| jasoncwarner_ | bryceh robert_ancell RAOF anyone who is up at the moment... | 03:17 |
|---|---|---|
| robert_ancell | yo | 03:17 |
| jasoncwarner_ | what would be the ubuntu-bug command for someone to use if htey are having issues with hibernate or suspend | 03:17 |
| jasoncwarner_ | ubuntu-bug somethingorother ? | 03:17 |
| jasoncwarner_ | ;) | 03:17 |
| robert_ancell | hmm, I'm not sure | 03:19 |
| jasoncwarner_ | ok...gonna have to do some digging...thanks... | 03:22 |
| jasoncwarner_ | I guess someone has to issue against the kernel, but 'ubuntu-bug kernel' isn't happy with that... | 03:23 |
| jasoncwarner_ | I'll just have people file it against X...after all, it is probably X related...cool with that bryceh and RAOF ? | 03:24 |
| RAOF | jasoncwarner_: ubuntu-bug linux | 03:28 |
| RAOF | I'm probably too late, but hibernate and suspend are almost uniformly kernel issues, and “linux” is the source package there. | 03:28 |
| jasoncwarner_ | RAOF: Too late! I just posted on twitter and Google+ for everyone to direct hibernate, suspend, unity glitches and weird color bugs to X ;) | 03:34 |
| jasoncwarner_ | j/k | 03:34 |
| jasoncwarner_ | thanks! appreciate it | 03:34 |
| RAOF | :) | 03:37 |
| pitti | Good morning | 04:20 |
| pitti | Ampelbein: hm, it built locally; presumably it behaves differently with -B, I'll try that here | 04:21 |
| pitti | ah, cyphermox got there before me, thanks! | 04:22 |
| cyphermox | pitti: yeah | 05:00 |
| cyphermox | pitti: was nice to be able to look into it; very interesting rules file :) | 05:00 |
| pitti | cyphermox: sorry for overlooking this; I didn't try with -B | 05:00 |
| pitti | cyphermox: heh, FSVO "interesting".. | 05:01 |
| cyphermox | fsvo? | 05:01 |
| cyphermox | oh | 05:01 |
| cyphermox | right :) | 05:01 |
| pitti | "for some value of" | 05:01 |
| cyphermox | also, -B? | 05:01 |
| micahg | jbicha: are you following about about the transmission depwait on libnatpmp-dev (needs an MIR) | 05:02 |
| jbicha | micahg: yes, I'm working on the mir now :-) | 05:03 |
| micahg | jbicha: great, thanks | 05:03 |
| pitti | cyphermox: of dpkg-buildpackage, binary-only build | 05:07 |
| pitti | cyphermox: -b builds arch:all as well, -B does not build them | 05:07 |
| cyphermox | yeah | 05:07 |
| pitti | cyphermox: our i386 buildds do -b (they build the arch:all ones), all other arches do -B | 05:08 |
| cyphermox | wasn't sure which app you were referring to | 05:08 |
| cyphermox | I usually always use bzr bd to build stuff (or at least, whenever possible) | 05:08 |
| cyphermox | ugh, some days NM is such a an annoyance ;_; | 05:08 |
| pitti | robert_ancell: hey Robert, how are you? | 05:08 |
| robert_ancell | pitti, hi, good | 05:09 |
| pitti | cyphermox: works there, too (bzr bd -- -B) | 05:09 |
| pitti | robert_ancell: I recently noticed an /usr/share/xsession/default.desktop in my system -- is that actually supported by lightdm? | 05:09 |
| pitti | robert_ancell: I wasn't sure where that was coming from, it might just have been a leftover from local experimentation | 05:10 |
| cyphermox | pitti: yeah; but i was to build all the packages, not just arch:(^all) ;) | 05:10 |
| robert_ancell | pitti, no, where did that come from? | 05:10 |
| pitti | cyphermox: right, but that bug only exposed itself with -B | 05:10 |
| jbicha | miniupnpc recommends minissdpd, do I have to do a MIR for recommends? | 05:10 |
| pitti | jbicha: yes | 05:10 |
| pitti | as they get installed by default | 05:10 |
| pitti | robert_ancell: ok, so probably that was my doing | 05:10 |
| pitti | robert_ancell: I was pondering bug 806064 | 05:10 |
| ubot2 | Launchpad bug 806064 in lightdm "lightdm needs something like gdm-set-default-session" [High,New] https://launchpad.net/bugs/806064 | 05:10 |
| pitti | robert_ancell: obviously we don't want other packages to change lightdm.conf | 05:11 |
| pitti | robert_ancell: but https://launchpad.net/ubuntu/+spec/desktop-o-cd-localization has a WI for you "implement lightdm support for /usr/share/xsessions/default.desktop symlink" | 05:11 |
| cyphermox | dear NM, please complete establishing ipv6 when ipv4 settings are applied, kthxbye. I 'd really like to see you connect within less than a minute... | 05:11 |
| pitti | robert_ancell: and once we have that, edubuntu could just install that symlink? | 05:11 |
| robert_ancell | pitti, yes, I'll do that then. | 05:12 |
| pitti | robert_ancell: or is there a better way to set the default session? | 05:12 |
| robert_ancell | No, I think that's best | 05:12 |
| pitti | robert_ancell: gdm-set-default-session worked because custom.conf wasn't a conffile | 05:12 |
| pitti | but lightdm.conf isn't, so in practice it's actually not all that useful | 05:12 |
| pitti | erm, ligthdm.conf _is_ | 05:12 |
| pitti | robert_ancell: i. e. if there's no explicit default in lightdm.conf, it'd check default.desktop first? | 05:13 |
| robert_ancell | pitti, did you respond to the idea to package the conf file for each derivative? Does that make sense? | 05:13 |
| pitti | oh, I didn't see that question | 05:13 |
| micahg | bug 799754 | 05:13 |
| ubot2 | Launchpad bug 799754 in lightdm "Please let vendors easily provide their own config" [High,Triaged] https://launchpad.net/bugs/799754 | 05:13 |
| robert_ancell | LightDM needs a hardcoded default, it's currently gnome, but I think changing it to "default" makes more sense. So a symlink would work (or you could make an actual default.conf) | 05:14 |
| pitti | robert_ancell: how would that look like, /etc/lightdm/local.conf or so? | 05:14 |
| pitti | which overwrites settings from lightdm.conf? | 05:14 |
| robert_ancell | pitti, no, there'd be a lightdm-config-ubuntu package | 05:14 |
| robert_ancell | and a lightdm-config-xubuntu | 05:14 |
| robert_ancell | they'd have to conflict with eachother | 05:14 |
| pitti | robert_ancell: conffiles are pretty sticky, though | 05:14 |
| pitti | removing a package doesn't remove the conffile | 05:14 |
| robert_ancell | that's fine | 05:14 |
| pitti | and if it's already there, and you install a new package, you'd get a nasty prompt | 05:14 |
| pitti | I think it'd be easier to do with a local.conf | 05:15 |
| robert_ancell | that's also acceptable - it's not "normal" to switch derivatives | 05:15 |
| robert_ancell | LightDM doesn't need a config file, so the "lightdm" package really shouldn't provide one, so where do we put our config? | 05:16 |
| pitti | robert_ancell: the main thing that isn't commented there is [GuestAccount]; do these also have internal defaults/ | 05:16 |
| pitti | ? | 05:16 |
| robert_ancell | yes, everything has defaults | 05:17 |
| pitti | (I guess so, as making the scripts configurable doesn't sound too useful actually) | 05:17 |
| pitti | robert_ancell: so, just drop it from the package then? | 05:17 |
| robert_ancell | pitti, so where does the ubuntu one come from then? | 05:17 |
| pitti | robert_ancell: then xubuntu can create theirs as a config file in xubuntu-default-settings | 05:17 |
| robert_ancell | is it baked into the image? | 05:17 |
| pitti | robert_ancell: do we need one? | 05:17 |
| robert_ancell | pitti, yes, we need to set the guest account settings | 05:18 |
| pitti | robert_ancell: I thought the ubuntu flavor by and large should just install the greeter it wants, but that can happen with seeding? | 05:18 |
| didrocks | good morning | 05:18 |
| pitti | robert_ancell: I'm confused -- robert_ancell | yes, everything has defaults | 05:18 |
| pitti | hey didrocks | 05:19 |
| didrocks | guten morgen pitti! | 05:19 |
| pitti | robert_ancell: the guest account settings don't sound very flavor specific to me? | 05:19 |
| robert_ancell | pitti, yes, but guest session is disabled by default | 05:19 |
| robert_ancell | perhaps I should make it work if the startup script exists | 05:19 |
| pitti | robert_ancell: if we are going to enable it by default in all conffiles, why not just enable it in the package? | 05:20 |
| pitti | then we can ship without a config, and xubuntu-default-settings etc. can ship their own small snippet (as a non-conffile) for their tweaks | 05:20 |
| robert_ancell | what is a snippet? | 05:21 |
| pitti | robert_ancell: I mean a config file with the settings they want to change | 05:22 |
| robert_ancell | right | 05:22 |
| pitti | if you want to keep the conffile, https://launchpadlibrarian.net/74364166/support-vendor-config.patch also looks appropriate to me | 05:22 |
| pitti | (that's similar to what I meant by "local.conf" above) | 05:22 |
| pitti | but it might be more elegant to not have one at all | 05:22 |
| pitti | realistically we aren't going to be able to change it in the future, so internal defaults are better IMHO | 05:23 |
| robert_ancell | ok, I think I'm agreed. I'll make the defaults "do the right thing". We'll provide a symlink to unity in xsessions, and then the lightdm package won't provide any config by default (I'll move the config file to /usr/share/doc) | 05:25 |
| pitti | robert_ancell: sounds good | 05:25 |
| robert_ancell | I have 7 items, then I can release 0.9.0 | 05:26 |
| robert_ancell | It was 5 :P | 05:26 |
| jbicha | could lightdm use the Debian alternatives system? | 05:26 |
| pitti | robert_ancell: actually, with a non-conffile lightdm.conf, we wouldn't strictly need the default.desktop symlink part | 05:27 |
| pitti | robert_ancell: as xubuntu-default-settings or the cd-localization ones could then just ship a lightdm.conf with the session | 05:27 |
| robert_ancell | pitti, oh, about the incompatibility breaks, previously I made no API/ABI guarantees so it wasn't suprising half upgrades broke things. From 0.9.0 there will be those guarantees | 05:27 |
| jbicha | I imagine some people will come up with lightdm themes they want to distribute and requiring the user to hand-edit lightdm.conf isn't user-friendly | 05:28 |
| pitti | jbicha: please not -- alternatives for conffiles are even easier to break than either of those by themselves | 05:28 |
| pitti | robert_ancell: which breaks? | 05:28 |
| robert_ancell | jbicha, themes wont need to edit lightdm.conf | 05:28 |
| robert_ancell | pitti, mismatches between liblightdm and lightdm | 05:28 |
| RAOF | But actually *using* the theme will require editing lightdm.conf, right? Otherwise how does LightDM choose which theme to load? | 05:29 |
| jbicha | robert_ancell: I had to edit lightdm.conf to change to the unity switcher | 05:29 |
| pitti | RAOF: if there's just one, it picks that? | 05:29 |
| robert_ancell | jbicha, right. Greeters are now stored in /usr/share/xgreeters, and we can set it to "default" the same way as xsessions | 05:29 |
| pitti | if there's more, and no config, it should pick a predictable one, like alphabetically first one or so | 05:30 |
| RAOF | pitti: Won't ubuntu-desktop depend on the unity greeter? | 05:30 |
| robert_ancell | so a symlink could switch the default greeter | 05:30 |
| pitti | RAOF: sure, but ceratinly not on the example gtk one | 05:30 |
| pitti | symlink is nice, too | 05:30 |
| robert_ancell | pitti, are you dissing my example greeter? :) | 05:30 |
| jbicha | but if users want to use some random ppa theme, they'd have to remove ubuntu-desktop or edit the .conf, right? | 05:30 |
| pitti | robert_ancell: I know it's a marvellous piece of artwork, but it crashes all the time! | 05:31 |
| robert_ancell | they could ln -s mytheme.desktop /usr/share/xgreeters/default.desktop | 05:31 |
| robert_ancell | pitti, I haven't been reproducing these crashes | 05:31 |
| RAOF | robert_ancell: Your example greeter is certainly worth dissing compared to the unity greeter. That's funky! | 05:31 |
| jbicha | until the unity greeter gets an update and overrides the link, right? | 05:31 |
| robert_ancell | jbicha, no, the unity greeter package won't own the link | 05:32 |
| robert_ancell | I've been a little hesitant to do pitti's idea of just picking another greeter if the requested one doesn't exist. I don't think you should randomly run a greeter, it might be innappropriate | 05:33 |
| jbicha | ah, the good old days when there was a GUI to change the login theme! ;-) | 05:33 |
| robert_ancell | jbicha, we need to update GNOME tweak tool to configure this :) | 05:33 |
| pitti | robert_ancell: I meant that under the assumption of not having a lightdm.conf | 05:33 |
| robert_ancell | pitti, even still | 05:33 |
| pitti | without a conffile, picking the only one available seems appropriate to me | 05:33 |
| robert_ancell | pitti, would you log the user into another session if their requested one didn't exist? | 05:34 |
| pitti | robert_ancell: well, requested == .dmrc, which is a config file | 05:35 |
| robert_ancell | I think I'm going to put the responsibility on the sysadmin, or in this case perhaps a package hook to set it right | 05:35 |
| pitti | robert_ancell: for first login, yes | 05:35 |
| pitti | if a user never logged in and never chose a session, I'd log him into an available session instead of failing completely | 05:35 |
| robert_ancell | what will happen if the greeter does not exist is failsafe X should start and say "FIX YOUR SYSTEM!" | 05:35 |
| pitti | corresponding to "if there is no lightdm.conf, and only one greeter, use that" | 05:35 |
| pitti | since that's still friendlier than not allowing anyone to log in at all? | 05:36 |
| robert_ancell | pitti, that one greeter might log you into a remote system | 05:36 |
| pitti | robert_ancell: well, I have no strong opinion about it. a default symlink for the greeter sounds fine to me | 05:36 |
| cyphermox | time to go to bed, ttyl all! | 05:39 |
| micahg | robert_ancell: the fix is for lightdm to depend on one greeter with an alternative to a virtual greeter package to make sure one is always installed | 05:39 |
| robert_ancell | cyphermox, later | 05:40 |
| robert_ancell | micahg, yes, but should lightdm scan the list of available greeters and just pick one if you didn't configure one? | 05:40 |
| micahg | robert_ancell: I'm with pitti, go for a sane default that can be overridden | 05:42 |
| robert_ancell | but what is the sane default | 05:42 |
| micahg | robert_ancell: maybe lightdm can provide a failsafe greeter | 05:42 |
| pitti | I think "pick the first one available alphabetically" is predictable and robust; but only if it's not set in lightdm.conf | 05:43 |
| robert_ancell | micahg, that's failsafe X | 05:43 |
| pitti | then you can just do a 00default symlink etc. | 05:43 |
| micahg | robert_ancell: no, I'm referring to a greeter that works, but isn't customized, not a recovery console of sorts | 05:43 |
| micahg | failsafe should've been in quotes :) | 05:43 |
| micahg | I think the example greeter could qualify | 05:44 |
| robert_ancell | hmm | 05:44 |
| robert_ancell | will think about it | 05:44 |
| didrocks | yeah, what was puzzling I guess (similar for me to the session choice), was the harcoded "gnome.desktop" in gdm, alphabetical if no default provided seems more something sysadm can understand | 05:47 |
| didrocks | robert_ancell: btw, do you think we will have some time this week to talk about the session detection/fallback? | 05:48 |
| didrocks | robert_ancell: that can either be now or wait for the BOF days of desktop summit | 05:48 |
| robert_ancell | didrocks, is all we need to talk about "when are you going to implement it Robert?" | 05:48 |
| didrocks | robert_ancell: I remember I had some actions as well that you needed to turn into bug reports for me :-) | 05:49 |
| didrocks | robert_ancell: the ones you wrote on your notebook :) | 05:49 |
| robert_ancell | it's point #4 on my 0.9.1 features list | 05:49 |
| didrocks | and as I told you, if I can help you anywhere! | 05:49 |
| AfC | My laptop just made 3-4 alert/notification sounds at me. But I have no blinkies in Empathy, or Pidgin, or anything notifications. In fact, there's no longer a notification area. #epicfail | 05:50 |
| robert_ancell | didrocks, ok, please make a merge request :) | 05:51 |
| didrocks | robert_ancell: so, where should it be implemented in? I remember you didn't decide exactly where it should be. As long as I can do something you are pleased with… :-) | 05:52 |
| robert_ancell | didrocks, I'm always pleased with patches. I'll have a look and email you tomorrow some pointers | 05:52 |
| didrocks | robert_ancell: excellent!, waiting for that then :) | 05:52 |
| jbicha | transmission mirs filed http://pad.lv/813308 http://pad.lv/813313 http://pad.lv/813318 | 06:02 |
| pitti | jbicha: thanks | 06:08 |
| pitti | jbicha: which of these were previously bundled in the source? it might make sense to point this out in the MIR, as we factually already had that code in main before | 06:09 |
| jbicha | I think they still are in the source but let me check | 06:09 |
| jbicha | it doesn't look like minissdpd was bundled with Transmission but the other two definitely are, noted on bugs | 06:18 |
| micahg | jbicha: were they bundled with 2.13 (that was what shipped in natty)? | 06:18 |
| jbicha | micahg: they've been bundled since transmission 1.0 :-) | 06:20 |
| micahg | jbicha: great | 06:20 |
| didrocks | mvo: hey, FYI, ppa:ubuntu-desktop/ppa will get in a few, the awesome bschaefer's work for xapian and cjk. I just tested there is no regression and it seems my usc and unity-places are happy :) | 06:54 |
| mvo | didrocks: awsome! | 06:57 |
| mvo | didrocks: I saw his work, but did not had a chance yet to test it | 06:57 |
| Sweetshark | Morning all! | 07:15 |
| Sweetshark | https://launchpad.net/~bjoern-michaelsen/+archive/libreoffice-oneirictest-20110718/+build/2634822 => mood: excellent | 07:16 |
| Sweetshark | seeing the i386 build => mood: moderate | 07:16 |
| chrisccoulson | good morning everyone | 07:18 |
| desrt | Sweetshark: lulz :) | 07:31 |
| pitti | hey Sweetshark | 07:35 |
| didrocks | hey Sweetshark! | 07:35 |
| didrocks | good morning chrisccoulson | 07:35 |
| pitti | hey chrisccoulson | 07:36 |
| pitti | hey desrt, good morning | 07:36 |
| chrisccoulson | hi pitti and didrocks, how are you? | 07:36 |
| seb128 | hey | 07:37 |
| didrocks | chrisccoulson: I'm fine, thanks, and you? | 07:37 |
| didrocks | salut seb128 | 07:37 |
| seb128 | hey pitti desrt didrocks chrisccoulson | 07:37 |
| seb128 | how are you? | 07:37 |
| pitti | bonjour seb128 | 07:37 |
| chrisccoulson | didrocks, yeah, good thanks | 07:37 |
| chrisccoulson | hey seb128 | 07:37 |
| seb128 | didrocks, just saw your unity-application-place schemas update, do you know if anyone is working on moving unity keys there as well? | 07:39 |
| seb128 | didrocks, it's using desktop.unity for some stuff it seems | 07:39 |
| seb128 | dconf-editor shows that key namespaces are a bit disorganized | 07:39 |
| didrocks | seb128: yeah, all is still on /desktop/unity unfortunately. I was hoping after next release (so, like on Friday) to try to move everything back in com.canonical… | 07:40 |
| didrocks | seb128: agreed, sanity is needed there! | 07:40 |
| seb128 | we perhaps need schemas guideline, mvo moved update-notifier to com.ubuntu when he commited as well it seems | 07:41 |
| didrocks | indeed | 07:44 |
| mvo | seb128: much agreed, I think its not ideal to have the current mix | 07:46 |
| mvo | seb128: I'm happy to move that to com.canonical if that is consistent with everything else, but really Iike com.ubuntu better as it shows that its really a community project | 07:52 |
| mvo | (but not better enough to be inconsistent with the rest :) | 07:52 |
| seb128 | mvo, well it landed like that, I don't think it's worth the renaming, at least it's com.ubuntu not desktop.something as others did ;-) | 07:53 |
| mvo | well, I'm fine renaming it if its the only com.ubuntu project in dconf, that is really a bit silly | 07:55 |
| seb128 | mvo, let's see, I guess things like ubuntu-tweak should land in that namespace as well | 07:56 |
| * mvo nods | 07:56 | |
| mvo | the longer term plan (I know I keep talking about this since 4.10) is to get rid of u-n for ubuntu anyway and use something like upstart session jobs | 07:57 |
| mvo | then u-n will move to universe for other derrivatives who like the notification area | 07:57 |
| mvo | and the namespace will fit again :) | 07:57 |
| mvo | so its all part of long term planning! | 07:57 |
| === pitti_ is now known as pitti | ||
| bschaefer | didrocks: Here is a screen shot of the software center, it looks to be matching positively for the word 'text' http://i.imgur.com/YhbC6.png | 08:07 |
| didrocks | bschaefer: awesome! | 08:09 |
| didrocks | mvo: ^ | 08:09 |
| mvo | bschaefer: \o/ | 08:11 |
| mvo | bschaefer: out of curiosity, is this going (or already went) to upstream xapian? | 08:11 |
| bschaefer | mvo: Do you mean is it getting patched in Xapian? | 08:12 |
| mvo | bschaefer: I assume its a patch against libxapian, right? | 08:12 |
| mvo | bschaefer: so I was wondering if that patch is send to the xapian project already | 08:13 |
| bschaefer | mvo: There are a few things he wants me change but I updated the ticket. http://trac.xapian.org/ticket/180 | 08:13 |
| mvo | bschaefer: aha, great! that was my question :) | 08:13 |
| bschaefer | mvo: Yeah, he seems to like it but I need to change few things to make him happy haha | 08:13 |
| seb128 | rodrigo_, hey | 08:15 |
| mvo | bschaefer: I just looked over his comment and the code and it seems like its not that much work to make him happy, really excellent news! | 08:16 |
| rodrigo_ | morning | 08:16 |
| seb128 | rodrigo_, how are you? | 08:17 |
| bschaefer | mvo: Yeah the only part that I think will take a little time is his last one, about merging in a new CJKTERM for the Parse. Everything else is done already | 08:17 |
| rodrigo_ | seb128, I'm fine, and you? | 08:17 |
| seb128 | rodrigo_, I'm great thanks | 08:17 |
| seb128 | rodrigo_, so question for you, when do we get a new g-s-d version? ;-) | 08:17 |
| rodrigo_ | seb128, next week | 08:17 |
| rodrigo_ | 3.1.4 is next week | 08:18 |
| seb128 | rodrigo_, hum ok | 08:18 |
| seb128 | guess we can wait | 08:18 |
| rodrigo_ | seb128, you wanted earlier? | 08:18 |
| seb128 | rodrigo_, the dx guys made indicator-power depends on the gpm to gsd move | 08:18 |
| rodrigo_ | ah | 08:18 |
| seb128 | so updates for it are blocked on a new gsd to land | 08:18 |
| rodrigo_ | hmm | 08:19 |
| seb128 | would be benefit from git snapshoting this week to give it a bit of testing before next week tarball? | 08:19 |
| seb128 | be->we | 08:19 |
| rodrigo_ | yes, I guess we can do a snapshot | 08:19 |
| rodrigo_ | btw, what other apps use g-p-m? | 08:20 |
| rodrigo_ | I am planning to update them all toether with g-s-d | 08:20 |
| rodrigo_ | together | 08:20 |
| seb128 | rodrigo_, well maybe push in the ubuntu-desktop ppa the snapshot this week and next week tarballs to oneiric? | 08:26 |
| rodrigo_ | yes | 08:26 |
| seb128 | rodrigo_, good question, if you give me a dbus namespace to grep on or something I can do it from the dc side and grep through main for it | 08:26 |
| rodrigo_ | seb128, org.gnome.PowerManager | 08:27 |
| seb128 | rodrigo_, ok, will have a grep on it | 08:28 |
| jibel | didrocks, I filed bug 813365 about compiz becoming slower over time. | 08:31 |
| ubot2 | Launchpad bug 813365 in compiz "compiz leaks memory, becomes sluggish and unusable after using it for few hours" [Undecided,New] https://launchpad.net/bugs/813365 | 08:31 |
| didrocks | jibel: and you added the lovely 0.9.5.0 tag, thanks! | 08:32 |
| jibel | didrocks, yw | 08:32 |
| seb128 | jibel: hey | 08:35 |
| jibel | salut seb128 | 08:35 |
| seb128 | jibel: is your chart fta's one or one you did? | 08:36 |
| jibel | seb128, one I did, that is a pmap of compiz process every minute | 08:37 |
| jibel | I didn't knew fta did one already | 08:37 |
| seb128 | jibel: is there a tool to build those or you just scripted it locally in some way? | 08:37 |
| seb128 | jibel: http://ftagada.wordpress.com/2011/07/18/leaky-unity-in-oneiric/ | 08:37 |
| jibel | seb128, it is a very simple script http://paste.ubuntu.com/648072/ then I draw the chart with libreoffice calc. | 08:39 |
| seb128 | jibel: ok, thanks | 08:39 |
| seb128 | jibel: btw could you try if you get the issue with "Dialog Handler" activated in ccsm? | 08:39 |
| jibel | if you're interested I can do something with gnuplot instead to remove the manual step | 08:39 |
| seb128 | jibel: it's the unity dialog thing | 08:39 |
| seb128 | jibel: no, that's ok thanks, I was just wondering if there was some nice leak tracking tool I didn't know about | 08:40 |
| seb128 | would be handy ;-) | 08:40 |
| didrocks | jibel: also, try with the dialog handler removed, and gtk-w-d instead of u-w-d | 08:40 |
| seb128 | didrocks, heh, I just said that :p | 08:40 |
| seb128 | with->without in what I said though | 08:41 |
| didrocks | seb128: hence the precision of deactivated :) | 08:41 |
| didrocks | and gtk-w-d can help as well! as a second step | 08:41 |
| seb128 | didrocks, seems like there is also a bug somewhere, some users get 100% cpu use and nothing displayed with the update | 08:41 |
| didrocks | seb128: jono reported that, did you see other people (100% CPU for omer as well, but it's displaying) | 08:42 |
| seb128 | happens only when using the unity decorator | 08:42 |
| seb128 | didrocks, Cimi was having it | 08:42 |
| didrocks | oh, you get them testing that? | 08:42 |
| seb128 | until he moved libdecor.so away | 08:42 |
| seb128 | then he tried with gtk decorator which works | 08:42 |
| didrocks | hum, libdecor is different from gtk-d-w | 08:43 |
| didrocks | ok :) | 08:43 |
| didrocks | so we have a workaround for now | 08:43 |
| didrocks | sam isn't there this week | 08:43 |
| seb128 | but his compiz doesn't start with the unity decorator | 08:43 |
| seb128 | he was going to figure what commit broke it when I went to bed so maybe check with him if he figured it out | 08:43 |
| didrocks | hum, I'm pretty sure all this unity dialogs change for u-w-d is guilty | 08:43 |
| seb128 | well "to figure", "to try to figure" | 08:43 |
| seb128 | i.e he started on bzr builds to find the revision which broke it | 08:44 |
| seb128 | not sure how much he got done | 08:44 |
| seb128 | DBO was somewhat trying to look at finding the revision with him when I left | 08:44 |
| didrocks | seb128: he told me he didn't have the time this morning to look at the perf regression and he's at a conference now | 08:46 |
| seb128 | didrocks, ok | 08:46 |
| didrocks | or do you mean Cimi? | 08:46 |
| seb128 | either of them | 08:46 |
| seb128 | well Cimi said he would find the commit | 08:46 |
| didrocks | ok, let's try to see with them :) | 08:46 |
| seb128 | not sure if he did or not ;-) | 08:46 |
| didrocks | thanks for the investigation | 08:46 |
| seb128 | didrocks, I'm just giving the infos I have from yesterday evening in case if they are useful | 08:46 |
| seb128 | well at least the gtk decorator trick should work for jono | 08:47 |
| didrocks | it is, the gtk-w-d is something that stroke me yesterday evening and thought "oh, I should have asked jono for it" | 08:47 |
| Amoz | hey, I'm tyring to package a simple gnome-shell extension, but I'm a little lost | 08:58 |
| jibel | didrocks, also bug 813359 is new | 08:59 |
| ubot2 | Launchpad bug 813359 in unity "'Super' shortcuts for the launcher doesn't work anymore" [Undecided,New] https://launchpad.net/bugs/813359 | 08:59 |
| Amoz | do I need a .install file to tell dh where the extension files should go? | 08:59 |
| didrocks | jibel: you're right, I guess it's because of the new compiz, unity didn't change apart from a rebuild | 09:00 |
| didrocks | jibel: confirming | 09:00 |
| === zyga-afk is now known as zyga | ||
| rodrigo_ | is anyone able to boot into a 3.0 kernel? it panics for me all the time | 09:33 |
| seb128 | I didn't try, I tend to be conservative with non desktop upgrades :p | 09:33 |
| didrocks | rodrigo_: it works here, just I can't use the latest kernel with nvidia | 09:34 |
| rodrigo_ | didrocks, ah, a good reason to not use it on my desktop machine, as it needs nvidia | 09:35 |
| didrocks | rodrigo_: you can with the -3 kernel though | 09:35 |
| rodrigo_ | anyway, a good thing we keep the old kernels installed :) | 09:35 |
| rodrigo_ | although too many, need to do some cleanup | 09:36 |
| jbicha | rodrigo_: I get a kernel panic when trying to boot Rawhide with 3.0 but Ubuntu works fine for me | 09:58 |
| didrocks | seb128: oh, you answered to bug #723861? I didn't challenge that :) | 09:58 |
| ubot2 | Launchpad bug 723861 in nautilus "Right clicking on the desktop still displays the “Create Launcher...” option." [Wishlist,Fix committed] https://launchpad.net/bugs/723861 | 09:58 |
| rodrigo_ | jbicha, ok, bad luck for me, I guess :) | 09:58 |
| seb128 | didrocks, yeah, I got annoyed by the "I know better" comment which was wrong | 09:58 |
| didrocks | indeed, especially that my comment on the binary clearly showed that it's a separate binary :) | 09:59 |
| jbicha | I've got a branch for gnome-games & aisleriot, not sure if I did everything right | 10:05 |
| seb128 | jbicha, send it for review ;-) | 10:07 |
| jbicha | https://code.launchpad.net/~jbicha/+junk/aisleriot-3.1.0 | 10:07 |
| jbicha | https://code.launchpad.net/~jbicha/ubuntu/oneiric/gnome-games/gnome-games-3.1.3 | 10:07 |
| seb128 | jbicha, could you do a merge request for the second one, would make the diff easier to review ;-) | 10:08 |
| seb128 | jbicha, would be nice to open a bug for the first one as a sponsoring request, review board | 10:08 |
| seb128 | jbicha, I will try to have a look after lunch | 10:08 |
| jbicha | seb128: it won't let me because it looks like there are no gnome-games Ubuntu branches which is weird | 10:09 |
| seb128 | jbicha, lp:~ubuntu-desktop/gnome-games/ubuntu ? | 10:09 |
| seb128 | https://code.launchpad.net/~ubuntu-desktop/gnome-games/ubuntu | 10:09 |
| jbicha | yes, but there isn't a ubuntu:gnome-games so launchpad is not giving me a merge button | 10:10 |
| jbicha | maybe if I push it to a +junk branch... | 10:10 |
| seb128 | hum | 10:11 |
| seb128 | jbicha, would be a good question for #launchpad I guess | 10:14 |
| jbicha | I figured it out, it was because of the ubuntu/oneiric in my branch name | 10:14 |
| seb128 | jbicha, i.e why can't you merge propose a diff against lp:~ubuntu-desktop/gnome-games/ubuntu when your vcs is derivated from it | 10:14 |
| jbicha | https://code.launchpad.net/~jbicha/gnome-games/gnome-games-3.1.3/+merge/68516 | 10:28 |
| jbicha | bug 813428 | 10:29 |
| ubot2 | Launchpad bug 813428 in ubuntu "[needs-packaging] aisleriot" [Undecided,New] https://launchpad.net/bugs/813428 | 10:29 |
| seb128 | jibel: thanks, how did you get around it to file the merge request? | 10:36 |
| jbicha | left out the ubuntu/oneiric part of the URL | 10:37 |
| seb128 | ok | 10:38 |
| jbicha | I'm going to be out for several hours | 10:40 |
| === _LibertyZero is now known as LibertyZero | ||
| === tkamppeter_ is now known as tkamppeter | ||
| mdeslaur | pitti: no progress from upstream policykit on our two open bugs? | 11:50 |
| * rodrigo_ lunch | 12:14 | |
| === MacSlow is now known as MacSlow|lunch | ||
| pitti | mdeslaur: I'll ping David again, he promised to do the new release last week | 12:29 |
| mdeslaur | pitti: cool, thanks! | 12:29 |
| jibel | seb128, didrocks, situation is worse with 'dialog handler' enabled +58.12 MB/hour over 3:45 hours | 12:39 |
| seb128 | jibel: well, it's enabled by default, did you try without it? | 12:39 |
| jibel | seb128, it was not enabled by default | 12:39 |
| pitti | oh dear, no wonder my machine is slow as hell -- it's swapping out like mad, apparently compiz has used almost a GB of RAM :( | 12:40 |
| seb128 | ok, I guess you got hit by the "upgrades don't change the configs in compiz" | 12:40 |
| seb128 | jibel: can you try without it and with the gtk decorator? | 12:40 |
| seb128 | pitti, cf jasoncwarner_'s email | 12:40 |
| pitti | yeah | 12:40 |
| jibel | seb128, how do I change the decorator ? | 12:41 |
| seb128 | jibel: gtk-window-decorator --replace | 12:41 |
| seb128 | (should work) | 12:41 |
| jibel | bbl | 12:44 |
| cyphermox | good morning! | 13:00 |
| pedro_ | morning cyphermox | 13:01 |
| cyphermox | hey pedro_ | 13:03 |
| === m_conley_away is now known as m_conley | ||
| seb128 | hey cyphermox | 13:06 |
| seb128 | ola pedro_ | 13:07 |
| seb128 | how are you guys, | 13:07 |
| seb128 | ? | 13:07 |
| pitti | hey cyphermox, hi pedro | 13:07 |
| didrocks | pitti: you can try to gtk-w-d as well btw | 13:08 |
| pedro_ | hola seb128 pitti | 13:11 |
| pedro_ | seb128, doing good and you? | 13:11 |
| pedro_ | salut didrocks | 13:11 |
| didrocks | hey pedro_! | 13:12 |
| didrocks | salut cyphermox | 13:13 |
| cyphermox | salut didrocks, ca va? | 13:14 |
| didrocks | cyphermox: ça va, et toi? | 13:15 |
| cyphermox | ouais | 13:15 |
| * pitti wonders if that is really a word -- 80% vowels? | 13:16 | |
| didrocks | pitti: coloquial way to tell "oui", which is 100% vowels :-) | 13:17 |
| pitti | four vowels in a row, *shaking head* | 13:18 |
| seb128 | pitti, don't worry we don't spell the s only the vowels ;-) | 13:20 |
| pitti | seb128: you mean "pronounce"? | 13:20 |
| seb128 | ups, yes, sorry | 13:20 |
| seb128 | it's pronounced "ouai" :p | 13:20 |
| pitti | weiaourd | 13:20 |
| seb128 | ;-) | 13:20 |
| seb128 | french is fun! | 13:20 |
| pitti | Pas de doute! | 13:22 |
| seb128 | I don't know what you guys are doing with your compiz | 13:22 |
| seb128 | mine seems stable in memory use with unity dialogs and unity decorator | 13:22 |
| pitti | I'd describe it as "bearing it" | 13:22 |
| seb128 | it's running since 9:35 and it's using 199mb of vsz and 95mv or rss | 13:23 |
| seb128 | jibel, pitti: do you have non standard indicators like the multiload one? | 13:23 |
| didrocks | I'm not the only one to get compiz running with reasonable size. Maybe it's because I don't open/close too many applicatino and I tend to reuse my windows | 13:23 |
| === MacSlow|lunch is now known as MacSlow | ||
| pitti | seb128: no, I removed multiload again | 13:24 |
| seb128 | didrocks, no, I keep closing and reopening evolution, firefox, gedit and some others | 13:24 |
| pitti | gtimelog, keyboard, nm-applet, message, battery, clock, broken thingy displaying my name, session | 13:24 |
| seb128 | I've been switching workspaces a lot using the keyboard and using the application place and alt-f2 a bunch of times | 13:24 |
| pitti | I don't actually open/close apps very much | 13:24 |
| pitti | I spend pretty much all my time in terminal windows, building one live CD after another | 13:25 |
| seb128 | hum weird | 13:25 |
| seb128 | you are on intel as well right? | 13:25 |
| jibel | seb128, yes multiload, it helps me to know when I will run out of memory ;-) | 13:25 |
| jibel | seb128, and weather of course | 13:25 |
| seb128 | jibel: once you are done testing with the gtk decorator could you try without multiload | 13:25 |
| jibel | I can remove all of them | 13:25 |
| didrocks | jibel: you should link "unity" command your multiload indicator :-) | 13:26 |
| seb128 | it's funny how the thing that you use to see when you run out of memory tends to be the one putting you there :p | 13:26 |
| pitti | seb128: if that was for me, yes I'm on intel | 13:26 |
| seb128 | pitti, yes it was, thanks | 13:26 |
| jibel | I'm on nvidia with proprietary driver | 13:26 |
| seb128 | jibel: indicator-multiload tends to exercise the indicator loader leaks | 13:26 |
| seb128 | if you get less leak without it that would mean part of the issue is in the indicator loader again | 13:27 |
| seb128 | though that didn't change with the recent uploads so it shouldn't be it | 13:27 |
| pitti | seb128: do we have that library again which allocates an extra 5 MB of GL context on nvidia? | 13:27 |
| pitti | ah, cairo | 13:27 |
| Sweetshark | french is nothing: how do you pronounce SwSpzFrmFmts? http://opengrok.libreoffice.org/xref/writer/sw/inc/docary.hxx#63http://opengrok.libreoffice.org/xref/writer/sw/inc/docary.hxx#63 | 13:28 |
| pitti | Sweetshark: that sounds line Win32 API, not a human language | 13:28 |
| jibel | it's 15:30, lost 30MB in 45min, and removed the multiload indicator. Let see if that makes a difference in the next 45 min | 13:28 |
| seb128 | pitti, yes | 13:29 |
| seb128 | jibel: you will get a 5mb extra use by gtk process running | 13:29 |
| seb128 | not sure if those count in compiz or xorg though | 13:29 |
| pitti | seb128: client-side | 13:29 |
| seb128 | 725434 | 13:30 |
| seb128 | is the bug | 13:30 |
| pitti | i. e. compiz, gedit, or whatever your program is | 13:30 |
| Sweetshark | pitti: thats czech notation: remove all vowels, they hurt runtime! SwSpzFrmFmts means "Star Writer Spezialle Frame Formate" btw -- a wonderful mix of german and english and vowel removal ... | 13:30 |
| pitti | *tsk* | 13:31 |
| chrisccoulson | rodrigo_, can we split evolution-couchdb to separate the eds backend and the evolution plugin? | 13:33 |
| seb128 | jibel: ok, so pitti has a good point, we turned back gl support in cairo which due to an nvidia binary driver issue lead to a few mbs "leak" by running gl process | 13:33 |
| rodrigo_ | chrisccoulson, yes | 13:33 |
| chrisccoulson | rodrigo_, thanks | 13:34 |
| chrisccoulson | rodrigo_, want me to do that? | 13:34 |
| rodrigo_ | chrisccoulson, I can do it, but feel free to if you want | 13:34 |
| chrisccoulson | rodrigo_, thanks. i'll maybe look at it later if i get a chance | 13:34 |
| rodrigo_ | chrisccoulson, ok, if not, I'll do it tomorrow | 13:34 |
| jibel | seb128, I can try with an intel card on the same system if that helps. | 13:35 |
| seb128 | jibel: well I'm just pointing out that it might be part of the issue | 13:36 |
| seb128 | jibel: but well as pitti pointed it, it's the client application that will see the "leak", so it's each process that will have an extra use, they should not reflect on the wm | 13:36 |
| didrocks | jibel: did you notice so many issue/increase with gtk-w-d? | 13:37 |
| pitti | but that nivida GL leak doesn't grow over time | 13:37 |
| seb128 | jibel: it would be useful if you could spot what action trigger leaks | 13:37 |
| seb128 | like if opening and closing the dash increase it | 13:37 |
| seb128 | pitti, right, it's just that's it's proportional to the number of things you have open | 13:38 |
| seb128 | pitti, well anyway there is a leak somewhere for sure, it's not only nvidia | 13:39 |
| jibel | seb128, switching desktop is one of them. I don't use the dash. | 13:42 |
| seb128 | hum | 13:42 |
| seb128 | I switch desktop like 25 times an hour | 13:43 |
| pitti | hm, I can't confirm this here | 13:43 |
| seb128 | so I switched desktop around 125 times today | 13:43 |
| seb128 | and I've no leak | 13:43 |
| pitti | mem usage keeps stable even when I wildly change desktops | 13:43 |
| seb128 | jibel: can you unity --reset to ensure you have a stock config? | 13:44 |
| didrocks | same here, I switch desktop a lot with the keyboard and have no issue | 13:44 |
| jibel | seb128, I switched a lot during last minute and lost 4.8MB | 13:44 |
| seb128 | in vsz or rss? | 13:45 |
| jibel | private area of compiz | 13:45 |
| seb128 | ok | 13:45 |
| seb128 | jibel: can you unity --reset? you didn't have the unity dialog on so your clearly don't have a config matching the current default one | 13:46 |
| seb128 | not sure if that could be due to some other option you turned on | 13:46 |
| seb128 | would be good to see if that happens with a default profile | 13:46 |
| seb128 | you can change the number of workspaces and keybindings those should not make a difference | 13:46 |
| chrisccoulson | rodrigo_, does ensure_couchdb_contacts_source_group in evolution-couchdb create the couchdb addressbook? | 13:47 |
| chrisccoulson | i want to provide the equivalent of the evo-couchdb plugin for thunderbird ;) | 13:47 |
| didrocks | seb128: jibel: I confirm the memleak here | 13:48 |
| seb128 | didrocks, when doing what? | 13:48 |
| didrocks | changing ws | 13:48 |
| jibel | ahah, compiz segfaults after a --reset :-) | 13:49 |
| didrocks | like crazy :) | 13:49 |
| seb128 | didrocks, what did you change? | 13:49 |
| didrocks | jibel: hum? I tried it just after latest release | 13:49 |
| didrocks | seb128: I just change them like crazy (20 times in 10s) | 13:49 |
| seb128 | 199392 94568 | 13:50 |
| seb128 | 199412 94592 | 13:50 |
| seb128 | that's after 15 ws1-ws3-ws1-ws3-ws1 by keybinding | 13:50 |
| seb128 | vsz rss | 13:50 |
| didrocks | I'm only looking for rss | 13:51 |
| seb128 | that's my second column | 13:51 |
| seb128 | 94672 after another 15 switches | 13:51 |
| seb128 | so it leaks maybe 10k by switch here | 13:51 |
| didrocks | that's weird if I change slowly, I don't see any increase | 13:51 |
| pitti | so it seems my compiz is stable at 752888 vss | 13:51 |
| didrocks | if I really force changing ws a lot, I see that increase | 13:51 |
| pitti | so it's huuge, but stably so | 13:51 |
| jibel | See you in a minute, --reset killed my laptop, invisible windows and such. BBL | 13:52 |
| seb128 | stupid compiz | 13:52 |
| seb128 | doing ws1-ws2-ws3 key bindings in a crazy random way leads to dialog landing on other workspaces | 13:52 |
| seb128 | where it should just be switching between ws | 13:53 |
| seb128 | hum | 13:54 |
| seb128 | does right clicking on the titlebar works for others? | 13:54 |
| seb128 | it used to open a menu which had the "send to workspace <n>" entries | 13:54 |
| pitti | seb128: it does strange stuff | 13:54 |
| pitti | shuffling and unmaximizing my windows | 13:54 |
| pitti | oh, window title bar | 13:54 |
| pitti | yes, that's broken | 13:54 |
| pitti | I clicked on the panel | 13:55 |
| didrocks | seb128: IIRC, right clicking has been removed on purpose by sam, but better to check again with him | 13:58 |
| seb128 | ok | 13:58 |
| jibel | --reset fixed the problem in some way. I lost all the shortcuts and can't switch between desktops anymore :-) | 14:04 |
| seb128 | jibel: you need to reconfigure those ;-) | 14:04 |
| mterry | seb128, btw, you asked about datetime preferences yesterday -- there is no way right now. I'm trying to get to the point where I can flip the switch on parsing Unity/GNOME OnlyShowIn values (I think I will today) | 14:08 |
| didrocks | jibel: yeah, the shortcuts configuration is broken with GNOME3 (no more g-c-c integration) | 14:08 |
| seb128 | mterry, ok, that's what I though, thanks | 14:08 |
| seb128 | didrocks, jibel: you should be able to change those in ccsm | 14:09 |
| didrocks | yeah, and it works (that's what I've done yesterday when trying --reset before pushing the package) | 14:09 |
| jjardon | dobey: Hello, I'm crious about the Ubuntu One configuration panel. will It use the new online accounts panel? | 14:10 |
| dobey | jjardon: i don't think we have any plans to do so, no | 14:12 |
| jjardon | dobey: Could I ask why not? Any technical limitation? | 14:15 |
| jibel | I also lost the application and files lenses. How do I readd them? | 14:16 |
| dobey | jjardon: we haven't scheduled anyone to work on it, direction of the new thing in gnome was unclear at UDS, etc. i don't know about what technical issues there may be yet, as i haven't looked at how that dialog works at all, but i am guessing it doesn't quit fit well with services that only allow a single account to be logged in at a time | 14:17 |
| dobey | jjardon: and i don't think it actually solves any of the real problems with having multiple on-line services accounts | 14:18 |
| seb128 | jibel: it seems to happen sometimes on unity restart here but I never get it on a new session | 14:20 |
| jjardon | dobey: http://people.freedesktop.org/~david/gnome-online-accounts-3.1.0/index.html | 14:20 |
| jibel | I'll look at it later, it is not like is 10.04.3 was due tomorrow | 14:21 |
| jibel | is->if | 14:21 |
| seb128 | jibel: ok, yeah better to focus on priority, thanks for taking some time to help on figuring the oneiric leak issues | 14:22 |
| didrocks | jibel: seb128: really? I never had that one, the places reconnects correctlyl after manual unity restart | 14:22 |
| seb128 | didrocks, I had it a few times recently in oneiric | 14:23 |
| jjardon | dobey: and http://davidz25.blogspot.com/2011/04/gnome-online-accounts.html | 14:23 |
| dobey | jjardon: right, i don't think Ubuntu One really fits into that design | 14:23 |
| jibel | I'm monitoring compiz without dialog handler, after a reset and default indicator. I'll update the report with the results. | 14:23 |
| jibel | next thing will be to test with another card on the same system to eliminate the graphics driver. | 14:24 |
| didrocks | seb128: ok, I have sometimes, the place not answering on first run, but it's running generally | 14:24 |
| didrocks | jibel: maybe I miss it, but did you report with gtk-w-d? | 14:25 |
| jjardon | dobey: Did you send any feedback upstream? I guess they would be happy to adjust it if Its needed | 14:25 |
| dobey | jjardon: no, like i said; i haven't looked at it in depth really. but for ubuntu one, it would basically be throwing away a bunch of our existing code, rewriting it all to be a bit more complex, and ending up with probably a worse user experience, in the end | 14:26 |
| dobey | jjardon: the only thing it does that u1 could make use of, is storing the authentication token somewhere, and we already do that. | 14:27 |
| dobey | jjardon: so i'm not seeing any benefit to using it | 14:28 |
| jjardon | dobey: I more worried about use another panel to configure Ubuntu One, when It can be done through the standard online accounts one | 14:28 |
| jibel | didrocks, not yet. I have some iso tracker stuff to do before. | 14:29 |
| didrocks | jibel: sure, just think about that combination (it seems it will avoid the computation in the decorator which makes everything slow) | 14:30 |
| didrocks | pitti: maybe you are interested as well ^ | 14:30 |
| dobey | jjardon: i don't understand what you mean by that i guess. gnome-online-accounts doesn't really configure anything related to the actual use of the accounts, afaict | 14:31 |
| seb128 | pedro_, is there a way to mark but as "to forward upstream" in a way you notice? | 14:32 |
| seb128 | pedro_, like is doing "also affect component" something you would notice? | 14:33 |
| pedro_ | seb128, you can open an empty upstream task , i'm looking those to forward upstream | 14:34 |
| pedro_ | seb128, that way they'll also appear in the 'upstream report' | 14:34 |
| seb128 | pedro_, ok great, I'm thinking that bugs tagged oneiric with an upstream component non filled should be upstreamed | 14:35 |
| seb128 | pedro_, basically a bunch of nautilus and g-c-c issues that seem would be useful to send there | 14:35 |
| dobey | seb128: i didn't get around to bugging kees last night, btw. if you can poke at my merge proposal for ubuntuone-control-panel, that would be great. converted it to dh_python2 for oneiric | 14:35 |
| seb128 | dobey, oh right, doing that now | 14:35 |
| dobey | seb128: thanks! | 14:36 |
| seb128 | dobey, yw | 14:36 |
| pedro_ | seb128, sure, i'll send those and if you notice any other just leave an upstream task open and i'll take care of those as well | 14:36 |
| seb128 | ok | 14:37 |
| seif | seb128, didrocks i am on the zeitgeist stuck now | 14:39 |
| seb128 | stuck? | 14:40 |
| didrocks | seif: integration in g-c-c for activity log manager? | 14:40 |
| mterry | TheMuso, your gnome-orca upload contains a debian-changes-3.1.3-0ubuntu2 patch that seems to undo some of the other patches? | 14:41 |
| dobey | mterry: hooray for format 3.0 (quilt) being dumb. | 14:43 |
| mterry | dobey, yeah, those debian-changes patches are a pain | 14:43 |
| Laney | you do get a lintian warning about them | 14:46 |
| seif | didrocks, yes sir | 14:46 |
| seif | setting up the resources here | 14:46 |
| didrocks | seif: nice, with your script included there and John providing you design advice? | 14:47 |
| seif | yep | 14:55 |
| didrocks | seif: just ping me if you need support/packaging then :) | 14:57 |
| didrocks | seif: how big this is, btw? | 14:57 |
| seif | well the hardest part is integration with gnome control center | 14:57 |
| seif | its something new for us | 14:58 |
| seif | i need to figure out where to start | 14:58 |
| seif | and if they support python stuff | 14:58 |
| didrocks | seif: maybe you should check with mterry and rodrigo? I'm not sure it's supported | 14:58 |
| mterry | seif, no python yet | 14:59 |
| seif | what do i need to use then | 14:59 |
| mterry | seif, C (or something like Vala that compiles to C) | 14:59 |
| seif | oki doki | 14:59 |
| seif | this is gonna be difficult since my C is not that good | 15:00 |
| seif | but i am willing to try | 15:00 |
| seif | i might find some1 esle in the team better with ti | 15:00 |
| mterry | seif, if you want an example of a third party plugin, see my app deja-dup | 15:00 |
| seif | mterry, any documentation around | 15:00 |
| mterry | seif, I an walk you through it | 15:00 |
| seif | mterry, awesome can you send me the deja-dup stuff | 15:00 |
| seif | ? | 15:00 |
| mterry | seif, not so much documentation. Basically you just need a shim to stick a widget of some kind in there | 15:00 |
| SpamapS | Has anyone else complained about nvidia+twinview not working? | 15:00 |
| mterry | seif, yeah, hold on a sec | 15:00 |
| didrocks | SpamapS: it's not working with latest kernel for me, but otherwise, it's fine | 15:01 |
| SpamapS | I reported bug #813343 last night against unity.. but thinking about it I wonder if it is compiz fault | 15:01 |
| ubot2 | Launchpad bug 813343 in unity "nvidia drivers, second monitor covered by black" [High,Incomplete] https://launchpad.net/bugs/813343 | 15:01 |
| mterry | seif, here's the shim you need: http://bazaar.launchpad.net/~deja-dup-hackers/deja-dup/20/view/head:/preferences/PreferencesPanel.c | 15:01 |
| mterry | seif, in the deja_dup_preferences_panel_init() function, you'll see we just stick a widget (which can be anything) into the panel | 15:02 |
| mterry | seif, see the Makefile.am and the toplevel configure.ac for how to get the CFLAGS, LIBS, and install directory you need | 15:03 |
| mterry | (the preferences/Makefile.am that is) | 15:03 |
| seif | i only need gtk and zeitgeist | 15:03 |
| seif | :) | 15:03 |
| mterry | seif, yeah, but I meant the CFLAGS and such for libgnome-control-center-dev | 15:04 |
| seif | mterry, i got it bookmarked | 15:05 |
| seif | will start hacking tomorrow | 15:05 |
| seif | right now just trying to organize myself | 15:06 |
| mterry | seif, cool, I'll be around for questions too | 15:06 |
| seb128 | seif, mterry: speaking of third party g-c-c upstream dropped the public api in git (well at least the include and .pc) | 15:07 |
| seb128 | so those will only work on distro that revert that commit | 15:07 |
| seif | seb128, wtf? | 15:07 |
| mterry | seb128, yup :( deja-dup is already built to handle that (falls back to dialog) | 15:07 |
| mterry | seif, oh yeah, sorry, assumed people knew about that controversy :) | 15:08 |
| seb128 | seif, "design decision, nobody else than the gnome-control-center maintainers should be able to put anything in gnome-control-center" | 15:08 |
| seb128 | things that go there should be design reviewed, approved and shipped in the g-c-c source | 15:09 |
| mterry | seif, Ubuntu is committed to continuing to allow 3rd party panels (perhaps with some whitelist) | 15:09 |
| mterry | seif, if you're still interested in having a panel on the distros that support it, deja-dup has fallback code (including code to write install .desktop files, with appropriate OnlyShowIns to show the dialog some places, the panel others) | 15:10 |
| seif | mterry, yeah | 15:11 |
| seif | i am going to fix other issues first | 15:11 |
| mterry | :) | 15:11 |
| seif | then we will drop to the ui | 15:11 |
| chrisccoulson | these logos are awesome: http://mozilla.seanmartell.com/tb/tb-channels.png :) | 15:28 |
| climbe2 | Problem! If anyone is able to help.... running 10.04, recently upgraded to kernel 2.6.32-33, and can't boot up!! See http://ubuntuforums.org/showthread.php?t=1807978 for my ongoing thread. Any suggestions?!?! | 15:33 |
| seb128 | chrisccoulson, nice ones! | 15:35 |
| chrisccoulson | seb128, yeah, i think i prefer those to the firefox channel logos | 15:35 |
| seb128 | climbe2, hi, doesn't seem a desktop issue so that's not the best place to ask, maybe try #ubuntu-kernel or #ubuntu-devel rather | 15:36 |
| climbe2 | ok, thanks | 15:36 |
| pitti | good night everyone! | 15:39 |
| seb128 | 'night pitti | 15:40 |
| seb128 | vuntz, how is the wnck review going? there is an update for the bug waiting for you for over a week ;-) | 15:49 |
| chrisccoulson | wth, my monitors just changed mode for no apparent reason :/ | 15:52 |
| chrisccoulson | they suddenly became mirrored whilst i was working | 15:52 |
| seb128 | chrisccoulson, your system know better what you need ;-) | 15:53 |
| chrisccoulson | seb128, it's really annoying. it did it a couple of weeks ago too, and what really sucks is that compiz messes up when i try to restore my display configuration (in the same way as it does when i connect my monitor) | 15:54 |
| chrisccoulson | g'ah, now my session is totally screwed | 15:58 |
| mterry | Are problems logging into Unity 3D known? I just get a black screen with a cursor when I try | 16:47 |
| seb128 | mterry, sort of, try disabling or moving away libdecor.so or unity-window-decorator | 16:48 |
| seb128 | mterry, using gtk-window-decorator should get you out of it | 16:48 |
| seb128 | it happens to some people | 16:48 |
| seb128 | there it's just really slow, like it takes 15 seconds to log in instead of 3 seconds | 16:48 |
| mterry | maybe I didn't wait long enough... | 16:48 |
| bryceh | chrisccoulson, sounds like compiz or gnome-settings-daemon messed up and fell back to X default (which is mirrored). Check your logs for compiz or gnome errors relating to screens. | 16:49 |
| chrisccoulson | bryceh, thanks, will look in a bit | 16:49 |
| bryceh | chrisccoulson, also was it completely random or is it possible you had plugged something in? | 16:49 |
| chrisccoulson | seb128, oh, that's what it is? i thought my laptop had frozen, it seemed to hang for ages when i logged in | 16:50 |
| chrisccoulson | bryceh, it was completely random | 16:50 |
| seb128 | chrisccoulson, right, it's also really slow to open any dialog on session start | 16:50 |
| didrocks | chrisccoulson: indeed, sounds like the same symptom | 16:51 |
| seb128 | it gets better once they got open once | 16:51 |
| bryceh | chrisccoulson, ok good luck. If you do find an activity that triggers it, I'd be interested to know. | 16:52 |
| mterry | seb128, no good... moving it out of the way didn't help. manually running gtk-window-decorator didn't help | 16:52 |
| seb128 | mterry, moving libdecor.so didn't "fix" it? | 16:53 |
| seb128 | mterry, ok, so not a known issue | 16:54 |
| mterry | seb128, oh, I just moved the unity-window-decorator | 16:54 |
| mterry | I'll move libdecor.so too. Thought you said 'or' not 'and' | 16:54 |
| seb128 | mterry, well I though it was a or solution but check by moving libdecor.so to be sure | 16:55 |
| mterry | seb128, success! That worked, though obviously now I have no decorations | 16:56 |
| seb128 | mterry, try putting it back and running gtk-window-decorator by hand, maybe that works | 16:57 |
| mterry | seb128, I tried that with libdecor.so in place and it didn't work | 16:57 |
| seb128 | I didn't get the issue, just slowness, my debug hints are based on what I read around on IRC ;-) | 16:57 |
| mterry | but maybe without libdecor.so it would | 16:57 |
| seb128 | mterry, what happen with libdecor in place if you try to run the gtk decorator? | 16:57 |
| mterry | seb128, nothing. I still was stuck on a black screen | 16:58 |
| seb128 | well from your working compiz I mean | 16:58 |
| mterry | and without libdecor, nothing happens when I run it as well | 16:58 |
| didrocks | mterry: did you set gtk-w-d as the default one? | 16:58 |
| mterry | Right, nothing happens now | 16:58 |
| seb128 | does it hang compiz? | 16:58 |
| mterry | seb128, no | 16:58 |
| seb128 | didrocks, he moved u-w-d away | 16:58 |
| mterry | didrocks, no | 16:58 |
| mterry | didrocks, yeah, unity-w-d and libdecor.so are moved | 16:58 |
| seb128 | didrocks, shouldn't do that lead to the same? | 16:58 |
| didrocks | seb128: compiz-decorator doesn't fallback to gtk-w-d | 16:59 |
| seb128 | didrocks, what does it do if there is no decorator? | 16:59 |
| didrocks | seb128: it's just… do nothing :) | 17:00 |
| seb128 | didrocks, well mterry still has the hang with u-w-d moved away | 17:00 |
| seb128 | didrocks, he had to move libdecor.so away to workaround it | 17:00 |
| seb128 | mterry, well, I suggest you check with dx from there on | 17:00 |
| * mterry goes to #ayatana | 17:00 | |
| seb128 | mterry, Jason was looking to find what commit was an issue and Cimi has the bug you are running into | 17:01 |
| didrocks | mterry: can you just try last one thing? | 17:01 |
| seb128 | but he got the gtk decorator to work I though | 17:01 |
| mterry | didrocks, sure | 17:01 |
| didrocks | mterry: move libdecor.so back | 17:01 |
| didrocks | open ccsm | 17:01 |
| didrocks | in the decor plugin, replace /usr/bin/compiz-decorator with /usr/bin/gtk-window-decorator | 17:02 |
| mterry | didrocks, ok, brb | 17:02 |
| didrocks | seb128: normally, if no decorator is found, metacity is launched, but maybe compiz doesn't release the X window lock at this stage… | 17:02 |
| didrocks | mterry: thanks :) | 17:03 |
| seb128 | didrocks, yeah, I though that no decorator binary would workaround the decorator issue | 17:04 |
| seb128 | didrocks, but it means it's not an issue for the u-w-d code itself if that happens without the decorator installed | 17:04 |
| didrocks | seb128: it should anyway, metacity --replace should be launched and should be working there | 17:05 |
| seb128 | didrocks, no decorator means no compiz? | 17:05 |
| didrocks | but I bet compiz doesn't allow that now when initialazing | 17:05 |
| mterry | didrocks, worked | 17:05 |
| didrocks | seb128: indeed, that's what's in compiz-decorator | 17:05 |
| seb128 | ok I though compiz would just run :p | 17:05 |
| didrocks | mterry: ok, so confirmed and confimed again it's u-w-d then! | 17:06 |
| seb128 | mterry, ok, so you have the same issue than Cimi | 17:06 |
| didrocks | mterry: so yeah, better to debug with Jason on #ayatana! | 17:06 |
| seb128 | check with DBO | 17:06 |
| seb128 | or with Cimi | 17:06 |
| seb128 | not sure if they figured something | 17:06 |
| didrocks | seb128: so, no, it's either libdecor -> no decoration or no decorator -> no compiz :) | 17:06 |
| didrocks | no cheese and desert :) | 17:06 |
| seb128 | ;-) | 17:06 |
| didrocks | but oh, a desrt there! | 17:06 |
| didrocks | :-) | 17:06 |
| seb128 | didrocks, shush, let him work | 17:06 |
| didrocks | seb128: he's compiling, I don't call that work :p | 17:07 |
| seb128 | he's fixing all those things people complain about for years today | 17:07 |
| seb128 | he must by inspired by Lyon or something | 17:07 |
| didrocks | heh, you have no idea how many beers this costs me :-) | 17:07 |
| seb128 | ;-) | 17:07 |
| dobey | mterry: was it you who i saw yesterday? making a comment about gnome-keyring-daemon being broken? | 17:07 |
| seb128 | speaking of which it's beer time for you guys | 17:07 |
| seb128 | or it should be beer time | 17:07 |
| didrocks | yeah, we will fix that really soon I guess! | 17:08 |
| mterry | dobey, no, not yesterday | 17:08 |
| seb128 | brb | 17:08 |
| dobey | mterry: hrmm. | 17:08 |
| dobey | oh, was james_w | 17:08 |
| === m_conley is now known as m_conley_away | ||
| === m_conley_away is now known as m_conley | ||
| * mterry can finally cross off the work item "get our datetime preferences squared away" by patching the entire world around | 18:03 | |
| james_w | dobey, yeah, it was me | 18:12 |
| dobey | james_w: i don't think it's supposed to be setuid, as it's not in 11.04 either it seems. were you seeing an issue where it wasn't starting up with the session at log-in? | 18:13 |
| james_w | dobey, yeah | 18:13 |
| james_w | dobey, and trying to start manually complained about not being able to get capabilities | 18:14 |
| dobey | james_w: with --daemonize option also? | 18:15 |
| james_w | dobey, no, didn't try that | 18:15 |
| dobey | james_w: does it happen every log-in, or only some of the time? | 18:16 |
| dobey | james_w: also, is there a bug # for this? :) | 18:16 |
| james_w | dobey, seemed to be every login | 18:17 |
| james_w | I only tried a couple before making it setuid | 18:17 |
| james_w | I didn't file one yet | 18:17 |
| dobey | james_w: please file. have someone with the same issue wondering why they can't log into u1 :) | 18:18 |
| mterry | seb128, how does versions.html find new upstream versions? I didn't think it just looked at debian/watch... | 18:18 |
| mterry | pitti, ^ | 18:18 |
| desrt | seb128: units policy changed in glib now | 18:19 |
| desrt | seb128: so you can happily drop that patch completely :) | 18:19 |
| desrt | seb128: also... you'll probably find the eventfd patch no longer applies to the just-released GLib. i recommend attempting to drop it, since the recent changes may have fixed the issue. | 18:19 |
| desrt | seb128: let me know if not... | 18:19 |
| SpamapS | Hrm.. this unity bug w/ twinview is really killing my ability to use Oneiric | 18:25 |
| SpamapS | xfce, lxde, kde, all suck. I want my Unity back. :/ | 18:25 |
| dobey | SpamapS: you should try ratpoison :) | 18:31 |
| SpamapS | dobey: has it improved in the last 4 years? thats the last time I tried it. :) | 18:33 |
| dobey | SpamapS: how do you improve on perfection? ;) | 18:33 |
| SpamapS | mmmm... so true | 18:33 |
| seb128 | re | 18:33 |
| seb128 | desrt, you shouldn't be there, you should be out seeing Lyon and drinking a beer with didrocks ;-) | 18:34 |
| seb128 | mterry, no it doesn't | 18:34 |
| seb128 | mterry, what versions? ;-) | 18:34 |
| mterry | seb128, http://people.canonical.com/~platform/desktop/versions.html | 18:34 |
| mterry | seb128, does it check ftp.gnome.org directly? | 18:35 |
| seb128 | desrt, thanks for the update, I will try dropping the eventfd change on the next update and clean the unity patch ;-) | 18:35 |
| seb128 | mterry, the vcs is at the bottom of the page if you want to read the source | 18:35 |
| seb128 | mterry, basically it uses those: | 18:35 |
| seb128 | http://people.gnome.org/~vuntz/tmp/versions/versions-unstable | 18:35 |
| seb128 | http://people.gnome.org/~vuntz/tmp/versions/versions-unstable-extras | 18:35 |
| seb128 | that's for GNOME | 18:35 |
| seb128 | then url encoded in the sources for everything else | 18:36 |
| mterry | neat | 18:36 |
| seb128 | basically looking for the most recent tarball in defined dirs | 18:36 |
| seb128 | mterry, if you have any fix to do feel free to commit | 18:36 |
| mterry | seb128, no correction yet | 18:36 |
| seb128 | mterry, you will see the source has a list of sources and matching urls (or a special variable for GNOME tarballs which tells it to check vuntz's list) | 18:37 |
| seb128 | mterry, it's versions.py in the source which has the list and logic | 18:37 |
| seb128 | mterry, there is a special list to force a source to an url or serie | 18:38 |
| mterry | seb128, where is the source? | 18:39 |
| seb128 | we don't use it often but it can be useful if you want to not track an unstable serie for some reason | 18:39 |
| seb128 | mterry, see bottom of http://people.canonical.com/~platform/desktop/versions.html | 18:39 |
| seb128 | "You can download this script from lp:~ubuntu-desktop/+junk/versions" | 18:39 |
| seb128 | mterry, ^ | 18:39 |
| mterry | seb128, heh, never get down that far. I see green and I'm like "that's all" | 18:40 |
| seb128 | ;-) | 18:40 |
| seb128 | mterry, you got lazy, you didn't deserve to see the source :p | 18:40 |
| mterry | heh | 18:40 |
| seb128 | mterry, the script is quite hackish, it started from a small bit of code to list GNOME updates and we added things over time to it, but well it's mostly doing the job ;-) | 18:44 |
| bdrung | desrt: thanks for committing the units policy fixes to glib. | 18:57 |
| bdrung | desrt: it would be nice to have a extra long full version with base-2 and base-10 as shown in http://overbenny.files.wordpress.com/2011/01/nautilus-file-properties-libkibi.png | 18:57 |
| bdrung | desrt: second request: can you recycle the code from http://bazaar.launchpad.net/~libkibi-dev/libkibi/trunk/view/head:/kibi/kibi.c ? | 18:58 |
| bdrung | desrt: it has two advantages: 1. it rounds correctly (rounding to floating point and then round the division leads to rounding errors) and 2. it makes sure that at least three significant digits are present | 19:00 |
| bdrung | let me know if i should prepare a patch | 19:00 |
| mclasen | bdrung: don't make him push his luck | 19:01 |
| mclasen | :-) | 19:01 |
| bdrung | re significant digits: 1021 has four, 0.58 has two, 1.5 has two, 465.5 has four, and so on | 19:01 |
| bdrung | mclasen: isn't he glib dev? i think my proposed changes shouldn't lead to bikeshedding. | 19:02 |
| bdrung | mclasen: what do you think about my idea at http://overbenny.files.wordpress.com/2011/01/nautilus-file-properties-libkibi.png ? | 19:03 |
| mclasen | not going to comment; I said I'd hold my nose while desrt is committing this... | 19:04 |
| chrisccoulson | pitti - you'll probably notice soon enough, but thunderbird just grew by 3MB ;) | 19:12 |
| chrisccoulson | i'll fix it in the next upload though | 19:12 |
| desrt | bdrung: i'm happy to take rounding fixes | 19:12 |
| desrt | bdrung: i really am not in favour of the "show both units" approach. it's really confusing for users -- we should just always say SI units. | 19:12 |
| desrt | bdrung: as for 3 significant figures, i'm not sure how i feel. i think i prefer to just have 1 place after the decimal always. | 19:13 |
| chrisccoulson | hi desrt, how are you? | 19:14 |
| broder | desrt: SI being base-2 or base-10? | 19:16 |
| desrt | SI is base 10 | 19:16 |
| * broder nods | 19:17 | |
| desrt | chrisccoulson: pretty good | 19:17 |
| bdrung | desrt: one place after the decimal means two to four significant digits | 19:25 |
| desrt | bdrung: indeed... | 19:25 |
| bdrung | which i don't like - two significant digits are sometimes not enough | 19:29 |
| chrisccoulson | g'ah, all my PPA's keep running out of space - https://edge.launchpad.net/~mozillateam/+archive/thunderbird-next/+packages | 19:35 |
| micahg | chrisccoulson: yeah, all those PPAs should be 4GB, not 2GB | 19:36 |
| chrisccoulson | yeah, i guess i'm going to have to request more space again | 19:37 |
| chrisccoulson | they will hate me soon ;) | 19:37 |
| james_w | dobey, bug 813755 | 19:48 |
| ubot2 | Launchpad bug 813755 in gnome-keyring "gnome-keyring-daemon fails to start as it can't get capabilities" [Undecided,New] https://launchpad.net/bugs/813755 | 19:48 |
| dobey | james_w: thanks! | 19:50 |
| dobey | james_w: i wonder if maybe a kernel/udev/something upgrade resulted in a permissions change on the PRNG device or something? | 19:51 |
| james_w | maybe | 19:51 |
| james_w | I'm not sure which caps it is after | 19:51 |
| james_w | I assumed it was talking about the kernel capability system | 19:52 |
| dobey | not sure either. | 19:52 |
| desrt | bdrung: my concern is (the someone common) case of having a bunch of sizes lined up in a column | 19:52 |
| dobey | but if setuid "fixes" it for you, then i suspect a permissions issue somewhere | 19:52 |
| desrt | bdrung: in that case having the decimal spot in always the same spot makes the sizes a lot easier to scan | 19:52 |
| desrt | bdrung: of course that only works if the file sizes are of similar units | 19:53 |
| dobey | james_w: curious. is there anything in syslog or dmesg that looks related? :) | 19:56 |
| james_w | dobey, not sure | 19:57 |
| james_w | dobey, I'll test in a minute | 19:57 |
| * desrt disappears for the evening | 19:58 | |
| === scott-work_ is now known as scott-work | ||
| bdrung | desrt: yes, the units needs to be similar and aligned right | 20:11 |
| dobey | hrmm, is pitti the best person to ask about apport architecture stuff? | 20:16 |
| === m_conley is now known as m_conley_away | ||
| TheMuso | mterry: ah crap, thanks for the heads up. | 22:10 |
| jasoncwarner_ | morning everyone... | 22:11 |
| RAOF | Good morning. | 22:11 |
| RAOF | Isn't it a bit early for Adelaide? | 22:11 |
| jasoncwarner_ | RAOF: 7:42 here...just getting done with email. I usually try to get started around 6 my time to catch US folks afternoon. Especially east coast US, though, at least two of them are in boston so I don't know if I WANT to! | 22:13 |
| mterry | TheMuso, no worries, I uploaded a fix for the OnlyShowIn part, but left the rest as it wasn't obvious what to do with it | 22:14 |
| jasoncwarner_ | chrisccoulson: around? meant to ask you last night during TB meeting why firefox grew in size last week... | 22:14 |
| chrisccoulson | jasoncwarner_, the upstream default optimizations are broken | 22:16 |
| chrisccoulson | it should be fixed already though | 22:16 |
| chrisccoulson | (https://bugzilla.mozilla.org/show_bug.cgi?id=655003#c28) | 22:16 |
| ubot2 | Mozilla bug 655003 in Build Config "Sort out optimization defaults" [Normal,Resolved: fixed] | 22:16 |
| jasoncwarner_ | chrisccoulson: ah, so it was only a temporary jump in cd space ? | 22:17 |
| chrisccoulson | jasoncwarner_, yeah. i distro patched it to bring it back to normal | 22:17 |
| chrisccoulson | note, the new thunderbird just grew by the same size, but i'll apply the same patch there too | 22:18 |
| jasoncwarner_ | chrisccoulson: are they going to be properly fixing so you don't have to carry the patch forward? (if that would even be needed) | 22:18 |
| chrisccoulson | jasoncwarner_, it's already fixed in aurora (ie, firefox 7) | 22:18 |
| jasoncwarner_ | cool.... | 22:18 |
| chrisccoulson | i'm not sure if they will fix it in the current beta. the size increase doesn't affect them, as they turn on PGO for their builds | 22:18 |
| jasoncwarner_ | chrisccoulson: yeah, I'm using aurora...trying to see what we'll actually be shipping oneiric ;) (that is the version you told me, right?) | 22:19 |
| chrisccoulson | yeah :) | 22:19 |
| chrisccoulson | it's final release is 2 days before final freeze ;) | 22:19 |
| RAOF | Uuuuh… | 22:19 |
| RAOF | :) | 22:19 |
| chrisccoulson | the dates are here btw - https://wiki.mozilla.org/RapidRelease/Calendar | 22:20 |
| chrisccoulson | and the release dates are pretty much set in stone ;) | 22:20 |
| chrisccoulson | except firefox 9 | 22:20 |
| jasoncwarner_ | alrighty :) | 22:21 |
| bryceh | jasoncwarner_, RAOF, btw had a concall with Intel this morning, which included a discussion of the new SNA acceleration architecture | 22:33 |
| * RAOF is all ears. | 22:33 | |
| bryceh | they said it's not enabled by default yet, is not fully validated, and "not ready for prime time" | 22:34 |
| bryceh | they said they will be proceeding with testing/enabling but do not have a timeline for when it will be officially part of the infrastructure | 22:34 |
| bryceh | they're making it available to try out, but dont' consider it part of the release yet | 22:34 |
| bryceh | IOW, their advice is not to ship it yet. | 22:34 |
| RAOF | Did they mention what will happen to the existing UXA support once it SNA *is* ready for prime time? | 22:35 |
| bryceh | in addition to the validation todo's there are kernel regressions they know of that will need fixed. | 22:35 |
| bryceh | RAOF, 3 guesses? ;-) | 22:35 |
| RAOF | Poof! | 22:36 |
| bryceh | no, they didn't say specifically, and I was too polite to bring it up ;-) | 22:36 |
| bryceh | RAOF, but they did ask for feedback so I could follow up with a request they keep it at least through our LTS if you think that would be the thing to do | 22:37 |
| RAOF | At least for current cards that would be good. | 22:37 |
| bryceh | maybe something we need jasoncwarner to give us direction on | 22:38 |
| bryceh | mesa rc3 is coming on the 25th, release is july 29th. | 22:38 |
| RAOF | It wouldn't be difficult to have SNA only support new chips, like IvyBridge, build two intel drivers, and only autoload the SNA one on the new chips. | 22:39 |
| RAOF | rc3 on the 25th? rc2 has only just been cut! | 22:39 |
| bryceh | they're almost entirely focused on gpu hangs now it sounds like; ordinary rendering bugs sound like they're getting deferred for next release | 22:39 |
| bryceh | RAOF, yeah I still like the two driver approach. | 22:40 |
| bryceh | oh, also, they said to take full advantage of the mesa 7.11 work esp. for ivybridge, we'd need -intel 2.16 | 22:40 |
| bryceh | RAOF, is that on your todo list? | 22:41 |
| RAOF | It is once it's released. | 22:41 |
| RAOF | The 2.16 will be in the Q3 release? | 22:42 |
| bryceh | great | 22:42 |
| bryceh | that's the gist I got, might be worth doublechecking | 22:42 |
| bryceh | I haven't even been keeping track of what work was done there | 22:42 |
| RAOF | Lots of sna, a bit of IvyBridge, and some miscellaneous rendering fixes. | 22:44 |
| bryceh | doesn't sound too scary, assuming the sna bits are adequately segregated from the rest | 22:47 |
| RAOF | I haven't exhaustively checked them, but they live in a subdir that's only built when sna is enabled, so should be pretty well segregated. | 22:48 |
| RAOF | On the other hand, we could get approximately the same effect by cherry-picking the 3 or so commits that *aren't* SNA on to 2.15 :) | 22:49 |
| bryceh | heh | 22:57 |
| bryceh | I'd be down with that | 22:57 |
| bryceh | otoh, if we're going to ship dual -intel drivers, may as well have them both up to current so the sna variant has the bestest bits | 22:57 |
| RAOF | You think we should ship dual intel drivers now? | 22:58 |
| bryceh | RAOF, to do this would we have dual source packages, or could it be done by creating additional binary packages? | 22:58 |
| RAOF | Additional binary packages. | 22:58 |
| RAOF | brb, moving to somewhere where my hands can defrost. | 22:58 |
| bryceh | RAOF, well, it is looking pretty clear that we don't want sna enabled for oneiric, and conceivably not for the lts either; however users are going to demand having access to the performance improvements | 23:02 |
| bryceh | and like you said, for newer hardware, it may basically require sna | 23:03 |
| bryceh | so adding an xserver-xorg-video-intel-sna seems like the most sensible solution | 23:03 |
| RAOF | and there's no particular reason not to do that now, right. | 23:07 |
| bryceh | jasoncwarner_, any thoughts on the above plan? | 23:07 |
| jasoncwarner_ | bryceh: reading backscroll, one sec | 23:11 |
| jasoncwarner_ | bryceh: to recap, we might not have sna for oneiric, or P either? | 23:13 |
| bryceh | jasoncwarner_, the plan would be to ship a separate driver with it enabled, that users could opt-in to if they want it | 23:14 |
| bryceh | jasoncwarner_, but by default we would be shipping UXA, not SNA | 23:14 |
| jasoncwarner_ | that seems reasonable, but what would all that mean for LTS in terms of drivers and what not? do we have a mechanism to get LTS uses latest and greatest drivers and fixes? | 23:15 |
| bryceh | well, we'd want to stick with the same arrangement for LTS if possible | 23:15 |
| bryceh | UXA by default, with SNA for opt-in | 23:15 |
| bryceh | but it is a good point that post-LTS, if we want to do driver updates or even bugfix backports, we might only be able to do that for the SNA driver | 23:16 |
| bryceh | jasoncwarner_, however... | 23:16 |
| bryceh | jasoncwarner_, SNA is only a concern of the 2D ddx driver. We could still ship fixes and updates to mesa and the kernel, SNA or no SNA | 23:17 |
| bryceh | and mesa/kernel is where I expect most hardware support updates to come from anyway | 23:17 |
| RAOF | Alhough it's entirely possible that some mesa features will be tied to SNA. | 23:18 |
| bryceh | RAOF, possibly yeah | 23:18 |
| jasoncwarner_ | bryceh RAOF I might have missed that...UXA is going away sometime post LTS? | 23:18 |
| jasoncwarner_ | (you said we might only be able to update SNA driver?) | 23:18 |
| bryceh | jasoncwarner_, well no one has explicitly said it's going away, but Intel has a history of aggressively demoting legacy acceleration architectures | 23:19 |
| RAOF | UXA will certainly get less love once SNA reaches prime time. Based on past behaviour, UXA will also be dropped from the tree after a while, too. | 23:19 |
| RAOF | Since we can't reasonably switch default drivers in an SRU I'd suggest that (at least) IvyBridge+ default to the SNA driver. | 23:20 |
| bryceh | that seems reasonable | 23:21 |
| RAOF | Since that's new hardware it's not possible for it to be a regression :) | 23:21 |
| bryceh | :-) | 23:21 |
| jasoncwarner_ | yeah, from my limited scope, it does...though we should be sure to check with pitti.. | 23:21 |
| RAOF | I'm not sure, but we may also want to do the same for SandyBridge. They're likely to get the most driver development investment. | 23:22 |
| TheMuso | mterry: Again thanks, thats fixed. Its things like this that make me think less of using a patch system in the form of v3 quilt. | 23:23 |
| TheMuso | Because you can't just drop the patch from the directory, and assume its gone away. | 23:24 |
| TheMuso | But I guess I haven't had to do that in enough packages yet to make popping the patches first a habbit. | 23:24 |
| bryceh | alrighty, sounds like a plan | 23:26 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
