didrocks | good morning | 05:36 |
---|---|---|
Saviq | hello | 06:29 |
oSoMoN | good morning | 07:05 |
=== API is now known as Guest68170 | ||
=== Guest68170 is now known as apinheiro | ||
didrocks | kamstrup: so, the crash I can get easily has been duped to bug #840758 | 07:21 |
ubot5 | Launchpad bug 840758 in unity (Ubuntu) "compiz crashed with SIGSEGV in dee_model_get_tag()" [Medium,Confirmed] https://launchpad.net/bugs/840758 | 07:21 |
didrocks | kamstrup: if you are interested, I got it in https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/851628 | 07:22 |
ubot5 | Ubuntu bug 840758 in unity (Ubuntu) "duplicate for #851628 compiz crashed with SIGSEGV in dee_model_get_tag()" [High,Triaged] | 07:22 |
didrocks | (dee_sequence_model_find_tag()) | 07:22 |
kamstrup | didrocks: looking | 07:22 |
=== chrisccoulson_ is now known as chrisccoulson | ||
apinheiro | njpatel, I have a easy branch to review (two lines), could you take a look ? | 07:56 |
apinheiro | https://code.launchpad.net/~apinheiro/unity/Bug851103/+merge/75591 | 07:56 |
njpatel | of course, looking now | 07:56 |
njpatel | apinheiro, approved, no need to wait for alex, feel free to merge it | 07:57 |
apinheiro | njpatel, ok, thanks | 07:57 |
didrocks | kamstrup: where did you see it has been freed? | 07:58 |
didrocks | kamstrup: oh, the tag address? | 07:59 |
kamstrup | didrocks: yeah, segv analysis | 07:59 |
kamstrup | didrocks: of course, it's just a guess | 07:59 |
kamstrup | didrocks: no, the tag address is actually a stuffed uint, so that one is ok | 08:00 |
didrocks | just a guess for the priv address then :) | 08:00 |
kamstrup | didrocks: but the segv analysis says: source "(%edx)" (0x00000006) not located in a known VMA region (needed | 08:00 |
kamstrup | didrocks: and on the crashing line we're accessing a GPtrArray 'row' | 08:01 |
kamstrup | didrocks: so it looks like 'row' may have been freed | 08:01 |
didrocks | kamstrup: yeah, indeed, but you don't know if it's self, out_row_tag or out_tag which is invalid, isnt it? | 08:03 |
didrocks | ah, the crashing line | 08:03 |
didrocks | ok, then, seems the row :) | 08:04 |
* didrocks didn't open the source | 08:04 | |
kamstrup | didrocks: line 788 looks like this: | 08:04 |
kamstrup | row_tag_iter = g_ptr_array_index (row, n_cols); | 08:04 |
kamstrup | that's what gives the segv | 08:04 |
didrocks | yeah, I doubt n_cols is a pointer :) | 08:04 |
kamstrup | so really, only options is that 'row' is bad | 08:05 |
didrocks | indeed | 08:05 |
kamstrup | indeed | 08:05 |
kamstrup | indeed indeed | 08:05 |
didrocks | kamstrup: so, in this case, (the last crash I got) | 08:05 |
didrocks | it wasn't on a search result, it was when I expanded the result | 08:05 |
kamstrup | didrocks: with this trace? | 08:05 |
didrocks | kamstrup: well, from the duplicated one, yeah: https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/851628 | 08:06 |
ubot5 | Ubuntu bug 840758 in unity (Ubuntu) "duplicate for #851628 compiz crashed with SIGSEGV in dee_model_get_tag()" [High,Triaged] | 08:06 |
didrocks | same line | 08:06 |
kamstrup | didrocks: i'll try adding some more fine grained tests to dee, and maybe add some clever detecting for access to freed values | 08:08 |
didrocks | kamstrup: indeed, but it would be good to know what freed those first | 08:08 |
didrocks | weird that I never ever triggered that and only with the new music lens, on u1 search | 08:09 |
kamstrup | didrocks: I've added a bunch of unit tests to dee but it seems rock solid. I also added some sanity checks in the code that should help pinpoint any problems | 09:02 |
didrocks | kamstrup: nice! but that only workarounds this issue isn't it? | 09:03 |
kamstrup | didrocks: do you want a release? We already have a heavily distro patched dee, and there is a fixed GI bug that hasn't been released either | 09:03 |
kamstrup | didrocks: it doesn't work around this issue, but it will help a great deal in nailing it down if we can't do it today | 09:03 |
didrocks | kamstrup: well, all upload needs to be approved and sign with blood, if it can't fix the issue but add more debugging for it, I'm happy to push to a ppa? | 09:04 |
didrocks | kamstrup: I mean, the user will still get the crash, isn't it? | 09:04 |
kamstrup | didrocks: to be precise: if/when we see the crash with this we can say with 99.9999% procent that it's because unity accesses a freed row | 09:05 |
didrocks | or just won't get the row because you check if it's freed? | 09:05 |
kamstrup | didrocks: or if you get the crash yourself, you can just run lp:dee/0.5 locally yourself | 09:05 |
njpatel | guys, it's unity :) | 09:05 |
didrocks | kamstrup: so, maybe it's better that I package that, keep debug symbols and report a stacktrace? | 09:05 |
kamstrup | didrocks: if you can get a trace with lp:dee/0.5 then that would be awesome | 09:05 |
kamstrup | didrocks: and capture xsessionerrors as well | 09:06 |
didrocks | kamstrup: sure | 09:06 |
kamstrup | njpatel: do you get the crasher yourself? | 09:06 |
didrocks | kamstrup: on it | 09:06 |
kamstrup | njpatel: yeah, but I added some sanity checks to dee to help pinpoint what exactly unity is doing :-) | 09:07 |
didrocks | kamstrup: hum, is it building? | 09:11 |
didrocks | waow, this is a weird linkage issue | 09:11 |
didrocks | kamstrup: http://paste.ubuntu.com/690632/ | 09:12 |
kamstrup | njpatel: i'll merge cimi's darken branch, ok? | 09:14 |
njpatel | kamstrup, yep, please | 09:15 |
didrocks | kamstrup: libdee-1.0.la isn't built, the message is real | 09:16 |
njpatel | "the message is real". | 09:16 |
didrocks | njpatel: I meant, it's not a path thing or the compiler is lying :) | 09:17 |
njpatel | THE COMPILER ALWAYS LIES | 09:17 |
kamstrup | >njdone | 09:17 |
kamstrup | "njpatel: done" | 09:18 |
kamstrup | didrocks: wtf?! | 09:19 |
didrocks | kamstrup: wondering, did I tell you that I always hate those .la stuff? :) | 09:19 |
didrocks | kamstrup: trying in a pbuilder, maybe I have a local .la installed which triggered a "I can't build .la files" | 09:19 |
kamstrup | didrocks: so are you compiling in some weird packaging setup or something? | 09:20 |
didrocks | kamstrup: no, we normally build everything, and at the end, we remove the .la file | 09:20 |
didrocks | kamstrup: but that's on the install part | 09:20 |
kamstrup | didrocks: i just double checked a clean lp:dee/0.5 checkout and it compiles without a hitch here | 09:21 |
didrocks | kamstrup: yeah, let's see how the pbuilder build will go | 09:22 |
greyback | mardy: hey I did a fresh checkout of your branch, and I can still reproduce that bug :( | 09:25 |
mardy | greyback: hi! | 09:25 |
greyback | mardy: anything I can do to help? | 09:25 |
mardy | greyback: maybe yes: when I first requested the merge, I wrote about a few lines of code to add in order to visually debug the issue | 09:26 |
mardy | greyback: you could add those lines, and take the video again | 09:26 |
greyback | mardy: no prob | 09:27 |
mardy | though I'm not sure it will be of help | 09:27 |
didrocks | kamstrup: FTBFS in a pbuilder as well | 09:31 |
kamstrup | didrocks: and there is no other error message anywhere? | 09:32 |
didrocks | kamstrup: no, it just doesn't build the .la file | 09:32 |
didrocks | kamstrup: one sec | 09:32 |
kamstrup | didrocks: and it's not because of some conflict with the distro patches? Ken put a tonne in there from trunk in order to get gwibber running | 09:32 |
didrocks | http://paste.ubuntu.com/690641/ | 09:33 |
AlanBell | I made a little video of using Unity in Oneiric with Orca http://people.ubuntu.com/~alanbell/unityorca.ogv not great. | 09:33 |
didrocks | kamstrup: I tried to rebuild the previous version (the one in Oneiric), same thing, I guess the issue is in something triggering/not triggering the .la buid | 09:33 |
didrocks | kamstrup: see my pastebin for the full build log | 09:34 |
didrocks | kamstrup: the previous successful build seems ok though https://launchpadlibrarian.net/75306938/buildlog_ubuntu-oneiric-i386.dee_0.5.18-1ubuntu1_BUILDING.txt.gz | 09:34 |
kamstrup | didrocks: not building the .la will definitely cause GI to fail as dee/Makefile.am declares --library=dee-1.0.la hard coded | 09:39 |
didrocks | kamstrup: indeed, I saw that | 09:39 |
kamstrup | I don't know if/how to solve that... | 09:39 |
didrocks | kamstrup: I'm just wondering why it doesn't, the last successfull build has exactly the same output | 09:40 |
didrocks | kamstrup: yeah, neither do I, I just know that .la are a PITA and it depends on other .la file in the distro | 09:40 |
kamstrup | didrocks: just toying with a solution, hang on | 09:42 |
kamstrup | didrocks: pull lp:dee/0.5 | 09:45 |
didrocks | kamstrup: ahah nice trick! | 09:50 |
kamstrup | didrocks: i am breathlessly awaiting the trumpets heralding the success... | 09:51 |
didrocks | kamstrup: succes! | 09:52 |
didrocks | sucess* | 09:52 |
didrocks | even | 09:52 |
didrocks | success* | 09:52 |
didrocks | grrrrr :p | 09:52 |
didrocks | kamstrup: so, making it crash with that, one sec | 09:52 |
didrocks | kamstrup: will you need the backtrace again? | 09:53 |
kamstrup | didrocks: trace and xsessionerrors | 09:53 |
didrocks | oki | 09:53 |
* didrocks rebuilds quickly without stripping the symbols | 09:54 | |
didrocks | kamstrup: interesting, so I have no crash from the music lens | 10:01 |
didrocks | kamstrup: but no cover | 10:01 |
kamstrup | didrocks: so you have generic icons for covers, or? | 10:01 |
didrocks | yeah, generic icons | 10:01 |
didrocks | and nothing in .xsession-errors | 10:02 |
kamstrup | didrocks: not impossible that the crasher is related to icon loading in fact | 10:02 |
didrocks | kamstrup: yeah, that would totally makes sense with what I observed | 10:02 |
kamstrup | didrocks: because the icon loading is async, it might be that we get some callback on the icons *after* a row has been removed, trying to access a tag | 10:02 |
didrocks | kamstrup: ahah, yeah! well thought :-) | 10:03 |
kamstrup | didrocks: after a quick look at the unity code I am willing to bet this is it :-) | 10:04 |
kamstrup | gord: ^^ ? | 10:04 |
didrocks | kamstrup: it couldn't be dee, it's never dee! :-) | 10:04 |
kamstrup | gord any chance that the icon loader races in its callback to ResultRendererTile::IconLoaded with the removal of rows - making it access a removed DeeModelIter, and thus a freed DeeModelTag? | 10:06 |
gord | kamstrup, nope, unless there is a async timeframe between a row being removed from the model and the row-removed signal being called | 10:08 |
kamstrup | gord: there's not | 10:09 |
gord | kamstrup, we'll do a mumble about this once neil starts falling over, you should hop on | 10:09 |
kamstrup | gord: Dee.Model enmits row-removed and afterwards removes the row, in the same stack frame | 10:09 |
kamstrup | gord: here's the race i imagine: | 10:10 |
kamstrup | 1) user types "foo" | 10:10 |
kamstrup | 2) model populates | 10:10 |
kamstrup | 3) start loading icons | 10:10 |
kamstrup | 4) user quickly deletes an o -> "fo" | 10:11 |
kamstrup | 5) icon loading resulots from 3 arrives | 10:11 |
kamstrup | 6) kaboom | 10:11 |
gord | kamstrup, nope | 10:11 |
gord | kamstrup, when a row is removed, its icon loader slot is removed | 10:11 |
kamstrup | gord: there's no spin of the mainloop in the async icon loading? | 10:11 |
kamstrup | gord: oh | 10:12 |
kamstrup | gord: the stack trce on https://bugs.launchpad.net/ubuntu/+source/unity/+bug/840758 seems to suggest you might not be correct in all circumstances | 10:13 |
ubot5 | Ubuntu bug 840758 in unity (Ubuntu) "compiz crashed with SIGSEGV in dee_model_get_tag()" [High,Triaged] | 10:13 |
kamstrup | this also makes perfect sense with that it's the slowly loaded http covers that triggers the bug | 10:14 |
gord | kamstrup, yup seen those, but literally i can't see any logic where that could happen, not dismissing the bug, its just not as obvious as you might think ;) | 10:15 |
* kamstrup hates non-obvious bugs | 10:15 | |
kamstrup | :-) | 10:15 |
greyback | AlanBell: thanks for the video. You're using Unity-2D there, and yes accessibility still needs work. | 10:16 |
gord | course if this was gobject i could just FOO_IS_BAR on the damn thing and it would handle it all fine, but oh well ;) | 10:16 |
AlanBell | greyback: dammit, thought that was 3d! | 10:22 |
AlanBell | oh, 3d won't run in a live session, need guest additions for the drivers still | 10:23 |
greyback | AlanBell: yes unfortunately | 10:24 |
Kaleo | AlanBell: thanks for the video! | 10:24 |
AlanBell | I will do another video with it installed and running 3d | 10:24 |
Kaleo | AlanBell: https://bugs.launchpad.net/unity-2d/+bugs?field.tag=a11y lists all the known accessibility bugs | 10:24 |
Kaleo | AlanBell: I have not finished watching the video but so far I think all the ones you spotted are reported and on the way to be fixed | 10:28 |
AlanBell | I filed a couple, I will reassign to unity-2d or close if they are duplicates | 10:29 |
Kaleo | AlanBell: thank you! | 10:29 |
AlanBell | wish we had the drivers for virtualbox guest on the CD | 10:30 |
mhr3 | didrocks, is the desktop.*.cache file still used in oneiric? | 10:41 |
didrocks | mhr3: yeah, the patch has been ported AFAIK, but better to check with pitti on Monday | 10:42 |
mhr3 | didrocks, ported in what way? | 10:42 |
didrocks | mhr3: ported to GNOME3? | 10:42 |
mhr3 | ah, k | 10:43 |
Kaleo | mardy: I reviewed https://code.launchpad.net/~mardy/unity-2d/rtl/+merge/71305 | 10:54 |
Kaleo | mardy: thanks! | 10:54 |
mardy | Kaleo: great! I'll fix those things you mentioned, but the other two would be better reported as bugs, just to get this thing in :-) | 10:59 |
Kaleo | mardy: absolutely fine | 11:01 |
Kaleo | mardy: it's going to require a freeze exception to land that in Oneiric. Can you please make sure https://wiki.ubuntu.com/FreezeExceptionProcess is followed? | 11:02 |
mardy | Kaleo: I'll try :-) | 11:02 |
Kaleo | see "UserInterfaceFreeze Exceptions" | 11:03 |
Kaleo | mardy: thank you | 11:03 |
=== MacSlow is now known as MacSlow|lunch | ||
=== apinheiro is now known as apinheiro_lunch | ||
=== apinheiro_lunch is now known as apinheiro | ||
jbicha | JohnLea: with bug 764744 are you intending to remove the System Settings link from the Session menu, or merely add System Settings to the launcher? | 13:14 |
ubot5 | Launchpad bug 764744 in unity "[UIFe] Add system setting icon to Launcher" [High,New] https://launchpad.net/bugs/764744 | 13:14 |
jjardon | tedg: hi, seems that https://code.launchpad.net/~jjardon/indicator-datetime/fix-837440 was not merged but Its marked as that | 13:18 |
jjardon | tedg: also, I'd like to request a package for indicator-power, Is it kenvandine on holidays? | 13:20 |
tedg | jjardon, Yeah, he is. | 13:31 |
tedg | jjardon, The status is merged... | 13:31 |
mardy | greyback: hi, what happened to your MR about alf+f2? lp now cannot find the page | 13:32 |
greyback | mardy: I just saw what you were doing, and realised your fix was better :) | 13:33 |
mardy | greyback: I disagree :-) | 13:33 |
greyback | mardy: hmmm | 13:34 |
mardy | greyback: what's wrong with your fix? It's simpler and more clear | 13:34 |
greyback | mardy: true, but it requires reloading the lens, no? | 13:34 |
greyback | I figured your way was more efficient | 13:34 |
mardy | greyback: mmm... maybe | 13:35 |
jjardon | tedg: grep -r GnomeWall * doesnt return nothing here | 13:35 |
greyback | mardy: I'll resubmit it and let Florian choose? | 13:35 |
tedg | jjardon, Yes, so I merged it but had to back it out as the GNOME Wall clock isn't in the GTK2 version of the gnome libraries. | 13:35 |
tedg | jjardon, Also, I looked at the GNOME Wall clock code and it's doing the same things we are doing right now. | 13:35 |
mardy | greyback: yep | 13:37 |
mardy | greyback: but I really like yours better; even if mine were more efficient, it just seems reasonable to reset that field when deactivateLens is called | 13:38 |
jjardon | tedg: do we support timerfd? | 13:39 |
jjardon | http://git.gnome.org/browse/gnome-desktop/commit/?id=5f65d7d58188ffc75ffded71c741892352411a0c | 13:39 |
greyback | mardy: ok so. I'll do merge request again (sorry for the noise). | 13:39 |
mardy | greyback: np. BTW, I want to ask you something: | 13:40 |
greyback | mardy: shoot | 13:41 |
mardy | greyback: press alt-f2, type "gedit", press enter | 13:41 |
mardy | then close gedit end repear | 13:41 |
mardy | does it work all the time? | 13:41 |
mardy | greyback: for me it works every other time | 13:41 |
mardy | greyback: bus I see with dbus-monitor that we are activating the lens item every time | 13:41 |
greyback | mardy: yeah you're right, it only works every second time. Hmmm | 13:42 |
tedg | jjardon, Not sure, but I think it makes more sense to wait for the GLib API (which would be stable) than using an unstable API. | 13:43 |
mardy | greyback: I went to debug unity-applications-daemon, but then I noticed that when it's started from the command line it always works :-O | 13:43 |
greyback | mardy: yuk. Did you see if it happens in Unity? | 13:45 |
mardy | greyback: no, I wrote to Mikkel (kamstrup) and he said that it works fine in Unity | 13:46 |
Kaleo | greyback, mardy: I am not sure what you guys are doing but one thing: the lenses are supposed to be stateful | 13:47 |
mardy | greyback: unfortunately I don't find him online ATM | 13:47 |
Kaleo | (they are not really right now) | 13:47 |
Kaleo | that is, keeping the current search and filters | 13:47 |
mardy | Kaleo: yep, the fix will make them more stateful :-) | 13:48 |
Kaleo | mardy: good :) | 13:50 |
greyback | mardy: I just tried Unity (couple of days old, updating now) & get the same problem. | 13:52 |
mardy | greyback: good to know, thanks! I'll search LP to see if there's a bug reported | 13:52 |
tedg | jjardon, I think we could take the timerfd stuff if it's in the Ubuntu kernel, but I don't really want to depend on the unstable GTK3 only libs. | 13:55 |
greyback | mardy: no prob, that's a good catch | 13:55 |
jjardon | tedg: unstable GTK3 libs? | 13:58 |
tedg | jjardon, libgnome-desktop | 13:59 |
tedg | jjardon, The API for the wall clock is behind the UNSTABLE build flag, right? | 14:00 |
tedg | jjardon, Not unstable like buggy, unstable API | 14:00 |
Trevinho | tedg: this is something you should look I guess: https://bugs.launchpad.net/indicator-appmenu/+bug/705653 | 14:02 |
ubot5 | Ubuntu bug 705653 in unity (Ubuntu) "Menus for parent window are still present when dialog is focused" [High,Triaged] | 14:02 |
tedg | Trevinho, Yeah | 14:04 |
tedg | I forgot about that one :-/ | 14:04 |
jjardon | tedg: ah, yeah. Well we can ifdef the code if you want and use GnomeWallClock only if It's available. Using GnomeWallClock fix quite a lot of bugs in indicator-datetime | 14:06 |
Trevinho | tedg: read my last comment.. .I also experienced that when working with panel-service to add some optimizations... | 14:06 |
Trevinho | the dobule entry-added/removed signal was creating problems, so I had to use a workaround | 14:07 |
mardy | greyback: seems to be a gedit bug, other apps work fine | 14:07 |
tedg | jjardon, It seems like it'd be a smaller patch to just use timerfd directly, no? | 14:07 |
greyback | mardy: why didn't I try other apps??? gah! | 14:08 |
tedg | Trevinho, Yeah, I'm more worried about showing the menus. They should be insensitive... | 14:08 |
tedg | Trevinho, But, yeah, we could optimize that case. | 14:08 |
tedg | Trevinho, Just in the code when it detects a switch it doesn't handle that case well. | 14:08 |
JohnLea | jbicha; the system settings link will remain in the Session menu, the change in that bug is in addition, not instead of | 14:12 |
Trevinho | tedg: I didn't look at the code (sorry), but maybe a small timeout on Window switch could help... | 14:17 |
mardy | greyback: https://bugs.launchpad.net/unity/+bug/840097 | 14:17 |
ubot5 | Ubuntu bug 827414 in gedit (Ubuntu Oneiric) "duplicate for #840097 gedit fails to start on first try, but does on the second" [Medium,Confirmed] | 14:17 |
greyback | mardy: how bizarre | 14:18 |
Trevinho | Hi JohnLea please give a look to this too: https://bugs.launchpad.net/bugs/818788 ;) | 14:19 |
ubot5 | Ubuntu bug 818788 in unity-2d (Ubuntu) "The sequence of the applets should be easily editable by the user." [Wishlist,Incomplete] | 14:19 |
jbicha | JohnLea: thanks, since we're adding launcher items, how about bug 778289? :-) | 14:20 |
ubot5 | Launchpad bug 778289 in unity (Ubuntu) "Help button in Unity default launcher" [Undecided,Incomplete] https://launchpad.net/bugs/778289 | 14:20 |
tedg | Trevinho, Well, FYI, it's this function, I think the comment pretty much addresses your issue: http://bazaar.launchpad.net/~indicator-applet-developers/indicator-appmenu/trunk/view/head:/src/indicator-appmenu.c#L1085 | 14:20 |
Kaleo | mardy: you seem to well understand what greyback did, how about you take care of the review? | 14:22 |
mardy | Kaleo: I'll approve it, then: for me it's fine | 14:24 |
Kaleo | mardy: great, thanks! | 14:25 |
Kaleo | mardy: I'm removing https://bugs.launchpad.net/ubuntu/+source/unity-2d/+bug/821919 from the milestone | 14:25 |
ubot5 | Ubuntu bug 821919 in unity-2d (Ubuntu) "Oneiric (live cd only): Alt+F2 sometimes not working in Unity 2D" [Medium,Triaged] | 14:25 |
mardy | Kaleo: no complaints from me :-) | 14:28 |
Kaleo | :) | 14:31 |
Kaleo | mardy, greyback, Saviq: when you send a MR out please make sure that somebody specific is assigned as a reviewer | 14:35 |
mardy | Kaleo: ack | 14:35 |
Kaleo | Saviq: replied with a fix to https://code.launchpad.net/~saviq/unity-2d/fix-panel-leak/+merge/75744 | 14:36 |
mardy | Saviq: BTW, welcome! :-) | 14:38 |
Kaleo | chrisccoulson: for https://bugs.launchpad.net/ubuntu/+source/nautilus/+bug/804435 do you think we'll be able to get into Oneiric? | 14:40 |
ubot5 | Ubuntu bug 804435 in gtk+3.0 (Ubuntu) "Wallpaper is loaded twice with different alignment by gnome-session and nautilus (Oneiric)" [Low,Triaged] | 14:40 |
=== MacSlow|lunch is now known as MacSlow | ||
chrisccoulson | Kaleo, sure. i'd like upstream to comment first though | 14:40 |
Kaleo | chrisccoulson: great! | 14:40 |
Kaleo | Cimi: any idea about https://bugs.launchpad.net/ubuntu/+source/nautilus/+bug/839611 ? | 14:44 |
ubot5 | Ubuntu bug 839611 in unity-2d "menu is first displayed as part of nautilus' desktop window before being hosted by the top panel thus shifting the wallpaper during the startup sequence" [Medium,Confirmed] | 14:44 |
Cimi | Kaleo: yup | 14:44 |
Cimi | Kaleo: that's a bug somewhere in my gtk patch | 14:45 |
Cimi | Kaleo: I might need to rewrite it | 14:45 |
Cimi | Kaleo: the thing is that out current design fails with gtk+ architecture, that's why I couldn't come up with a solution to this yet | 14:46 |
Cimi | Kaleo: but I had an idea, probably worth spending some time trying that | 14:46 |
Kaleo | Cimi: I kind of see :) | 14:50 |
Kaleo | Cimi: what's your gtk patch? | 14:50 |
Cimi | Kaleo: the ubuntu menubar patch | 14:54 |
Cimi | Kaleo: what is different from natty is that I'm hiding the menubar widget | 14:55 |
Cimi | Kaleo: that works better in unity and unity-2d, but nautilus is suffering :-\ | 14:55 |
Kaleo | Cimi: I understand better | 15:00 |
Kaleo | Cimi: do you intend to fix it soon? | 15:00 |
Kaleo | chrisccoulson: do you think https://bugs.launchpad.net/ubuntu/+source/nautilus/+bug/839610 has the same cause (and patch in GTK) as https://bugs.launchpad.net/ubuntu/+source/nautilus/+bug/804435 ? | 15:01 |
ubot5 | Ubuntu bug 839610 in unity-2d "Background wallpaper briefly appears shifted horizontally by around 15 pixels on startup" [Medium,Confirmed] | 15:01 |
ubot5 | Ubuntu bug 804435 in gtk+3.0 (Ubuntu) "Wallpaper is loaded twice with different alignment by gnome-session and nautilus (Oneiric)" [Low,Triaged] | 15:01 |
chrisccoulson | Kaleo, yeah, i think they're the same | 15:02 |
Kaleo | chrisccoulson: thanks a bunch | 15:02 |
Cimi | Kaleo: maybe next week I'll have a quick look | 15:05 |
Kaleo | Cimi: thanks, it's quite important | 15:07 |
Kaleo | Cimi: I'll mark GTK as affected and assign it to you | 15:07 |
Cimi | Kaleo: there's already a bug filed for that | 15:11 |
Kaleo | Cimi: oh, great, what's the bug number? | 15:11 |
Cimi | https://bugs.launchpad.net/bugs/717358 | 15:11 |
ubot5 | Ubuntu bug 717358 in gtk+2.0 (Ubuntu Natty) "nautilus puts a menu bar at the top of the desktop" [Low,Fix released] | 15:11 |
Kaleo | Cimi: thanks | 15:11 |
Cimi | Kaleo: now affects gtk+3 too | 15:11 |
Kaleo | Cimi: should we reopen it then? | 15:13 |
Cimi | Kaleo: for oneiric | 15:16 |
Cimi | Kaleo: not sure I can fix that, though | 15:16 |
Cimi | Kaleo: Ill try with a workaround | 15:16 |
Kaleo | Cimi: it makes the login sequence very ugly | 15:17 |
Cimi | Kaleo: :-\ | 15:19 |
Cimi | Kaleo: i didn0t write that patch | 15:20 |
Cimi | Kaleo: i just tweaked to hide the menubar | 15:20 |
Kaleo | Cimi: who wrote the patch in question? | 15:20 |
Cimi | Kaleo: I was aware of that bug, but hiding the menubar has benefits for the theme | 15:20 |
Cimi | Kaleo: the gtk menubar patch was written by cody russel | 15:20 |
Kaleo | Cimi: I see | 15:20 |
Cimi | Kaleo: our mistake is that we show the menubar if the appmenu is not running, for fallback | 15:21 |
Kaleo | Cimi: even for the Nautilus desktop window I guess? | 15:21 |
Cimi | Kaleo: we should instead remove the menubar totally, like switching it off for the whole experience, and leave it enabled just in classic mode | 15:21 |
Cimi | Kaleo: the nautilus desktop window is just nautilus I guess, we can't hide for it | 15:22 |
Cimi | Kaleo: otherwise nautilus won't have the menubar | 15:22 |
Kaleo | Cimi: should not Nautilus have a piece of code that makes sure no menubar is created for the desktop window? | 15:22 |
Cimi | maybe | 15:22 |
Kaleo | tedg: what do you think? ^ | 15:29 |
Kaleo | tedg: https://bugs.launchpad.net/gtk/+bug/839611 | 15:29 |
ubot5 | Ubuntu bug 839611 in unity-2d "menu is first displayed as part of nautilus' desktop window before being hosted by the top panel thus shifting the wallpaper during the startup sequence" [Medium,Confirmed] | 15:29 |
tedg | Kaleo, Yes, there's a bug to update the patch for that. | 15:30 |
Kaleo | tedg: great! do you have the number? | 15:31 |
tedg | Kaleo, bug 805252 | 15:35 |
ubot5 | Launchpad bug 805252 in nautilus (Ubuntu Oneiric) "appmenu shows nautilus menu when desktop is displayed" [High,Triaged] https://launchpad.net/bugs/805252 | 15:35 |
Kaleo | tedg: thanks a lot! | 15:36 |
=== yofel_ is now known as yofel | ||
htorque | is there a small 101 on how to compile unity-2d from source? | 16:31 |
andyrock | htorque, https://wiki.ubuntu.com/Unity2D | 16:35 |
andyrock | btw it's quite easy | 16:35 |
andyrock | cmake . | 16:35 |
andyrock | make | 16:35 |
andyrock | ;) | 16:35 |
htorque | andyrock: oops, i think i should have found that via google :D | 16:37 |
htorque | thanks! ;) | 16:37 |
andyrock | htorque, ubuntu wiki is not so google friend | 16:38 |
htorque | andyrock: there's nothing i need to build prior to unity-2d? | 16:39 |
andyrock | no idea about this | 16:42 |
andyrock | htorque, btw i think sudo apt-get build-dep unity-2d | 16:43 |
htorque | i'll see if it explodes or not ;-) | 16:44 |
andyrock | ;) | 16:44 |
Ndhb | hi! I want to start developing and porting lenses for unity in python, where do I start? | 17:04 |
gord | Ndhb, great! https://wiki.ubuntu.com/Unity/Lenses should help get you started | 17:11 |
macer1 | Hi. I have a question...hmm...unity 2d starts looking better than 3d version...it is faster, have more animations, and gives better user experience...can 3d version be rewriteen in qt like unity 2d? | 18:37 |
macer1 | and unity 2d and 3d will be perfect | 18:37 |
macer1 | *merge of unity2d and 3d | 18:37 |
macer1 | like kde | 18:41 |
macer1 | you can run it with or without hardware acceleration | 18:41 |
macer1 | unity 3d looks now really bad compared to unity 2d :( | 18:42 |
macer1 | http://ubuntuone.com/4hJIk1CeAboeNXsKv0T1ZP - wow! unity looks great! i want to use it! no...wait...it is unity 2d...so no compiz... | 18:46 |
davidcalle | macer1, you can still use compiz with unity-2d ;-) | 18:47 |
macer1 | i know, but it is not the same... | 18:48 |
macer1 | please, do something with it :( | 18:48 |
macer1 | even scroolbar looks better in unity 2d... | 18:48 |
AlanBell | so use unity2d! | 18:54 |
macer1 | no | 18:55 |
macer1 | because i want ubuntu to look cool out of the box :) | 18:55 |
AlanBell | 2d is out of the box | 18:55 |
macer1 | but there isnt compiz integration OOTB... | 18:56 |
macer1 | user installs ubuntu. oh unity looks so cool!! installs propiertary drivers. reboot. and then "why is unity looking so bad now :( ?" | 18:57 |
AlanBell | hmm, I don't think it looks worse | 18:58 |
AlanBell | I think 2d is surprisingly good, and I often get confused about which one I am looking at | 18:58 |
davmor2 | AlanBell: just look to the rubbish bin my friend | 18:59 |
c10ud | macer1, you could ask why duplicating the effort, but if they're doing it, i guess they have a reason for it so.. | 18:59 |
macer1 | of couse in oneric, unity 2d looks good in oneiric | 18:59 |
AlanBell | I don't think 2d was expected to be so good and I don't think 3d was expected to be so problematic | 19:00 |
macer1 | i think main reason unity 3d to be gtk is gnome... | 19:00 |
macer1 | but unity 2d is made for gnome, too... | 19:00 |
AlanBell | it is nux | 19:00 |
macer1 | nux + gtk3? | 19:01 |
AlanBell | davmor2: so in 3d the trashcan is at the bottom of the launcher | 19:01 |
AlanBell | and it is in 2d | 19:03 |
AlanBell | and they both blur the background | 19:03 |
AlanBell | 2d is a fraction darker than 3d I think | 19:03 |
macer1 | nux isnt cool | 19:05 |
nhaines | More than a fraction. | 19:05 |
macer1 | *isn't | 19:05 |
macer1 | hmm...it does not have any advantages over qt i think | 19:05 |
AlanBell | nux is an architectural problem certainly | 19:05 |
davmor2 | AlanBell: No I said look to the bin, the icons are different :) | 19:06 |
macer1 | devmor2: Icons are not diffrent in oneiric now | 19:06 |
davmor2 | oh did that change | 19:07 |
AlanBell | look the same to me | 19:07 |
AlanBell | the main difference is that 3d breaks orca slightly more than 2d does | 19:08 |
macer1 | there is a "natural scrooling" on unity 3d launcher. 0_o | 19:09 |
AlanBell | and 3d breaks onboard totally | 19:09 |
AlanBell | or isn't typeable by onboard or other onscreen keyboards | 19:10 |
AlanBell | if you drag an app icon from the lens and drop it on the desktop you get an error while copying dialog | 19:12 |
AlanBell | error getting information about / | 19:12 |
AlanBell | I really hate that you can have stuff in the launcher that isn't running | 19:13 |
macer1 | the main think i have about ubuntu is that there is qt,gtk, and mono at the same time. it is not standarized, like in osx. | 19:13 |
macer1 | *i hate | 19:13 |
macer1 | and osx have multitouch everywhere. in other thinks ubuntu is better IMO | 19:16 |
macer1 | *things | 19:16 |
macer1 | so will ayatana fix ugly unity 3d? | 19:27 |
jbicha | macer1: qt, gtk, and mono run on OS X too | 19:40 |
macer1 | but its not included by default | 19:42 |
macer1 | banshee,tomboy,unity-2d,firefox... | 19:42 |
macer1 | all that are not native apps | 19:42 |
macer1 | thunderbird in oneiric | 19:42 |
macer1 | c | 19:42 |
jbicha | what do you mean "native apps"? | 19:43 |
macer1 | gtk integration | 19:44 |
macer1 | look at elementary OS | 19:44 |
macer1 | this is perfect OS with native apps, like OSX | 19:44 |
jbicha | they look native enough to me | 19:45 |
macer1 | but it is not perfect. and I like when things are perfect :D | 19:45 |
jbicha | you could use a pure GNOME distribution if you like, or just remove the non-GNOME parts of Ubuntu | 19:47 |
macer1 | i know that. | 19:48 |
jbicha | how does Banshee not fit in? it looks like GNOME to me | 19:48 |
macer1 | yes...but mono is sooo slow | 19:48 |
jbicha | macer1: are you sure? there was a bug with the Ubuntu One plugin that was slowing Banshee down but that shouldn't be an issue now | 19:49 |
macer1 | hmm | 19:51 |
macer1 | if ubuntu want to be better than osx and windows it needs to be standarized | 19:52 |
macer1 | jbicha, it is fixed but all mono is slow | 19:52 |
macer1 | elementary OS is doing it right. | 19:53 |
macer1 | they writed gtk email client, used gtk browser. | 19:53 |
macer1 | and writed gtk music player | 19:53 |
macer1 | ubuntu should do this like elementary OS... | 19:53 |
jbicha | mono is a just a programming language and it can use GTK just like Vala or C | 19:56 |
macer1 | ok | 19:57 |
macer1 | oh and libreoffice | 19:58 |
macer1 | but why not make a better music player, better browser, better mail client, better adress book? | 19:58 |
AlanBell | go for it | 19:58 |
macer1 | elementary OS know that firefox,banshee,tomboy,libreoffice does not integrate well. they deleted it or maked better version | 19:58 |
macer1 | we can use elementary programs | 19:59 |
macer1 | as default | 19:59 |
jbicha | macer1: there's no way we can use Marlin as default file browser for instance as it's not been released yet | 19:59 |
jbicha | also we have to use Nautilus to draw the desktop icons anyway so shipping 2 file browsers is a bit silly | 19:59 |
macer1 | I think marlin can do that | 20:01 |
jbicha | so, first upstream needs to release the apps, then they can get packaged and then they can be evaluated to see if they should replace the current default | 20:01 |
jbicha | it's extremely unlikely that anything other than firefox or chromium will be the default web browser | 20:02 |
jbicha | but I'll be happy to try Elementary OS out, it looks like that won't happen for at least 7 more months | 20:05 |
macer1 | sorry i needed to restart OS | 20:10 |
macer1 | so marlin in ubuntu 12.04 will be cool :D | 20:11 |
macer1 | and beatbox | 20:11 |
macer1 | and postler... | 20:11 |
jbicha | macer1: those don't have a chance at being default in a Ubuntu version if there isn't a usable release by the time of the preceding UDS | 20:56 |
AlanBell | are they in Ubuntu at all at the moment? | 21:02 |
AlanBell | or Debian even? | 21:02 |
Pendulum | \\\\/............... | 21:05 |
AlanBell | o/ scotia | 21:06 |
Pendulum | AlanBell: yeah, she just always has to get her paw into everything, even Unity ;-) | 21:08 |
hicham | anyone trying to upstream ubuntu menu proxy patch to gtk ? | 21:58 |
=== seif is now known as seiflotfy |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!