.google.com/107564545827215425270/WeihnachtsbackenMitStVo02#5677358274213291218 [08:10] Phear! [08:25] pitti, impressive! [08:32] morning [08:34] hi glatzor, pitti [08:35] morning rodrigo_ [08:42] morning mvo [08:44] hey glatzor! gooooood morning :) [08:44] good morning [08:44] hey didrocks [08:45] hey mvo, how are you? [08:45] hey mvo, guten Morgen [08:45] morning didrocks [08:45] didrocks: tired, but at cup of tea will help me here [08:45] hey pitti! guten morgen [08:45] bonjiour didrocks [08:45] and mvo [08:46] hey rodrigo_! [08:46] good morning glatzor, rodrigo_, guten morgen pitti [08:46] bonjour didrocks [08:51] didrocks: you are the approver of https://blueprints.launchpad.net/ubuntu/+spec/desktop-p-quickly, is that on purpose? if so, could you please have a look at it today? [08:51] pitti: sure, it was still drafting last time I checked, looking === tkamppeter_ is now known as tkamppeter [08:52] didrocks: cool, thanks [08:54] glatzor: I see you already did 0.43+bzr707-1ubuntu1 in ubuntu-precise, I will merge the changes from julian and upload, ok? [08:55] mvo, you could update to the latest snapshot [08:55] mvo, I made the (internal) simulate call async and added some mainloop iteration to make the daemon respond faster [08:56] glatzor: cool, will do [08:58] the worker/downloader separation should also be doable without threading - which will add download while installing support [08:58] glatzor: r712, right? [08:58] glatzor: man, you rock *so* hard [08:58] glatzor: \o/ [08:59] mvo, right r712 [09:00] * mvo prepares the package [09:01] glatzor: any word/opinion on reviews for debian ? [09:02] glatzor: hm, I get a error that it can't find org.freedesktop.PackageKit-aptd.service, maybe a forgoten bzr add ? [09:04] mvo, unluckily it has to be named org.freedesktop.PackageKit.service so python-aptdaemon.pkcompat conflicts with packagekit [09:05] glatzor: aha, so I need to fix the install file only, no problem [09:05] glatzor: meh, that is really a bit unfortuante that they can't be co-installed [09:05] mvo, wait I haven't yet pushed my local changes [09:05] glatzor: aha, ok [09:05] glatzor: shall I wait? [09:05] mvo, done [09:06] mvo, dbus activation seems to require that the service file has got the same name as the to be activated interface [09:07] mvo, gnome-packagekit needs an or dependency on aptdaemon :) [09:07] glatzor: heh, I'm sure rodrigo_ will be excited to hear that [09:07] mvo, python-aptdaemon.pkcompat actually [09:08] glatzor: I pushed my own changes now, I do a testbuild and if its good I will upload [09:08] hey [09:08] hey seb128 [09:08] seb128: aptdaemon/PK goodness from glatzor on the way [09:08] hey mvo [09:08] great [09:10] oh, it's a seb128! [09:10] hey pitti [09:10] seb128: enjoyed the long weekend? [09:10] how are you? had a good w.e? [09:10] yes, I did! [09:10] seb128: yeah, went to Dresden to visit friends and family again [09:11] went to the spas on friday that was nice, and mostly relaxed on satuday and sunday [09:11] seb128: https://picasaweb.google.com/107564545827215425270/WeihnachtsbackenMitStVo02#5677358274213291218 was one of the results :) [09:11] lol, nice one ;-) === popey_ is now known as popey [09:12] glatzor: hm, the system hangs for some seconds now e.g. on cache update in dbus-send org.freedesktop.PackageKit.StateHasChanged string cache-update" [09:12] glatzor: known issue? [09:13] seb128: what is "spass" [09:13] mvo, plurial spa, but maybe it's not a plural in english [09:13] sauna, jacuzzi, etc [09:14] aha, nice! [09:14] ;-) [09:14] mvo, oh yeah evil issue [09:15] glatzor: if evil and known, that is good enough for me for now [09:16] mvo, the script is called by aptdaemon in process so that we cannot receive the signal in the same process [09:20] good morning everyone [09:21] hey chrisccoulson [09:21] hi pitti, how are you? [09:21] chrisccoulson: quite fine, thanks! how about you? [09:22] glatzor: hm, hm, anything we can do about this? (e.g. sending it out of process?) [09:22] pitti - yeah, not too bad thanks. i think i'm finally getting over my ubuflu :) [09:22] good to hear -- that was a tough one! [09:23] mvo, it is called by /etc/apt/apt.conf.d/20packagekit [09:26] seb128: WDYT about re-enabling apport soon? as we don't have an unstable GNOME release, we shouldn't currently have loads of crashes which will automagically fix themselves [09:26] seb128: I'm working on client-side dupe detection this week, and for testing that on a wider scale enabling apport would be handy [09:26] pitti, works for me though I guess what we will collect is mostly duplicates of known issues [09:26] but we can still turn it off in a week if we think it's noisy [09:26] so go for it ;-) [09:26] and those are precisely what I'm interested in :) [09:27] seb128: ok, thought so; checking with Daviey and cjwatson for foundations/server first, too [09:28] glatzor: that isn't a blocker, I think, I will go ahead and upload so that we can enable the or-dep in gnome-packagekit and get testing [09:29] glatzor: plus (most importantly) so that the tools like gnome-control-center are unblocked [09:32] mvo, the issue will only appear if you install packagekit. perhaps adding a small timeout to the script could be a solution [09:32] mvo, the script is called three times by apt on failure [11:55] pitti: I think with the Quickly spec, I gained some new WI and so, I'm above the limit again, I'll drop/postpone some WI as I think in addition to this, other WI will be added as well… [11:56] didrocks: well, it's not a super-hard limit, but you really should be comfortable with the ones you have [11:56] didrocks: also keeping your unity RM duties in mind [11:56] pitti: indeed, rather than deleting WI, I think I'll make "target for opportunities" list in some blueprints so that I can grab them if needed [11:57] seif: hey, what was the status about libroffice zeitgeist datasource? I think you need to talk to Sweetshark about this (people are giving nice feedback on the gedit and totem ones btw ;)) [11:59] didrocks: btw, does quickly take the couchdb EOL into account? [12:00] pitti: indeed, the new quickly-widgets shouldn't dep on it (it's a community volonteer doing it) [12:00] we discussed it at the session, but it's like to port quickly-widgets to gtk3 anyway [12:03] someone please file an apport crash bug [12:03] but nothing crashes in Ubuntu right now, right? [12:04] getting closer to have client-side dupe detection work [12:04] but next episode after lunch, bbl [12:11] didrocks, gedit is getting my plugin upstream [12:11] :) [12:12] great :) === MacSlow is now known as MacSlow|lunch [12:52] seb128: fixed the retracer crashes, whopsie [12:52] pitti, danke [13:12] our sprint in Budapest ends on Friday, 13th. Not that I care too much, but UDS-O ended also on Friday, 13th ... [13:13] that's why we flew out on Saturday :) [13:13] ... and on the flight back I was seated in row 13 (labeled 14) on _both_ flights. [13:13] heh [13:13] * Sweetshark feels flying again on Friday would be pushing his luck .... [13:17] * rodrigo_ lunch [13:49] Sweetshark: you'll notice most airplanes don't actually have a row 13 :) [13:49] sure, they don't have a row number 13, but they have a 13th row (with number 14) :) === MacSlow|lunch is now known as MacSlow [13:54] i'm glad that i'm not superstitious :) [13:55] pitti: well, not necessarily the 13th physical row; some planes skip several numbers between business and coach; so 1-3 are business, and coach starts at 10 [13:55] I had a crazy continental flight with rows numbered 1-12, 14, 20-. I wonder why they had to keep row 14. For tradition? Or because they feared to run into troubled with people explicitly evading row 14, and then being scared by row 20? [13:56] Sweetshark: was 20 the exit row? [13:56] dobey: not sure, but i dont think so [13:58] Sweetshark: 20 was probably exit row, or the row just in front of it (which had immovable seat backs) === mterry_ is now known as mterry [14:20] hum, I should upgrade to precise [14:20] mterry, hey [14:21] mterry, can I abuse from your build fixing time to look at the cairo upload I just did? it builds fine on oneiric but failed to build on precise, the ltmain.sh patch doesn't apply after autoreconf run [14:21] it's likely trivial but I don't have a precise box yet to refresh that ltmain hack (I'm not even sure that's needed nowadays in fact with as-needed) [14:22] seb128, sure [14:22] mterry, thanks [14:35] seb128: ah, we can't merge/sync gnome-bluetooth? [14:38] pitti, we can merge I guess, but I try to clean outdated version first, will do merges later on [14:38] pitti, we can't sync no [14:39] pitti, we have an indicator patch [14:39] seb128: libindicate is in Debian [14:39] pitti, right but I doubt debian wants to transform the systray in an indicator [14:40] tehre's always ubuntu.series, but I guess trying to sync only really makes sense for actual libraries where we are less likely to diverge again [14:41] anyway, I was just curious [14:41] thanks [14:41] that patch is making upstream unhappy, they don't like us patching their code and introducing issues and bugs, I would prefer not to push that change in Debian as well [14:41] yw ;-) [14:41] pitti, if you want to do an update in debian we could sync, libsoup2.4 is on that list ;-) [14:42] queued [14:42] danke! [14:48] Hi. When I insert a SD disk or USB drive and try to open it via the unity launcher it tries to open the folder with the archive manager. How can I fix that? [14:52] tjader, right click on a directory in nautilus, properties, open with, what is selected in that tab? [14:54] seb128: it does not have that tab, but by right-clicking, open with, another application I was able to fix it [14:54] thanks [14:56] tjader, you're welcome [15:00] seif: i forget; to fix a package in proposed (that's already in proposed but verification-failed), do i need to avoid bumping the 0ubuntu2 or should i bump it? [15:00] damnit [15:00] not seif [15:00] seb128: ^^ [15:01] dobey, you can't use the same version twice, you have to bump the number [15:01] ok [15:01] but include both changelog entries on upload (i.e use -v) [15:01] since the previous one didn't go to -updates [15:19] seb128, btw, uploaded refreshed ltmain.sh patch in cairo. works for me in precise [15:33] dobey, seb128 [15:33] for some reason gtk in ubunut wont allow me to draw frames with borders [15:34] whats up with that [15:34] ? [15:34] frames with borders? [15:34] it works fine [15:34] you mean the theme doesn't draw borders for your frames? [15:35] shadows [15:36] dobey, in gtk3 [15:36] seif: using which theme? [15:39] adwaita [15:39] ? [15:39] seif: right, the GNOME 3 theme, and it probably doesn't draw the borders on frames [15:42] pitti: can libreoffice-3.4.4-0ubuntu1 be moved to oneiric now? the one launchpad issue has verification-done and it has 7 days in -proposed. [16:09] Sweetshark: we just got a single piece of feedback for this large update; I'd rather get some more confirmations from Ubuntu (not just one from Kubunut) [16:12] Sweetshark: i thought it was 7 days after verification-done :) [16:12] pitti: can you accept ubuntuone-client into oneiric-proposed? there was a stupid mistake i feel stupid for not catching earlier :( [16:15] dobey: will do soon (unless RAOF or Spammaps beat me to it) [16:16] pitti: ok, thanks [16:30] hmm, the timezone picker seems to have regressed in precise [16:31] took me about 20 clicks to land on Chicago on the map :P [16:43] slangasek: you should try it with orca and keyboard /o\ [16:43] or just keyboard, no mouse [16:43] AlanBell: I don't see how that could be any worse [16:45] pitti: well, the Kubuntu issue cant cause regressions in Ubuntu (libreoffice-kde is not installed in default Ubuntu), or do you mean a just a general "doesnt crash on startup"-verification? [16:46] slangasek: it was hard in oneiric, I will try precise if there are changes in that area, basically it isn't a standard dropdown control and it doesn't really work with keyboard to select the city [16:46] Sweetshark: yes, a smoketest [16:46] AlanBell: right, precise still doesn't have a dropdown; what has regressed in precise is that repeatedly clicking in a zone now no longer rotates through the set of related cities [16:46] I don't know how I managed to finally get 'Chicago', but it's not deterministic :P [16:47] ooh, maybe it has improved then :) [16:49] slangasek, AlanBell: what selector are you talking about? ubiquity? clock applet? [16:49] seb128: clock applet [16:49] well, s/applet/indicator/, I guess [16:49] seb128: ubiquity [16:50] ah, then we're talking about two separate selectors at the moment, I guess [16:50] I will give the clock applet a go, that isn't a dropdown widget either but looks like orca might have a better stab at it . . . [16:50] (planned to be merged this cycle, I think?) [16:52] slangasek, it was I think, but that was before ev and mterry got busy on other things so dunno if that still stands [16:53] ok [16:53] slangasek, can't you just type in the text entry for the location? [16:53] slangasek, seb128: I believe ubiquity currently uses a small library that ev broke out last cycle. indicator-datetime needs to use it too [16:53] there are about 6 Londons in the world [16:54] orca can't read the time and date applet selector very well, but it can be operated with the keyboard [16:56] good night everyone! [16:56] odd fake dropdown widgets really don't help matters for screen readers, it bypasses all the accessibilty stuff that is built into the standard widgets [16:56] 'night pitti [16:58] seb128: "type in the text entry" - how am I supposed to know what it wants me to type? [16:59] slangasek, well it says "Location:", just type a name? [16:59] like Portland [16:59] seb128: the first time I typed 'Chicago', it popped up a little "error" icon; now it works [16:59] weird, maybe the request timeouted or something [17:00] seb128: still, the primary widget presented here is the map, and the 'Location' field gives users no guidance; the map clearly needs to be fixed or dropped [17:00] bug #893197 filed [17:00] Launchpad bug 893197 in gnome-control-center "timezone map no longer allows reliable selection of the correct zone" [Undecided,New] https://launchpad.net/bugs/893197 [17:01] gah, nothing works :( it can see the frames but both orca itself and ubiquity are inaccessible === eeejay_is_afk is now known as eeejay [17:03] so much for testing precise :( [17:04] slangasek, indicator-datetime didn't change since oneiric so that dialog is the same it was in oneiric [17:04] slangasek, but I agree it could be better ;-) [17:05] seb128: it's possible that only US/Central has problems that I'm noticing; the last time I needed that timezone would've been Dallas in January [17:06] slangasek, you are travelling? you didn't get prompted to change your location? [17:06] ... prompted? [17:06] no, I haven't been prompted [17:06] yes, I'm travelling [17:06] bug #802384 should be fixed in precise :-( [17:06] Launchpad bug 802384 in indicator-datetime "not getting prompted to change timezone via geoclue" [Medium,In progress] https://launchpad.net/bugs/802384 [17:06] indeed, it should [17:07] slangasek, it should detect that your location is different from your tz and let you select the tz you are in [17:07] hey kenvandine [17:07] hey seb128 [17:07] seb128: do you mean active geolocation? [17:07] kenvandine, meet slangasek, an unhappy user [17:07] slangasek, meet kenvandine, the maintainer of datetime [17:07] now you guys can sort it ;-) [17:07] :) [17:07] hehe [17:08] seb128, kenvandine: does that mean bug #893197 should be assigned to indicator-datetime rather than g-c-c? [17:08] Launchpad bug 893197 in gnome-control-center "timezone map no longer allows reliable selection of the correct zone" [Undecided,New] https://launchpad.net/bugs/893197 [17:08] slangasek, yes [17:08] ok [17:08] slangasek, g-c-c only loads the datetime .so [17:08] reassigned then [17:08] the code is in datetime [17:08] thanks [17:09] kenvandine: am I misremembering that the datetime widget on the desktop used to allow you to click multiple times in the same location to rotate through the cities within that band? Was that only in ubiquity? [17:10] i don't think the datetime one did that [17:10] ok [17:10] it is pretty hard to select chicago though [17:11] absurdly hard, given that clicking at various points in central time causes zones not under my cursor to be chosen instead [17:11] long term, is this supposed to be replaced with the tz selector from ubiquity? [17:12] ev has dealt with all of these issues in the installer before, and that one works pretty slick [17:12] yeah [17:12] i know we've wanted a shared widget [17:12] the issue is that the installer one is python code, that one is C, I think the idea was to use the datetime one [17:13] but nothing prevent us to "port" the installer code to datetime [17:13] yes, I think that's what's needed here then [17:13] ev and mterry were discussing using the same codebase before Oneiric, not sure which one that was though [17:13] * slangasek doesn't either [17:14] seb128, what's the question? [17:14] mterry: which code was planned to be used for a unified tz map widget [17:14] mterry, what was the plan with indicator-datetime and ubiquity maps? [17:15] mterry, slangasek says the datetime map widget makes it really hard to pick a location, which is an issue the ubiquity one doesn't have [17:15] unless you are using orca, in which case both suck :) [17:16] seb128, kenvandine: as for geolocation, I'm on a plane, all bets are off :P [17:16] seb128, slangasek: ev made libtimezonemap last cycle. I believe ubiquity uses it, but indicator-datetime doesn't. What needs to happen is that indicator-datetime gets ported to use it (simultaneously moving some of its features/fixes to that code into the library) [17:16] aha [17:16] slangasek, ah... then i am sure that wouldn't work :) [17:16] mterry: ok, cool [17:17] slangasek, what's the specific issue that's difficult with picking maps in indicator-datetime? [17:17] kenvandine: in general I don't get warm fuzzies about the idea of being prompted to update my TZ based on geolocation [17:17] mterry: please try to click on Chicago :) [17:17] * mterry tries [17:17] slangasek, it doesn't popup and prompt you [17:17] it just adds a menu item to the indicator [17:17] oh [17:17] to let you switch it [17:17] yeah, that seems fair [17:17] :) [17:18] you have anticipated my curmudgeonly objections, well done ;) [17:18] mterry, it is insanely hard to select chicago... [17:18] slangasek, you want it to change automatically? [17:18] Still don't quite understand. :-/ [17:18] slangasek, or not changed at all? [17:18] I was able to select it in both the map and the locations dialog. Which one of those are we talking about? [17:19] mterry, the map [17:19] * AlanBell could select it on the map [17:19] click just bottom left of the lakes [17:19] it seems very close to others or something [17:19] i felt like i was holding my mouse still and between clicks it was changing [17:20] between chicago and vincennes which are in different timezones [17:20] kenvandine, ah! you're talking about clicking on chicago, not the dropdown [17:20] just since I saw you people talking about the indicator datetime map: the location of Barcelona is way off on that map [17:20] kenvandine, vincennes in france? [17:21] no :) [17:21] it needs to zoom really [17:21] ah :) [17:21] rodrigo_: "not change at all" is my preference, I don't trust geolocation [17:22] not for something like this [17:22] slangasek, right, my preference also [17:22] AlanBell: it wouldn't need to zoom if it would just allow cycling through the cities in the band on multiple clicks (i.e., if it were the ubiquity one :) [17:23] mterry: yeah, it seems to be doing the old and broken "find the city that's geographically closest to where you clicked" heuristic [17:23] except even then, I'm not sure why it thinks Indiana is closer than Illinois :) [17:24] oh! In fact, there appears to be a vertical offset for all of the cities [17:25] so something seems to be misaligned [17:25] fwiw [17:26] anyway, bug filed and my timezone has been set... I don't need to monopolize the channel further :) [17:27] thanks slangasek [17:28] * rodrigo_ out for a bit, bbl [17:28] * AlanBell goes to look where chicago actually is [17:28] London is positioned over edinburgh right now it seems [17:29] slangasek: actually below the red dot is the shaft of the pin, which is sticking in the right place in the map7 [17:30] the dot isn't the city position itself [17:30] ah [17:30] right; not very intuitive to me [17:30] not really, I wouldn't have noticed it myself without trying it with compiz zoom === eeejay is now known as eeejay_is_afk [17:36] wow, completely easy to miss for anyone without perfect eyesight === eeejay_is_afk is now known as eeejay === eeejay is now known as eeejay_is_afk === eeejay_is_afk is now known as eeejay === eeejay is now known as eeejay_is_afk === eeejay_is_afk is now known as eeejay === eeejay is now known as eeejay_is_afk === skaet_ is now known as skaet === eeejay_is_afk is now known as eeejay [18:24] hmmm, really need to figure out scrollbar stuff this week [18:25] oh, seb128, i looked at the shutdown on resume issue last week btw [18:25] for me, it's definitely related to that extra battery, and i tracked down the kernel commit which fixes it :) [18:25] so i don't know if other people have a separate issue [18:26] https://bugs.launchpad.net/ubuntu/+source/linux/+bug/852406/comments/37 [18:26] Launchpad bug 852406 in linux "Phantom battery appears after resume from suspend" [Undecided,Fix committed] [18:26] chrisccoulson, nice, thanks for tracking the kernel commit [18:41] have a good evening everyone === eeejay is now known as eeejay_is_afk === eeejay_is_afk is now known as eeejay === eeejay is now known as eeejay_is_afk === eeejay_is_afk is now known as eeejay [22:47] bg [22:47] opps [23:16] robert_ancell: Hi Robert! Do you have a few minutes to complete https://code.launchpad.net/~gunnarhj/lightdm/lang-chooser/+merge/78226 ? Martin has approved it after a few code improvements. [23:17] GunnarHj, sure [23:18] RAOF: poke [23:18] robert_ancell: Great. I'll be available here for possible follow-up questions. [23:18] smspillaz: Yo! [23:18] RAOF: are there any known caveats with linking to libdrm ? [23:18] Not as far as I'm aware. [23:19] cmake was being a bit weird so I'm just trying to link manually on the command line to see what I can get, but its failing bizzarely [23:19] RAOF: using -ldrm and the implict search path from gcc [23:19] Got a pastebin, or some code? [23:19] it looks like it actually finds libdrm.so but then just refuses to find any symbols in it [23:19] yeah, sure, hang on [23:19] GunnarHj, could you make an upstream merge proposal? It should be good to go upstream now [23:21] RAOF: http://paste.ubuntu.com/745454/ [23:21] robert_ancell: Sure, I can do that. Upstream, is that lp:lightdm? [23:22] RAOF: I'll get the code up somewhere, sec [23:22] GunnarHj, yup [23:22] GunnarHj, what was the conclusion on /usr/share/language-tools/language-options? Is that Ubuntu specific [23:23] smspillaz: Got an extern "C" {} around the include? [23:23] ahh that might foil it [23:23] robert_ancell: Yes. That call won't be there when we get at 12.04. Martin wants it to be replaced by some AccountsService method. [23:24] (I don't) [23:24] Name mangling! [23:24] RAOF: for the record: lp:~smspillaz/compiz-core/compiz-core.test_vblank_waiter [23:24] GunnarHj, is there something we can commit for upstream that is equivalent (and carry the change in a patch for Ubuntu?) [23:25] RAOF: :( you're right [23:25] and to think I spent 5 hours yesterday tearing through the cmake source to find out what was going on [23:25] robert_ancell: Actually it *can* be committed upstream - it won't break anything for non-Ubuntu users, since it falls back to 'locale -a' if the script isn't there. [23:25] i always felt like libraries should do the extern "C" {} bit themselves [23:25] They generally do. [23:26] they generally should [23:26] robert_ancell: Guess it's up to you if you can accept a temporary solution upstream. [23:26] except if you're xf86drm and use non standard include paths and all kinds of other voodoo [23:26] GunnarHj, yeah, I'll have a look. My preference is to only have "standard" things upstream [23:26] RAOF: of course, I could patch it upstream, but I imagine the response would be something like [23:26] "C++ sucks" [23:27] I'm actually a bit surprised; a fair amount of mesa code is C++. [23:27] I know [23:27] RAOF: well, libdrm != mesa [23:27] you lmean like http://cgit.freedesktop.org/mesa/drm/commit/?id=cfee5218b17a2741e5519ed44091171e01f0dbb2 ? [23:27] though its distributed under the same namespace [23:28] Sarvatt: if only .... :( [23:28] robert_ancell: thanks for sending the email to me. I'll get that to Rick. RAOF, at the end of the day could you send the auto-testing recap to me as well? thanks! [23:28] its in libdrm 2.4.27 [23:29] Yeah, I was probably looking at an older checkout. [23:29] (in precise) [23:29] jasoncwarner_: Certainly. [23:29] robert_ancell: Ok. I wait with the upstream MP then. [23:29] RAOF: thanks! [23:29] auto-testing? [23:29] GunnarHj, no, just do the MP, I can remove things from it and commit them. So then you do a merge and have a smaller diff [23:30] robert_ancell: Aha, sure. Then I disconnect for a while, because I have Precise on a separate partition. [23:31] GunnarHj, ok [23:32] bryceh: if that was directed at me, then, yes, it was for auto-testing [23:32] I'm trying to get the vblank code that vanvugt has done tested outside of an opengl implementation [23:32] smspillaz, ah [23:33] the hilarious thing is going to be when everyone on DX runs nvidia and can't run this test [23:48] * smspillaz slaps himself [23:48] I need to stop using unsigned int for everyhting [23:48] *everything [23:49] Doing some signed compares? :) [23:50] Bah. Dear cmake: I really didn't change any files outside of test/. Why are you rebuilding everything? [23:50] RAOF: easybuild! [23:50] Bake! [23:50] * RAOF wonders if Robert has that on highight :) [23:51] * desrt wonders if robert_ancell has 'robert' on highlight [23:51] oops. [23:51] desrt, I sure don't [23:51] Nice work :) [23:51] robert_ancell: easy-build development looks stagnant since last time we talked about it [23:51] is that an accurate reflection of reality or am i looking at the wrong project? [23:51] desrt, yes, been busy :( [23:52] fair enough [23:52] robert_ancell: i've been selling you to everyone i talk to [23:52] desrt, oh damn. Now I need to deliver ;) [23:52] robert_ancell: i did a recent talk at FSCONS about how to make a gnome application [23:52] it was embarassingly difficult [23:52] particularly setting up po/ and intltool [23:53] desrt, yeah, there's a mountain of crap to climb [23:56] smspillaz: to sign or to unsign is one of the endless debates [23:56] there are two schools of thought on the subject [23:57] i think they're both wrong [23:58] insofar as i don't think it's possible to come to *any* sane conclusion regarding what is essentially a completely broken system