[00:46] Hrmmm. Seems the nexus 7 daily images are not very zsyncable. [02:03] anybody here have any experience setting up a project to use valadoc? === hggdh_ is now known as hggdh [03:14] RAOF, jasoncwarner, hope you guys are keeping comfortable today. [03:15] * RAOF has just been offered an ice cream. He's pretty comfortable. [03:15] If its refreshing, thats all that matters I guess. [03:18] Hi all [03:19] I have a Radeon HD 6470M Graphics card are there drivers for it on ubuntu [03:25] cheebu: Yes; you're after #ubuntu for support questions, though [03:37] smspillaz: Hello? === Optichip` is now known as Optichip [04:16] duflu: howdy [04:17] sorry for the late response, I found on old TV on a verge and can finally play skyward sword after my tv was broken for about a year [04:19] desrt: do you know what the most sensible approach for projects using glib + gcc 4.8 are at this point ? [04:19] just detect the compiler and disable the warning? [04:19] I believe its PS policy to use -Werror, as much as I dislike it [04:20] desrt: FWIW, pretty much every other library that uses static asserts in its headers got screwed by this change too :) [04:20] boost, v8 etc etc [04:23] smspillaz: I was going to mention that I will be aiming to tag a 0.9.9.0 within hours/days. Only really bad regressions need to be squeezed in. The rest can go for 0.9.9.2. [04:23] duflu: sure thing. do compiler errors count as bad regressions ? [04:23] Also, I will be at a sprint from the end of next week and all the following week. [04:23] lovely [04:24] smspillaz: Only if it's the compiler version native to that release [04:24] duflu: does clang count ? [04:24] smspillaz: Depends on triviality of the fix and whether there is one already [04:24] only if you're running on BSD :) [04:24] duflu: hrm, I didn't know they were doing sprints in the mid-year [04:25] I thought they had canned the whole "lets fly the company everywhere" policy by the time I left :) [04:25] smspillaz: *shrug* [04:25] * duflu goes to finish lunch and dishes [04:26] duflu: can you check if either dbarth or iftikhar amed are going? I have a laptop to give them [04:26] and I can't post it because FedEx hates me [04:26] smspillaz: Yes dbarth confirmed. Drop it over I guess [04:26] * duflu goes afk === pitti_ is now known as pitti [05:39] Good morning [05:39] and happy new year everyone! [06:26] good morning! [06:29] Morning didrocks, pitti [06:29] hey duflu, how were your holidays? [06:30] didrocks: Very busy with Christmas and home improvement. But sleeping in was excellent. You?? [06:30] bonjour didrocks, je tu souhaite une bonne année! [06:30] hey duflu, how are you? [06:31] Good thanks pitti [06:31] How was the holiday(s)? [06:31] guten morgen pitti! Bonne année à toi aussi :) [06:31] duflu: was good as well, even had the opportunity to take a day for skying :) [06:31] I enjoyed ours very much; we'd been in Dresden [06:32] didrocks: what is skying? you had a parachute, I hope? [06:32] Heh, didrocks 007 [06:32] pitti: duflu: ok, always trapped by this one :p "go skiing" :) [06:32] ooh [06:33] English sucks. And don't forget vacuum [06:34] :) === Noskcaj is now known as Noskcaj_afk === Noskcaj_afk is now known as Noskcaj [07:29] hey Laney, how are you? [09:04] morning! [09:05] hey Laney! [09:05] hey [09:05] how's it going? well rested? :-) [09:06] not that much in fact, I couldn't disconnect mentally weirdly [09:06] even if I didn't work [09:06] well, I still have a week of skiing at the end of the month, so I'll use that :) [09:06] and you? how were your holidays? [09:07] good, spent some time with my family and then with my girlfriend's family [09:07] then went to the countryside in wales for new years [09:07] sweet :) [09:10] I received a record player for christmas too so I've been becoming a vinyl nerd :P [09:10] ahah, excellent! big collection of vinyls? :) [09:15] not yet, but hopefully soon! [09:16] * Laney spies much unread email [09:17] Laney: FYI, ubuntu-themes is automatically merged [09:18] Laney: so no need to merge yourself (actually it's worse because you have to bump the version yourself, and we don't have automagic changelog) [09:18] didrocks: aha, we did wonder about that [09:19] I was asking what to do about the changelog at the time [09:19] Laney: wasn't cyphermox around? [09:19] infact I don't know how to work with these auto packages any more really [09:19] he should be knowledgeable about it :) [09:19] I think I spoke with jibel but he was also unsure [09:20] Laney: ok, cyphermox should have it picked up and looked at the merge, anyway, no worry for this one, it's more a FYI :) [09:21] thanks, the information is good [09:21] is there documentation for how to work with these packages? [09:22] Laney: not yet, but normally the person responsible for the stack should pick it up when looking at the merge [09:22] Laney: I will check that cyphermox is looking at that one [09:23] not sure if he is or isn't - in this case it was because the contributor pinged in here so I thought I'd look at it [09:23] it's good for any developer to be able to make changes if they need to anyway [09:24] Laney: oh, they can and even upload [09:24] Laney: it's blocking the daily build if it's not synced [09:24] but as we have a merger, it's better to use it :) [09:26] so all that needs is an approve vote on the MP? [09:26] and are MPs expected to update d/changelog or is that done automatically? [09:30] Laney: approve vote and change the globale status to "approved" [09:30] Laney: if a bug is attached to a MP, the bug title will be use in debian/changelog [09:30] linked to the author of the merge [09:49] Happy new year everybody, I have an ubuntu-themes related question [09:49] https://bazaar.launchpad.net/~ubuntu-art-pkg/ubuntu-themes/trunk/view/head:/Ambiance/gtk-3.0/apps/gnome-terminal.css [09:49] does anybody know why all that file is commented out? [09:55] Hi, does anyone know if ubuntu for phones will be libre software? [10:09] jjardon__: #ubuntu-phone is the place for questions about this [10:35] Laney: thanks === tkamppeter__ is now known as tkamppeter [11:47] pitti, hi [12:00] hello tkamppeter, gesundes Neues! [12:15] hi pitti, how are you? [12:15] hey chrisccoulson, happy new year! [12:15] hey chrisccoulson! happy new year :) [12:16] pitti, didrocks, happy new year to you too :) [12:17] chrisccoulson: had some nice holidays? [12:19] pitti - yeah, it turned out ok, although i'm kinda glad it's over now. i had flu for the first week, and then 2 days before christmas we had 2 appliances fail - 1 of them being our refrigerator :( [12:19] so it's been an expensive few weeks ;) [12:19] urgh [12:19] I just recovered from a cold, but it was a light one [12:20] was nice to see family and friends again in Dresden [12:26] pitti, dir auch, danke. [12:26] pitti, I am trying to get this with dh_installinit working, but I have a problem. [12:29] pitti, Axxording to the man page one has to call it with "dh_installinit --name=cups", but I do not get the build process calling it this way, make targets "override_dh_installinit/cups-daemon:" or simply "override_dh_installinit:" get simply ignored. [12:29] tkamppeter: in postgresql-common I just name the file debian/postgresql-common.postgresql.init [12:30] which ends up as /etc/init.d/postgresql.init [12:30] and call it with dh_installinit --name=postgresql -u'defaults 19 21' -r [12:31] tkamppeter: oh, it's just override_dh_installinit: if you are using dh7 [12:31] tkamppeter: i. e. use the existing target [12:31] and just add the --name=cups [12:32] pitti, so debian/rules must contain: [12:32] override_dh_installinit: [12:32] dh_installinit --name=cups [12:32] Is this correct? [12:33] tkamppeter: no, we also still need the existing "upstart or not" magic [12:33] tkamppeter: just add it to the existing rule === Quintasan_ is now known as Quintasan [12:40] mmm, cups [12:40] oitti, where is this "upstart or not" magic? There is a common-install-prehook-impl:: which places the upstart file in the debian/ directory only on Ubuntu, but this should also work with the new cups-daemon package. [12:40] * ogra_ gets coffee [12:42] pitti: ^^ [12:45] tkamppeter: in override_dh_installinit: in debian/rules [12:45] tkamppeter: I'm looking at current git head, are you not? [12:45] tkamppeter: sounds like you are looking at a cdbs debian/rules [12:46] tkamppeter: oh indeed, raring has a different packaging [12:46] pitti, I am in the Raring package, cups_1.6.1-0ubuntu13 [12:46] seems between raring and debian git there's quite some divergence [12:46] tkamppeter: DEB_DH_INSTALLINIT_ARGS, just add it there [12:48] pitti, Debian did not upload cups for months because of the test suite not working. So I have done several releases of cups Ubuntu-only. I have always added my changes to the GIT, but perhaps not taken some recent GIT changes to Ubuntu. [12:49] tkamppeter: well, it's equally bad for the tests to not succeed in Ubuntu :) [12:49] before you know it they will show a really important failure, and nobody will notice [12:49] pitti found the DEB_DH_INSTALLINIT_ARGS place, so I add the --name=cups there. [13:01] pitti, thank you very much, it is working correctly now. I had overlooked the init magic in the beginning of debian/rules. [13:17] tkamppeter: great! (sorry, was off to lunch) === davidcalle_ is now known as davidcalle [14:40] didrocks: which merge did I miss? [14:41] hey cyphermox! [14:41] cyphermox: https://code.launchpad.net/~ubuntu-art-pkg/ubuntu-themes/trunk [14:41] https://code.launchpad.net/~mitya57/ubuntu-themes/remove-desktop [14:42] cyphermox: how are you btw? how were your holidays? [14:42] good, good [14:43] cyphermox: btw, I think you noticed there is another recent merge on ubuntu-themes (but maybe it needs more digging) [14:43] yeah [14:44] cyphermox: also FYI, I finally added the commit message for ido and approved https://code.launchpad.net/~robert-ancell/ido/lp582985/+merge/139791 [14:44] cyphermox: if you see things are forgotten or not moving, feel free to do this kind of things [14:44] ok [14:44] smspillaz: "don't use -Werror" [14:44] smspillaz: pretty simple universal advise for everyone. [14:45] you didn't get 'screwed' by a GCC change. you got screwed by your use of -Werror. [14:45] cyphermox: let's discuss on Monday how the indicator stack progressed between you and mterry looking at it the week before my holidays :) [14:46] hey desrt! [14:46] good holidays? [14:46] didrocks, !!! [14:46] didrocks, how are you? [14:46] hey larsu!!! ;) [14:47] larsu: I'm good, thanks! yourself? [14:47] didrocks, good, thanks [14:47] didrocks, is this your first day back? [14:47] larsu: right, and you? when did you get back? [14:47] I didn't manage to disconnect mentally during holidays though, so looking forward for this week of skiing at the end of month :/ [14:48] hey desrt, hello cyphermox! happy new year! [14:48] hey pitti, happy new year [14:50] didrocks, two days ago, but I haven't been very productive yet :) [14:50] didrocks, lots of catching up [14:50] hi cyphermox! [14:50] * pitti hands didrocks a big "off" switch for the brain [14:50] hey larsu, wie gehts? [14:50] catch up> OMGpainyes [14:50] * didrocks grabs it gladely and will use it next time :) [14:50] took me 5 hours to get my ubuntu mbox down to 2 and my lists folder < 30 [14:50] larsu: heh, here as well :) [14:51] pitti, gut danke. Et toi? [14:51] larsu: prima, danke! [14:51] :) [14:53] desrt: its not really my decision sadly [14:54] desrt: plus, -Werror is handy ... sometimes [14:54] desrt: and to be fair, we got screwed by a gcc change causing some broken header file to stop compiling :) [14:54] our code isn't broken :) [14:55] now it just happened there were a lot of broken header files out there [14:55] the header file isn't broken either [14:55] it's perfectly conforming to the C language specification [14:55] desrt: if the compiler is warning about an unused typedef then you should probably fix that [14:56] however pendantic it seems [14:56] the compiler (incorrectly) assumes that there is no reason to ever have unused typedefs [14:56] it knows that it may be wrong [14:56] that's why it's only a warning and not an error [14:56] desrt: I think the warning is stupid too [14:56] but its still a warning [14:56] and it still breaks anyone who uses -Werror [14:56] you ask for the unreasonable behaviour when you give -Werror [14:56] and i have no pity :) [14:56] no, we ask for the compiler to make us fix bad behaviour and not let it slip by [14:57] its not exactly FuckItJS [14:57] (which I read about the other day, and is like -Werror on steroids) [14:57] my point is that not every warning is a case of 'bad behaviour' [14:57] desrt: well, we could turn the warning off [14:58] desrt: unfortunately, in a marvel of gcc stupidity, its a new compiler option [14:58] if you fancy playing whack-a-mole :) [14:58] smspillaz: -Wno-error=.... [14:58] desrt: -Wno-error=unused-typedefs doens't work with old compilers [14:58] are you sure? [14:58] because .......... oh wait, that option never existed [14:58] yep, I've tried it [14:59] it also doesn't work with clang, which the powers the be insist that everything also compiles with [14:59] oh christ that's stupid [14:59] desrt: in any case, gnome will make a lot of people happy by just fixing the warning :) [15:00] and save N maintainers from having to add crappy compiler detection code to their buildsytem so that they can turn off a warning :) [15:01] i agree [15:01] desrt: I also concur that gcc would make N * M people happy if they didn't warn about unused typedefs in the first place :) [15:01] problem is, it's difficult to fix [15:01] desrt: yeah, I saw :( [15:01] for the same reason that you'd have to add compiler detection, so would we [15:01] and we'd have to figure out how to do it at compile time....... [15:01] yeah I saw, it sucks [15:02] what was that whole thing C++ having static_assert ? [15:02] C++11 mind you [15:02] I wonder if C11 got anything similar ...... ;-) [15:03] desrt: I wonder if it would make more sense to just put pressure back on the gcc folks to not warn about this [15:03] smspillaz: up to you :) [15:03] i guess you need a fix to come faster than they will reply, though [15:03] heh yeah [15:03] *sigh* [15:04] the alternative is that glib could stop caring about different compilers being used at glib-build and glib-user-build times [15:04] although in our new gcc/clang world, that seems kinda dangerous [15:04] I think this is why I've grown tired of compiled languages [15:04] or at least, languages with slightly varying compilers [15:04] and no integrated build system [15:04] * smspillaz misses go [15:05] if go is even vaguely successful in the medium/long term you will see more than one compiler, practically by definition :p [15:08] heh, it seems that smspillaz has the same level of love for gcc as i do atm ;) [15:10] desrt: hopefully not, the go runtime and tools cover a lot of functionality [15:10] that's just like saying there will be more than one ja---- oh wait never mind [15:11] looks like _Static_assert came in 4.6.0 [15:12] desrt: in other "other things breaking my build" news, we had a mesa version bump where they changed a function signature so that it wasn't compliant with the spec [15:12] and then didn't change anything else in the header file to indicate the change [15:13] C is the language of insanity [15:13] i'm fixing glib [15:13] desrt: woohoo! [15:14] now I just need to get boost to fix their headers [15:14] boost, possibly sigc ... [15:17] I'm looking at the FTBFS of libvirt-glib and as it's in the desktop-extra set, I'm asking for an opinion: the FTBFS is due the multi-arched Python and I wonder if should disable the build of the python module which isn't packaged at all or fix the build nonetheless (I will mail the DD about it in any case) [15:19] geser: fix. it's easy, there are three ways to do it. wiki.debian.org/Python/MultiArch [15:20] xnox: even if the module doesn't get included in any deb? (I've a patch ready as this way my first idea just to find out that no deb includes the module) [15:21] geser: I thought it is used, via gobject-interspection by utah for example. [15:22] although it's python-libvirt which they use. so no idea. [15:26] xnox: it would by the python module for libvirt-glib not libvirt itself [15:36] smspillaz: so uh.... ya [15:37] i have a patch [15:37] but it will cause trouble with another set of compiler flags [15:37] (specifically, -pedantic) [15:38] turns out, gcc is a big stupid head [15:39] it warns that this code is not valid ISO C because _Static_assert is not available everywhere: [15:39] #if _GCC_ > 4.6.0 [15:39] _Static_assert(...) [15:39] #end [15:40] seems like, pedantically speaking, GCC should not be defining _GNUC_ when using -pedantic... sigh. [15:40] hahha [15:40] desrt: well, we don't use -pedantic, but I can see how that would be a problem [15:41] ya... i don't watch to push a patch that helps one user or a ridiculous set of flags while huring another user with a different ridiculous set of flags [15:41] *of [15:42] desrt: well, I think in this case, you would do #if _GCC_ > 4.6.0 _Static_assert(....) #else // old behaviour #endif [15:42] in that case, problem fixed for everyone not using -pedantic [15:42] yes. that is what i'm doing. [15:42] and then everyone who uses -pedantic gets the old behaviour [15:42] but some people are using -pedantic [15:42] and, as of now, this works. [15:42] no... [15:42] desrt: that being said, I think you might find that -pedantic doesn't even have the warning anyways [15:43] the problem is that gcc still defines _GNUC_ to the proper version, even with -pedantic [15:43] smspillaz: i just checked... [15:43] at least gcc 4.7 does [15:43] desrt: argh [15:43] FFS [15:43] does it define anything else when you use -pendantic ? [15:44] i don't want to play this kind of whack-a-mole [15:44] neither do I :( [15:45] anyway... there are other issues too [15:45] see the bug [15:45] desrt: yeah, just saw [15:45] desrt: well, we can just detect the compiler compiz-side and disable the warning [15:46] maybe in the release notes it might be worth putting something about gcc bone-headedness [15:46] in slightly more diplomatic language [15:46] do you actually release tarballs with -Werror? [15:46] pitti: every build for all PS projects uses -Werror [15:46] some projects might want to use it for building from VCS, but for "make dist" it really shouldn't be on by default [15:46] that's a sure way to say "f*** you" for pretty much anybody packaging it [15:47] pitti: that's a good point [15:47] pitti: or jhbuilders [15:47] which is the 'build from VCS' case, really [15:47] pitti: maybe it might be worth turning it on in CI only now that we actually have a CI [15:47] in one of my projects I had some configure magic to flip some -Werror=.. stuff on or off depending on whether it was a release or git; if I could remember which one.. [15:48] desrt: yeah, wouldn't help with that [15:48] its a massive PITA for me locally too - having to do printf ("id: 0x%x\n", (unsigned int) window->id ()); is just annoying when Window is typedef unsigned long [15:48] FWIW, I fully agree that -Werror is not what most people want; gnome-common's set of -Werror=... is much more sensible [15:48] well, that's the kind of type check which actually should be on [15:49] pitti: do you have the relevant link to that ? [15:49] but e. g. turning deprecation warnings into errors is just wrong for releases [15:49] pitti: right, we disable deprecation warnings at least [15:50] smspillaz: ah, there: http://cgit.freedesktop.org/upower/tree/configure.ac#n110 [15:50] pitti: if you can grab a link to the gnome-common bit I'll post a message to unity-dev suggesting that we use those -Werror flags and only do it on CI builds [15:50] pitti: great, thanks [15:50] smspillaz: that's not gnome-common, that's "use -Werror when building from git" [15:50] pitti: ah okay. Do you have the gnome-common link ? [15:51] last one I saw was https://mail.gnome.org/archives/desktop-devel-list/2012-November/msg00001.html [15:51] smspillaz: uh... [15:52] desrt: uh ? [15:52] smspillaz: you know that not doing that printf cast would break your code on x86_64, right? [15:52] desrt: its for debugging purposes only [15:52] I never actually release code like that [15:52] how does gcc know that? :p [15:52] desrt: it bitches and won't compile my code, at least with its current compiler flags :) [15:53] smspillaz: http://git.gnome.org/browse/gnome-common/tree/doc/usage.txt, but you probably don't want to depend on gnome-common itself for compiz? [15:53] pitti: sure, just starting a discussion on which warnings we should turn into errors [15:53] pitti: fortunately, compiz has pluggable build systems [15:53] he can add this to the gnome-flavoured-automake backend [15:53] desrt: it actually used to [15:53] ... [15:53] no, really [15:53] wow [15:53] smspillaz: they are in http://git.gnome.org/browse/gnome-common/tree/macros2/gnome-compiler-flags.m4 [15:53] you used to be able to use either cmake or autoconf [15:54] desrt: I believe a lot of vodka was behind that decision [15:54] mmm vodka [15:54] smspillaz: that's how i feel about most of compiz :) [15:54] I know a few former developers were borderline alcoholics [15:55] desrt: well, its helped and hindered in places [15:55] desrt: I think a large part of the reason why compiz got complicated fast was because it wanted to be the one window manager to rule them all [15:55] a noble ambition, yet one that is basically impossible [15:56] indeed [15:56] things would be much better though if we just had one window manager [15:56] hence the reason why I'm not going to "port" it to be a wayland compositor [15:59] we do only have one window manager [15:59] (and its various forks) [15:59] desrt: when I said "we" I included everyone who develops on top of X11 [15:59] not just gnome :) [16:00] okay. one good window manager, then? :) [16:00] desrt: though, if I were to just include gnome, then I would think that the fact that you have people using libmutter is an achievement [16:00] desrt: well, I think the kwin and xmonad people would disagree :) [16:00] meh. [16:01] x is a dangerous and scary place [16:01] and i only know of one window manager that was written by havoc pennington and owen taylor :p [16:01] I never understood why the desktop scene is much more heavily fragmented than everywhere else in linux [16:01] ?! [16:02] you mean like the one unified webserver that we all use? [16:02] or how everyone agrees on scripting languages, shells and editors? [16:02] well, we all use the same kernel [16:02] yes... everyone using linux is using the same kernel [16:02] ...except those who aren't (android) [16:02] isn't android much closer to mainline these days ? [16:03] if you don't define yourself directly into that conclusion, however, you notice the various BSDs and solaris pretty quickly, though [16:03] of course [16:03] I guess what I meant was [16:04] what frustrates me the most about of the linux ecosystem is that when fragmentation occurrs, it often results in duplication of effort on disproportionate scales [16:04] the free software mantra: "anything worth doing is worth doing twice" [16:04] exactly that [16:04] I hate that [16:04] if ever there is only one of something it's because it's such a shitty job [16:04] like samba :) [16:04] i mean... thank god it exists... [16:04] but i'm sure glad i didn't have to write it :p [16:05] I think the entire fact that that we have to have specifications for window managers pretty much indicates how broken the entire process is [16:05] because usually, the specification means "the implementation has to do X" [16:05] i'll say again [16:05] 11:01 < desrt> x is a dangerous and scary place [16:06] desrt: by "x" you are referring to X11 right ? [16:06] yes [16:06] *sigh* yeah [16:06] i think you'd agree that writing a window manager is a non-trivial affair :) [16:06] yes, I'd agree [16:06] particularly one that supports modern notions of (for example) focus stealing prevention [16:06] ditto reparenting [16:06] ditto compositing that isn't slow [16:07] ditto damage handling [16:07] etc etc [16:08] what I'd really like to see is one window manager which handles all of the client <-> WM stuff and then a means to inject policy into it [16:08] I think libmutter is awesome, because its part of the way there [16:11] damage isn't so bad [16:11] non-slow compositing needs only a few tricks [16:11] but ya... most of the other stuff sucks [16:15] desrt: don't forget stacking ;-) [16:15] i consider that part of focus-stealing prevention [16:16] although they're two very different parts of the same beast, admittedly [16:16] very different [16:16] one is you deciding what to do and telling X [16:16] the other is you trying to wrap your brain around the insanity that results from that (and any other random events that may occur at the same time) [16:16] desrt: and override redirect windows [16:16] those are boring :) [16:17] the bane of my existence [16:17] well [16:17] they're boring for a window manager [16:17] interesting for a compositor [16:17] and confusing for a windowmanager/compositor [16:17] and infuriating for a maintainer [16:18] desrt: btw, damage tracking becomes a lot more fun once you implement support for GLX_EXT_buffer_age [16:18] "fun" [16:19] fair enough [16:19] i never tried to mix GL with damage [16:19] * desrt always XRendered [16:19] (and usually just via cairo, much less) [16:20] desrt: well, the main problem at least is that if you draw outside the damage region, you die [16:21] neat [16:21] in classical systems you obviously get to be as sloppy as you want [16:21] yep === sil2100_ is now known as sil2100 [17:47] When's there going to be a Raring PPA for the Canonical Qt5 Edgers team? Will we have to wait until April? [18:06]