/srv/irclogs.ubuntu.com/2011/12/05/#ubuntu-desktop.txt

smspillazhi00:10
RAOFHowdie smspillaz00:20
RAOFWhat's going down in groove town?00:20
smspillazgroove town ?00:34
smspillazits perth00:34
smspillaznothing happens here00:34
smspillazRAOF: well, I am grooving out by writing tons and tons of classes and abstractions to make testing work in compiz but its like00:39
smspillazthis is the one thing that makes me hate C++00:39
RAOFYup.00:39
RAOFDynamic languages like python love unit-testing, and low level stuff like C also works fine.  It's the intermediate languages like C++ where unittesting means fighting language features.00:40
smspillazits not really fighting its more like00:41
smspillazforcing you to do things properly00:41
smspillazexcept that nobody ever did things "properly" in compiz00:41
smspillazso you pretty much end up rewriting the whole thing00:41
desrt_smspillaz: hey.  can i convince you to use the compiz bell plugin as a procrastination strategy to avoid writing tests?00:56
=== desrt_ is now known as desrt
smspillazdesrt: lol01:02
smspillazdesrt: but then I'd have to write tests for the bell plugin!01:02
smspillazdesrt: and then tests for the tests!01:03
RAOFMove Xig!  For great justice!01:03
smspillazlol01:03
desrtsmspillaz: you could invoke proverbs to avoid writing a test for the bell plugin01:03
desrtwho would question the knowledge or proverbs?01:03
smspillazdesrt: proverbs01:03
desrt*of01:03
smspillaz?01:03
desrtsmspillaz: commit message: 'ask not for whom the bell tolls'01:03
smspillazlol01:03
robert_ancellRAOF, configure.ac:36: error: must install xorg-macros 1.14 or later before running autoconf/autogen - any ideas?02:29
RAOFrobert_ancell: Got xutils-dev installed?  That's where xorg-macros lives.03:56
robert_ancellRAOF, that's what I'm looking for, thanks!03:57
robert_ancellRAOF, hey, does anyone use colormaps anymore?03:58
* desrt eyes robert_ancell with a high level of suspicion03:59
* robert_ancell looks innocnet04:00
desrti hope you're not planning on doing anything that would make it more difficult for me to run starcraft!04:00
robert_ancelland spell good04:00
robert_ancelldesrt, does starcraft use colormaps?04:00
desrtyes04:00
robert_ancellvia wine?04:00
desrtwell04:00
desrtit uses windows colourmaps04:00
desrt"256 color mode"04:00
desrti think wine may automatically convert if the X server is running in 24/32bit04:01
desrtwhich generally causes the game to run much slower04:01
RAOFRight.04:01
desrtsince it can't just memory map straight through anymore04:01
RAOFNo one uses colormaps.04:01
desrtbut has to perform a mapping on each pixel04:01
RAOFBecause no one uses 8bpp.04:01
desrtstarcraft is the most important thing you can do with your computer04:02
RAOFTechnically, the Composite extension allows you to divorce the bitdepth of the windows from the bitdepth of the framebuffer.04:02
desrtwe should really have someone dedicated to making sure it runs efficiently04:02
RAOFSo, if you wanted fast starcraft what you should do is to propose an extension to Composite, and implement it.04:02
desrthmmmmm04:02
desrton the premise that mutter/compiz would be able to do the colour mapping faster than wine?04:03
desrt(probably a pretty good premise, actually)04:03
RAOFRight.04:03
robert_ancellis starcraft not fast enough anyway?  I mean, it was designed in the early pentium era right?04:03
RAOFThere have been discussions on the xorg devel mailing list along the lines of “Yeah, this is totally possible, and I'll implement it.  Right after learning Portugese”04:03
desrtrobert_ancell: having to map every single pixel as it's written to the screen is a non-trivial undertaking04:03
desrteven if there's only 640x480 of them :p04:04
RAOFActually, its really eas.04:04
RAOFAs long as you're allowed to use a GPU.  Which seems reasonable :)04:04
desrtright.  wine uses the cpu :)04:04
desrtand i think it does it on the same core as the game is already running on04:04
RAOFIt totally should be creating 8bit palletised textures.04:04
desrtso no help there04:04
desrthah04:04
RAOFIt's a supported GL format :)04:04
RAOF(Which I think (a) no-one uses and hence (b) no-one cares that it's slow)04:05
desrti don't think it has that level of information from starcraft04:05
desrti think it just sees a giant field of pixels04:05
RAOFNo, but Compiz certainly could.04:05
desrtah04:05
desrtyou mean dynamically update a 640x480-sized 8bit texture with the framebuffer content from the game04:06
desrtya.  that would probably work rather well04:06
RAOFYeah.  You'd advertise an 8bit window format to Wine, Wine would use it, and Compiz would do the upload as an 8bit palletised texture.04:06
desrti can imagine that it could be the least-optimised codepath in the GPU/driver and still be capable of delivering full motion at 640x480 on intel hardware04:07
desrtt wine: plz haz 8bit pixmap04:07
desrt - love, compiz04:08
RAOFFailing anything else, colourspace conversion is the poster-child for pixel shaders.04:08
RAOFWhich everything non-ancient has.04:08
desrtRAOF: so.. how much free time do you have this cycle to dedicate to making starcraft awesomely fast while using 1% of the CPU?04:08
RAOFI've already got an X extension on my plate, and it's not Composite :P04:09
desrtbtw04:09
desrtdid you see ajax finally pushed the xfixes extension for selective property notifications?04:09
desrti don't care what version of X you're shipping this cycle, you need to pick that patch04:10
RAOFI did see that, yes.04:12
desrtactually, now that i read the email, i don't care for it that much04:17
desrtit falls short in a couple of areas04:17
desrtbut it's still a million times better than the current situation04:17
desrtand it fits with the nature of X better than my own suggestion did04:18
robert_ancellRAOF, ooh, which extension is that?04:18
RAOFSomething for DX.04:19
desrtRAOF: how about my grab-detecting extension? :)04:19
desrtthat would actually be useful...04:20
desrt(well... the starcraft thing would be useful too... but i can understand if it's a bit of a hard sell)04:20
desrt((and honestly, would probably most easily be implemented as a nested X server which rendered itself using GL complete with scaling and the input mapping required to deal with that))04:21
robert_ancelllearning more about X really makes the old stuff in GTK+ make a lot more sense04:27
desrtbut also makes you want to kill it with fire all the more...04:27
desrtcolourmaps and visuals leaking out of the toolkit = pretty broken abstraction04:28
robert_ancellRAOF, No package 'xf86driproto' found ?04:35
RAOFx11proto-dri-dev?04:36
RAOFAh, sorry. x11proto-xf86dri-dev04:36
* RAOF is apparently today's apt-file replacement ☺04:36
robert_ancellso much friendlier than apt-file :)04:37
TheMusolol04:47
=== pitti_ is now known as pitti
pittiGood morning05:35
didrocksgood morning07:09
pittihey didrocks07:09
didrocksguten morgen pitti07:10
=== geser_ is now known as geser
SweetsharkMoin Desktopers!07:59
pittihey Sweetshark, how are you?07:59
pittididrocks: enjoyed your first "free" weekend? :-)07:59
didrockspitti: well, not as much as I wanted. Unfortunatly since last Thursday, my neck is totally blocked (as the top of my body), so I can't really moved and the pain is quite severe :/ So, the week-end has been video games mostly, trying to move as few as possible08:01
didrocksbut video games week-end can be nice sometimes…08:01
pittididrocks: erk, how did that come?08:01
pittididrocks: the video gaming sure, but certainly not the body pain?08:01
didrockspitti: I'm not sure at all how it came, made a vasovagal episod Wednesday evening on the way back from the rock danse. Julie told me that I was quite straight during this, so maybe I hurted some muscles back then08:03
didrockspitti: and you? how was your week-end? Went to Dresden?08:07
pittididrocks: no, to Halle; we had a family thing there (my in-law's parents)08:07
pittibut we went back on Saturday evening08:08
didrocksah ok, was a short visit then :)08:08
pittiso we spent yesterday with some gardening, and then when the storm/rain started, we just idled, did some reading, and I played Diablo II again :)08:08
didrocksheh, nice :)08:10
=== highvolt1ge is now known as highvoltage
smspillazdidrocks: weird, you have the same thing that I had last week08:17
didrockssmspillaz: oh, really?08:18
smspillazyeah, had a really stiff spine last week08:18
smspillazpain to do anything08:18
* smspillaz wonders if there was something going around in orlando08:19
Sweetsharkpitti: Fine, thanks. I had some guests and we partied a bit (officially because of my election to deputy director, inoffically because it is fun). And I am getting to old for this, leaving the flat at midnight to continue on the Kiez ...08:24
pittiSweetshark: heh, I symphatize; back then I usually got home when it was light outside again, but these days I just fall asleep around 2 on a Saturday night :(08:25
Sweetsharkwell, on the reeperbahn there are some clubs that open when everyone else closes to keep people partying until the others open again, so they can continue there ...08:27
Sweetshark... great for drugged up technocrowd that run around like duracellbunnies.08:28
pittihehe08:28
pittibut when it gets late, I also either need to go home or go dancing; just sitting and drinking doesn't work any more08:28
didrocksrickspencer3: hey, I think that you wanted to assign the dx team for the compiz bug, isn't it? bug #89782908:44
ubot2Launchpad bug 897829 in compiz "Assigning shortcut to "Move window to workspace X" doesn't ever take effect" [Low,New] https://launchpad.net/bugs/89782908:44
rickspencer3didrocks, well, I thought the desktop team might want to hide the option since it doesn't work08:44
rickspencer3didrocks, but just reassign as desired08:45
didrocksrickspencer3: sorry, read ccsm :)08:45
didrocksnot g-c-c08:45
didrocksyeah, you're right, or maybe I can bind the key to a compiz one, let me check08:45
rickspencer3didrocks, you don't *have* to respond08:46
=== Guest37602 is now known as jpds
chrisccoulsongood morning desktop team09:18
didrockshey chrisccoulson, how are you?09:25
pittihey chrisccoulson09:25
seb128hey09:37
pittibonjour seb128, ca va?09:37
didrockssalut seb12809:37
seb128hey pitti, I'm good thanks, how are you?09:37
seb128lut didrocks09:37
seb128hum, my connection is laggy today it seems :-( took me a while to connect to IRC and reading imap on the canonical imap is slow09:38
chrisccoulsonhi didrocks, pitti, seb12810:05
seb128hey chrisccoulson, how are you?10:05
chrisccoulsonseb128, yeah, not too bad thanks. how are you?10:05
seb128I'm good thanks10:06
pittiSweetshark: could you please upload a new openoffice.org  which doesn't build openoffice.org-base on armel?10:06
pittiSweetshark: it only exists on amd64, i386, powerpc10:06
pittiSweetshark: or alternatively, is the absence of libreoffice-base on armel a bug or can't it work on armel for some reason?10:07
pittiSweetshark: I'll just fix oo.o10:55
rodrigo_hello10:57
seb128hey rodrigo_, how are you?10:59
rodrigo_hi seb12811:02
pittiSweetshark: does http://paste.ubuntu.com/760293/ look ok to you?11:08
Sweetsharkpitti: yes, looking good11:10
pittiSweetshark: ok, hanks11:10
pittithanks11:10
=== MacSlow is now known as MacSlow|lunch
seb128hehe12:34
seb128"<Np237> mbiebl, is it really wise to switch to totem 3.2 in this situation?12:34
seb128<jordi> Np237: looks bad, here at least.12:34
seb128<Np237> Do we need to package totem 3.0 as totem-fallback too? o_O"12:34
seb128\o/ for not jumping on the clutter based version ;-)12:34
pittiseb128: :)12:36
rodrigo_lunch, bbl13:12
=== MacSlow|lunch is now known as MacSlow
seb128pitti, there is https://code.launchpad.net/~mitya57/nautilus/ubuntu-fix-lp-795708/+merge/84370 on version if you feel like doing desktop sponsoring ;-)14:19
pittiseb128: oh, please :)14:19
seb128that's likely a good candidate for a sru as well14:19
pittididrocks: ok for me to do https://code.launchpad.net/~om26er/ubuntu/oneiric/compiz-plugins-main/fix-sru-regressions-2/+merge/83590, or do you want to handle compiz-y stuff yourself?14:20
pittiseb128: so these don't appear on http://reports.qa.ubuntu.com/reports/sponsoring/ :(14:21
seb128pitti, no, because there are against a team vcs14:21
didrockspitti: just check that there are tests for both in trunk at least as we will need to copy c-p-m to precise and tests for compiz is part of the requirements14:21
seb128didrocks, pitti: can we please not block those sru regression fixes on precise work?14:22
didrocksif we don't have to push the SRU to precise, it's fine14:22
didrocksotherwise, we really can't push it14:22
seb128it's already embarassing enough that we broken those and didn't fix them in over a month of time14:22
didrocksagreed, but I asked confirmation about "not pushing anything not tested in precise", even if it's to fix a SRU, so if pitti agrees that it's only pushed to oneiric and not precise, that would be possible14:23
* didrocks is between a rock and a hard place for this :/14:23
pittididrocks: ok for me, but really unusual14:23
didrockspitti: thanks Martin :)14:24
=== m_conley_away is now known as m_conley
xclaesseseb128, empathy package does not seems to install mcp-account-manager-goa.so14:28
xclaesseit should go to /usr/lib/mission-control-plugins.0/14:29
xclaessein precise14:29
xclaessealso I guess it needs to do some magic to make apparmor let mission-control load that plugin14:30
xclaessebecause just installing it manually doesn't work14:30
jdstrandxclaesse: if this is in the default install, please file a bug and assign it to me14:30
xclaessejdstrand, copied package into a oneiric ppa14:31
seb128kenvandine, ^14:31
seb128xclaesse, you better ping kenvandine about empathy packaging issues, he's the one doing most of the packaging work on it14:31
xclaesseseb128, ok thx :)14:31
kenvandinei can fix that14:31
xclaessekenvandine, cool, thanks :)14:32
kenvandinenp14:32
xclaessekenvandine, FYI that plugin makes MC import accounts setup in gnome-online-account14:32
xclaesseso if you have a gtalk account in goa, it gets into empathy automatically14:32
kenvandinecool14:32
kenvandinexclaesse, what about accounts already in empathy?14:32
kenvandinedoes it detect the dupe?14:32
xclaesseno14:33
xclaesseyou'll just get both14:33
kenvandinebummer14:33
kenvandineok14:33
xclaessekenvandine, goa also support facebook accounts (if build with an app secret key) and soon MSN (via XMPP!) as soon as davidz review my patch :)14:34
kenvandineok14:35
pittididrocks: I assigned the precise tasks of these two to you, please assign to teh correct person in DX; thanks!14:35
didrockspitti: excellent, thanks!14:35
seb128om26er, pitti: the compiz sru is screwed :-( how come it doesn't include the fix for bug #876591 which is in the packaging vcs for some weeks?14:50
ubot2Launchpad bug 876591 in compiz-plugins-main "Ctrl-Alt-KP_9 adjusts window to upper half instead of top right corner (behaves like Ctrl-Alt-KP_8)" [High,Confirmed] https://launchpad.net/bugs/87659114:50
pittiseb128: oh, we have a bzr for oneiric?14:51
om26erseb128, this one https://code.launchpad.net/~om26er/ubuntu/oneiric/compiz-plugins-main/fix-regressions14:51
om26erthis branch also have the fix for 87659114:51
seb128pitti, no, we usually don't create a new serie if precise didn't get a newer version14:51
pittiseb128: so you mean s/screwed/incomplete/, or is it actually broken as-is?14:51
seb128om26er, how come you didn't propose that one for sponsoring?14:52
pittiseb128: we can reject this, and upload that other branch instead, fine for me14:52
seb128pitti, no, it just doesn't contain one of the fixes, so incompletes14:52
pittisorry, I didn't check trunk bzr, just the MP14:52
om26erit was creating a very big diff like a few thousand lines14:52
seb128om26er, you were proposing it against the wrong serie I guess?14:52
pittipresumably http://bazaar.launchpad.net/~compiz/compcomm-plugins-main/ubuntu is a custom branch14:53
pittiwhile om26er's MP was against the UDD one14:53
seb128pitti, don't bother, we will get the other bug fixed in the next one I guess14:53
pittiyou can't merge them to each other14:53
om26eri cant propose my branch to https://code.launchpad.net/~compiz/compcomm-plugins-main/ubuntu there is no option14:53
seb128I will just unsubscribe from the bug meanwhile14:53
pittiom26er: no, as I said you can't merge UDD branches against debian/ only ones, and vice versa14:54
pittithey are two entirely different things14:54
om26erpitti, alright so where should i propose https://code.launchpad.net/~om26er/ubuntu/oneiric/compiz-plugins-main/fix-regressions now?14:55
pittiom26er: you can't propose it anywhere else14:56
pittiom26er: if you branch from UDD, you have to merge to UDD14:56
pittiom26er: i. e. the "correct" way would have been to branch from ~compiz/compcomm-plugins-main/ubuntu, do additional changes there, and propose to merge against that again14:56
pittiom26er: but as seb128 said we can get to that patch in the next round14:56
seb128pitti, waouh, down to 23, good job (to you and others, 37 was already low, it was around 70 until recently)14:57
seb128om26er, or do another merge request just for this one14:58
seb128it might be good to get in because ctrl-alt-kp9 seems to be quite visibly broken and we have users who wonder why the 1 liner is still not landing after a month ;-)14:58
om26erseb128, sure, I can do that now14:59
seb128om26er, thanks14:59
om26erpitti, thanks for clearing things :-)14:59
om26erseb128, :-)14:59
jincreatorpitti: Hi, are you busy now?15:07
pittijincreator: I'm always busy, it's a matter of what I'm busy with :)15:07
jincreatorpitti: I saw you remove 69-language-selector-ko-kr.conf.15:08
pittijincreator: right, you said it was obsolete as we are using nanum now?15:09
jincreatorpitti: Yes, but only for when Un fonts is default at Korean font. Sorry for make confusing.15:09
pittijincreator: I thought the hack was for unfonts?15:10
jincreatorpitti: Yes, it is. But we are trying to change default font from unfonts to fonts-nanum.15:10
pittijincreator: right; so if the hack only applies to unfonts, why do we need it if we are using nanum?15:11
jincreatorpitti: Because fontconfig-config(and upstream at freedesktop) doesn't have setting for fonts-nanum. I report bug to fontconfig for add it, but it looks lots of time to applied.15:12
jincreatorSo fonts-nanum maintainer(in Debian) add fontconfig snippet to overwrite settings and make fonts-nanum as default font, but it seems not working at latest Ubuntu.15:14
pittijincreator: right, that's the <lang> issue we talked about15:14
jincreatorpitti: Yes, so if we patch fontconfig-config to add information about fonts-nanum, we can remove 69-language-selector-ko-kr.conf. But I think just modify 69-la... is easier and simple workaround. What do you think about it?15:15
pittijincreator: I think fixing the <lang> bug is better, as it will also unbreak other ttf-* packages15:16
pittijincreator: and NB that language-selector is going away, so we'll need to drop the hacks anyway15:17
pittior at least minimizing them so that we can port them to whatever replacement15:17
pittibut <lang> seems like a much better option than these static hacks, as they will be prone to being wrong15:17
pittiyou might have more than user on a machine, etc. (think universities, etc.)15:17
jincreatorpitti: You mean language-selector is no more exists at Precise?15:22
pittijincreator: still there, but we have tried to get rid of it for a few cycles now, and rodrigo is almost there now15:23
pittiwith adding enough functionality to the gnome-control-center module to replace it15:23
pittil-s is quite horrible and unmaintainable code15:23
jincreatorpitti: Ok, I see. I'll try to find another method to make font-nanum as default Korean fonts without using l-s.15:28
pittijincreator: I thought making <lang> work would do the trick?15:30
pittiwhatever breaks it, we should just fix that IMHO15:30
pittigood night everyone!15:33
jincreatorpitti: good night!15:33
didrockshave a good night pitti :)15:41
looldidrocks: Hey, zeitgeist-core used to ship zeitgeist-daemon.pc which unity-lens-files checks for in configure, but it's nowhere in the archive in precise15:52
looldidrocks: should we drop the ZEITGEIST_DAEMON test from configure.ac?  it seems unused15:52
didrockslool: that's more than possible that it's not shipped anymore with the new vala zg15:53
didrocksmhr3: I think you see both sides of this piece there ^15:53
mhr3didrocks, i'm not sure why files lens checked that15:55
didrockslool: let's wait on kamstrup then15:55
looldidrocks: Sure; give me a heads up if you want me to upload a packaging fix, but lazy as I am, fixing it upstream would be even nicer  :-)15:56
didrockslool: yeah, I think it's better to directly fix that upstream and less than a burden :)15:57
mhr3lool, didrocks, ok, it will be fixed upstream15:58
didrocksthanks :)15:58
loolmhr3: Thanks15:58
mhr3files lens doesn't need the check though15:58
didrocksmhr3: I can even propose a merge req if you want :)15:58
didrocksmhr3: will be the perfect time to boostrap the file lens automerge15:59
mhr3didrocks, merge req for files lens?15:59
didrocksyep15:59
mhr3didrocks, sure, why not15:59
didrocksmhr3: ok, will do that just after the call15:59
mhr3it'll fail though :P15:59
didrocksbecause of new libunity?15:59
mhr3yes :)15:59
didrocksok, I'll let you do the libunity stuff first :)16:00
=== m_conley is now known as m_conley_away

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!