[00:00] and there are a lot of discussions, I usually find it very hard to pick which ones to go to as there are many at the same time [00:01] I've never been to a summit before, I really need to get out and meet people more. [00:02] And I would be going through Denmark, Germany, France having fun and sampling the culture and cuisine once again [00:02] yeah, if you can afford to go it's a great experience [00:31] can we upload to karmic? [00:32] mdke: I don't think so yet [00:32] will things be held in a queue or just rejected? [00:32] (I see a few things on karmic-changes but maybe those are exceptional [00:34] mdke: the toolchain is still being finalized [00:34] mdke: I suppose you could give it a shot, but generally we do wait until the toolchain is up and working [00:36] ok, fair enough [00:36] I'll keep my releases in bzr only for a bit :) [01:53] mdke: feel free to upload; they'll be held in a queue [02:14] infinity: debootstrap 1.0.13 is on its way in now, with karmic support [02:19] karmic support? [02:20] yes [02:20] What does that mean... [02:20] the next Ubuntu release will be called "Karmic Koala" [02:20] its codename, that is [02:21] I know, but for debootstrap to support karmic, what does that mean. [02:21] I don't find anything through google about that... [02:21] debootstrap needs to have explicit support for each release it might be asked to install [02:22] in this case, it's pretty trivial, just a symlink [02:22] ohh, nice [02:22] I'm not sure why google would be helpful here [02:23] we have to do a bunch of stuff like this when opening each new release; it's pretty routine [02:23] I've never been all that involved in the development of a release [02:23] I managed to scratch the surface but I have no idea where to go to get really involved [02:23] See what is happening and all [02:24] http://wiki.ubuntu.com/NewReleaseCycleProcess is what we have to do to get each new release ready for developers to contribute [02:24] I subscribe to the announcements mailing-list(though an announcement is only made every month or so unless they want to say, please do 5 a day) [02:28] LordMetroid: http://launchpadlibrarian.net/25947235/debootstrap_1.0.12_1.0.13.diff.gz has some other changes as well, but the relevant change for this is right at the bottom [02:28] * cjwatson -> bed [02:29] nigh cjwatson [02:29] *night [04:19] tjaalton: are you familiar enough with the vmmouse driver to play "crack or not crack" with me? :) [04:19] so the problem is, it seems like modern vmware+vmmouse, when you click down the mouse it sends ButtonDown, MotionEvent(0px or 1px), ButtonUp [04:20] this combination of events causes a drag to be interpreted. [04:20] often times this means right-clicking for a context menu automatically selects the first thing [04:21] (it interprets it as you right-clicked, dragged to the first item, then let go) [04:21] now... how to correctly work around this? [04:21] I tried a simple one-liner that disallows posting both a move and a button change event at the same time, and it's pretty effective but not perfect [04:22] still even a miniscule move of the mouse while clicking gets caught as a drag. [04:22] would implementing a couple millisecs of time awareness into a buttondown in which motion events are discarded be insane? [04:43] jdong, have you tried 12.6.3? [04:43] bryce: I stole a tarball from Debian Sid whatever that version is [04:43] yeah 12.6.3 [04:43] same bug persists [04:44] holding the mouse as still I I can, clicking always generates a down, move, up event. [04:45] jdong, hrm [04:45] jdong, have you talked with upstream about this? I've been exchanging emails with philip langdale the past couple days on vmmouse [04:45] it might be my host (VMware fusion 2.0.2) [04:45] but in any event IMO the guest mouse driver shares some responsibility in filtering out nonsense and making it behave more like a mouse. [04:46] and no, I haven't had contact with upstream [04:46] fwiw I'm considering merging up a backport of vmmouse for jaunty... [04:46] if you're in regular contact with them and want to poke this issue to them too, it'd be well appreciated [04:46] however if it's too buggy [04:46] it'll require a bit of packaging changes (conflicting mdetect), other than that so far it looks good here [04:46] no, I don't really have the time to do a lot of troubleshooting (I never use vmware myself) [04:47] but if you work with them and get a patch, I'd be open to including it in the backport [04:47] ok [04:47] I can put you in contact with philip if you'd like to bypass the usual process [04:48] er, "directly" [04:48] that'd be cool [04:48] jdong@ubuntu.com ? [04:48] yep [04:51] jdong: ok, start a bug and give me the bug # and I'll include it in the email [04:59] bryce: bug 366521 [04:59] Launchpad bug 366521 in xserver-xorg-input-vmmouse "VMWare fusion host + vmmouse driver generates drag events on any click" [Undecided,New] https://launchpad.net/bugs/366521 [05:00] awesome, thanks [05:00] sure thing [05:00] yeah interestingly in this xev capture, it raised a completely null MotionNotify event... [05:00] i.e. motion to the exact same coordinate as the buttonpress one. [05:00] that's clearly wacky behavior ;-) [05:01] email sent [05:01] thanks [05:01] hopefully it'll be something simple :-) [05:02] for some reason I have a sinking fear it is due to something busted in xserver [05:03] yeah I'm not sure exactly what the physical mice drivers do [05:03] but it's always nailed the differecne between a click vs click-drag for me [05:03] but vmmouse is kinda ridiculous [05:03] right now EVERY click is a click-drag [05:03] and with my one-liner hack if you even BREATHE on the mouse it turns into a drag. [05:03] I've seen a bit of mailing list banter on QT's list about this [05:04] it in fact is bad enough that their click-to-sort table widgets don't work with vmmouse [05:04] you know, the symptoms you've described I had seen myself, although only with firefox, and it's long since gone away so I'm sure it' sunrelated... [05:04] heya andresmujica [05:04] but QT folks adopted a very novel approach by claiming MotionNotify is for "logical motion" [05:04] (aka it's not our problem fix it elsewhere) [05:04] in any case, if nothing else I sympathize with how irritating it is [05:04] haha thanks :) [05:05] it's crazy I'm the only one who's complained about it so far [05:05] not many people run Ubuntu in VMWare it seems [05:05] I think a lot of people moved from vmware to kvm [05:05] (at least not a GUI) [05:05] I used to get lots of bugs about ubuntu+vmware, pre-kvm [05:06] yeah, if I'm on a Linux host I'll jump to KVM first [05:06] but on this hardware OS X is the best native platform (the one that gives me least grief setting up) [05:06] so it'd be nice if I can run a couple Ubuntu instances [05:09] jdong: i use vmware with ubuntu [05:09] heya calc [05:10] calc: do you get the crazy mouse symptoms I see here? [05:10] jdong: i haven't noticed any problems with it yet with 6.5.2 other than the keyboard issue (it might be fixed for 6.5.2 i had to fix it in 6.5.0) [05:10] calc: a quick and dirty test: gnome-terminal, do a dmesg [05:10] hmm let me see [05:10] calc: then do a single left click on any leftover blankness of a line of output [05:10] see if the entire blank area gets selected. [05:10] or better yet, xev, click the mouse and see how many events fire [05:11] ok [05:11] should just be down then up [05:11] I get down, motion, up. [05:11] only one press/release when i hit a button [05:11] hmm let me try a few more times [05:11] calc: interesting, perhaps it's VMWare Fusion specifically that is implicated. [05:12] I did mention in my bug report it was on Fusion [05:12] i get notify events if i click inside the black square but only press/release for the other area [05:12] i am not completely up to date 9.04 in my vm though [05:13] hmm [05:13] * calc does an update and sees what it does [05:13] also i'm running amd64 [05:13] yeah this is 64-bit Ubuntu in my VM too [05:13] it's probably that your parent X server "numbs" motion events while clicking for you. [05:14] I'm betting OS X does use the GUI toolkit abstraction layer to weed out nonsense drags [05:14] I can imagine drag filtering being incredibly annoying to a graphics designer or a leet gamer. [05:14] ah maybe so, i'm running 9.04 amd64 -> vmware -> 9.04 amd64 [05:14] right, I suspect the parent X server's mouse driver is playing a role [05:15] well when i click i'm not making any motion at all [05:15] nor am I attempting to do so [05:15] i'm using a laptop so its simply a keypress and that is all [05:15] in my testing above the mouse is off the table (optical) [05:15] ah ok [05:15] and the Motion event is for the exact coordinate of the click :) [05:15] which makes it even more idiotic [05:16] heh [05:17] hmm apparently the servers are so loaded its causing even approx to hang for me, ugh [05:19] oh it's not a fun day at all in archive land :) [05:21] hi bryce! [05:21] hi all [05:22] andresmujica: o/ [05:33] cjwatson: Yay, thanks. [05:34] cjwatson: I'll update the livefs stuff and triple-check that it works sometime early in the weekend, so I can sleep soundly for the rest of the weekend knowing that my "new release" work is more or less done. :) === ScottK2 is now known as ScottK-desktop === MooKow is now known as LordKow === Sikon is now known as LucidFox [09:19] whom/where would i file ubuntu wiki bugs at? https://wiki.ubuntu.com/DebootstrapChroot does not work for jaunty. i believe it has to do with the fact that recommends are now installed by default. if you follow the how to exactly as it describes then you are left with a broken chroot env [09:21] cjwatson: thanks for the info === azeem_ is now known as azeem [10:01] LordKow: Well, it's a wiki - so you can do so yourself, but you can look at who did the previous edits for advice. [10:03] jpds: ah i see. i thought maybe the wiki was a little more controlled than that. im not just going to edit it... maybe there is some sort of commenting feature i can utilize. [10:04] LordKow, how is it broken, what problem are you getting ? [10:06] LordKow: Yes, there's a 'Comment for this edit:' field. [10:06] joaopinto: 'apt-get install wget debconf devscripts gnupg nano #For package-building' when installing devscripts it wants to install dbus which will most definitely not work in a chroot. apt errors out in the dbus postinst script when it tries to start the service. and then it errors out again when trying to remove it (prerm script) [10:08] im not really sure why devscripts needs dbus in the first place. im guessing there is a dep that maybe (incorrectly?) depends on dbus... i can't see how any devscript or any of it's deps would need dbus. that is a pretty rich package though so i could be missing something. [10:10] LordKow, you will need apt-get install --no-install-recommends ... [10:11] LordKow, it does not depend on dbus, it's a "recommends" relation [10:13] joaopinto: ah okay. so it was between intrepid and jaunty release that ubuntu went with the install recommends by default (following debian guidelines)? [10:13] I believe it was during intrepid [10:16] joaopinto: perfect, --no-install-recommends took care of it. now it should be passed along to the wiki page which im working on figuring out right now [10:24] joaopinto, jpds: thanks for the help === korn_ is now known as c_korn === tkamppeter_ is now known as tkamppeter === ogra__ is now known as ogra_ === asac_ is now known as asac [15:17] Where can I find some short examples on how to work with libapt-pkg? === Omegamoon is now known as Omegamoon|away [16:29] Quick one... Is there a different Kernel Between ubuntu and ubuntu NBR in 9.04? [16:30] no [16:30] thanks!!! i have been looking everywhere for that [16:31] something aint right with the ACPI and my eeepc 1000 isnt chrging [16:32] Ill submit a bug report and roll it back to 8.10 with the array kernel [17:48] I have to ask - why is Ubuntu 9.04 released with a dysfunctional intel xorg driver? [17:49] larsivi: basically because it was the best we got [17:50] LaserJock: I understand the complications of package and kernel dependencies, but there has been a constant regression since 8.04 meaning that maybe keeping with some older versions would have been a better compromise [17:51] larsivi: except there's problems matching Xorg and driver versions I believe [17:51] larsivi: I don't know that the old driver on newer Xorg worked any better [17:52] LaserJock: so then one could have opted for an older Xorg too [17:52] Am I wrong in thinking that most users have intel cards? [17:52] larsivi: at least as far as 2.4 vs 2.6 goes they both suck in different ways according to the phoronix performance tests anyway [17:52] larsivi: which would have caused problems for everybody else [17:52] larsivi: yes, you are [17:52] larsivi: I think anyway there are more nvidia/ATI out there presently, I could be wrong though [17:52] LaserJock: "most" is true but not an overwhelming majority [17:53] intel has by far the majority marketwide... not sure about ubuntu users specifically [17:53] calc: I wasn't suggesting a 2.4 kernel :P [17:53] calc: more intel than everything else combined? [17:53] larsivi: i'm talking about intel driver version number [17:53] calc: hehe, ok :) [17:53] LaserJock: not to that extent but iirc its over 40% intel (maybe 50%) [17:54] * calc looks at the data again [17:54] calc: yeah, I could believe that [17:54] my point was that holding Xorg back is not an easy decision [17:54] http://jonpeddie.com/press-releases/details/computer_graphics_chip_shipments_dive_in_q4_08_according_to_jon_peddie_rese/ [17:54] especially if you think the -intel drivers are going to be better than they were [17:54] intel has 47.8% according to that site for shipments for 2008 [17:55] calc: well, on new machines sure [17:55] 47.8% for Q42008, was 49.4% 2008Q3 [17:55] calc: I was just thinking of all our users out there, I'm guessing quite a few still have nvidia/ATI [17:55] there are quite a few machines still being used form before Intel was even really in the market [17:56] *from [17:56] LaserJock: yea for Ubuntu users specifically the makeup may be different but Intel has been at ~ 50% for many years [17:56] anyway, this appears to have been known for quite some time (more than 6 months), so I find it a bit curious that those that can (Canonical?) haven't spent more resources on it [17:56] nvidia seems to actually be losing marketshare as time progresses [17:57] larsivi: I think pretty much all who could, did what they could [17:57] and ! (AMD/Intel/Nvidia) percentage-wise is just noise [17:57] larsivi: we had several testings of different options, it just seemed like every change would help some but not others [17:58] according to the numbers for laptops (where most Intel IGP is used anyway) the numbers are actually ~ 60% for Intel [17:58] larsivi: http://keithp.com/blogs/Sharpening_the_Intel_Driver_Focus/ might be a good read [17:59] LaserJock: in 8.10 things were slow for me (couldn't use desktop effects), but now it is close to unusable (although desktop effects are fast enough now) due to graphical glitches and absurd cpu usage [17:59] LaserJock: I follow the xorg mailing list [18:00] larsivi: right, so helping the Ubuntu X team and upstream to figure out why you're having problems is probably worthwhile [18:00] larsivi: for me 8.10 was just fine, 9.04 I get GPU lockups when using desktop effects [18:00] such is life sometimes :/ [18:01] LaserJock: i think part of the reason we see more non intel users on Ubuntu is due to them being power-users and thus willingly to pay more for the nvidia binary only drivers :-\ [18:01] calc: surely there are a lot of people using older machines though too [18:02] calc: I only have one machine with an Intel card because most of my machines were built before Intel was commonplace [18:02] LaserJock: yeah, I have put forth my information in the relevant places [18:03] perhaps that is true for very old machines < 8 years old [18:03] calc: ?! [18:03] intel has dominated the graphics market since probably shortly after i810 [18:03] er > 8 years old (over i used the wrong symbol) [18:03] calc: I only really started to see them ~ 2 years ago [18:04] LaserJock: a lot of marketshare is in business desktops which has used integrated intel graphics nearly forever [18:05] calc: ah, that'd make sense [18:06] in 2005 it appears intel had somewhere ~ 35% market share with ati taking second place at 24% [18:06] larsivi: anyway, yeah, -intel suckage is kind of a known thing, but it wasn't for lack of trying to get something better [18:06] roughly the same in 2006 [18:07] calc: good grief, really? I'd never heard of an intel graphics card until probably 2007 [18:08] or wait, 2006 maybe [18:08] darn, had it been *that* long since Dapper?! :/ [18:08] 2004 mobile market intel had 40% (the 2005/2006 numbers were for desktop above) [18:09] I have the 965 now, before that the 915 which I bought in a laptop early 2005 [18:09] hey LaserJock [18:09] the 915 came a long time after the 810 stuff [18:10] LaserJock: so assume that 40% of ubuntu users have intel graphics cards, and maybe half of those experience issues, that would mean that 20% of jackalope users would have issues with it that are obvious and easy to notice - that is a blocker in any other software [18:10] it looks like 2003 overall share intel had ~ 32% [18:10] with ati at 25% [18:12] larsivi: yeah, but there wasn't much that could be done unfortunately [18:12] larsivi: hopefully 9.10 will be *much* better on that front [18:13] LaserJock: no offense meant, but that is bullshit - that is a fix to delay a release for [18:13] overall we need to do what our users require and using hwdb to track that would be useful :) [18:13] larsivi: for how long? [18:14] larsivi: the problem isn't even really known as far as I know (I'm not an X expert by any means) [18:14] * calc thinks the current ubuntu user breakdown would have lots more ati/nvidia than the overall marketshare numbers would otherwise indicate [18:14] LaserJock: I would have lost my job 10 times if I had released software that didn't work for 20% of the users [18:14] larsivi: I don't think there's any gaurantee that with a delay of even a month or two that things were going to be much better [18:14] larsivi: it's not our software, it's Intel's and Xorg's [18:15] but ubuntu pick it [18:15] maybe there'll be updates? [18:15] larsivi: we could have decided to not ship an -intel driver, yes [18:15] now I'm a private user, so I may not switch, but as a commercial user I would have dropped Ubuntu over this issue [18:16] sistpoty: hopefully [18:17] it'd really be nice to get things better for sure [18:17] larsivi: I assume as a commercial user you wouldn't run jaunty right after its release, but maybe go with hardy instead? [18:18] sistpoty: that is a point, but for the desktop I don't find the LTS that important [18:18] *shrug*, was just a guess [18:20] when that is said, as a proffesional user I actually find it a bit problematic that python haven't been upgraded to 2.6 [18:20] larsivi: you mean in 8.04? [18:21] LaserJock: yes - sorry for confusing matters a bit [18:23] yeah, adding a whole new python version isn't really -updates territory [18:26] anyway - will leave you with my bile for now :P [18:27] When is karmic development going to start? === RainCT_ is now known as RainCT [18:42] LaserJock: i think larsivi didn't get the fact that he would have to drop all Linux shipped ~ 2009 to avoid this intel problem [18:42] LaserJock: already has [18:42] LaserJock: the toolchain is being worked on now [18:51] calc: right, yeah, I guess the perception is that we just picked the wrong version and that's it [18:52] * calc read the keithp post, sounds good going forward [18:52] hopefully no one decides to rewrite xorg again next month :\ [18:53] LaserJock: fortunately places like phoronix have made it clear that there wasn't really a choice [18:59] Do the UDS cost anything to attend? [18:59] <[reed]> no, they are free... of course, travel is yours to pay unless you get sponsored [19:00] <[reed]> LordMetroid: https://wiki.ubuntu.com/UDSKarmic [19:00] thanks [19:06] jcastro: Thanks for the link to the Intel video blog post. [19:07] I'd feel better if he'd mentioned UXA is pretty unstable for a lot of people and they were fixing that. [19:08] hmm. UXA is the thing that made compiz eat my memory I think :-) [19:17] calc> Ah, I see... https://edge.launchpad.net/ubuntu/karmic [19:17] What I meant was, when is it going to be unfrozen? [19:17] https://wiki.ubuntu.com/KarmicReleaseSchedule has April 30 as "toolchain uploaded" [19:19] LucidFox: the toolchain is being worked on right now [19:20] Yes, I realize this :) [19:23] LucidFox: When it's ready. [19:23] * LucidFox nods [19:23] LucidFox: You can upload stuff now, it'll just get held until the archive opens. [19:55] This is ridiculus [19:56] Ubuntu has only gotten slower and slower, now gedit is even unable to render the characters at the same moment I type them. [19:57] This problem needs to be addressed or is Ubuntu only intended for high end computers... [19:58] hello, what changed since kernel 2.6.28, because i cannot establish a ppp <-> nas interface and connect to internet anymore === Omegamoon|away is now known as Omegamoon [20:00] They replaced the kernel with Ms code [20:05] LordMetroid: do you have Compiz enabled? if so, disable it. [20:05] Nope [20:05] LordMetroid: what does (h)(a)top say? [20:07] (h)(a)? [20:07] htop atop [20:07] choose yer poison [20:08] brb, installing one of them [20:10] LordMetroid: before you join a developer channel and rant about the lack of quality and performance problems, using basic tools to actually check your own machine is a best practice. [20:11] I have run top [20:11] and? [20:12] what process is utilizing the most CPU cycles? [20:12] X.org [20:12] and what video chipset and grpahic driver are you using? [20:13] An ATI Radeon Mobility 7600 [20:13] and whose driver? [20:13] X is occupying like 40% [20:13] The default from installation [20:16] alt-F2 [20:16] metacity --replace [20:16] restart X === Guest87757 is now known as NCommander === NCommander is now known as Guest85976 [20:24] * mneptok whistles the "Jeopardy!" theme [20:33] cr3! [20:41] mneptok: hey dude! [20:43] cr3: our final weekend in .qc, we leave on Tuesday. got back from the MySQL conf @ 3am last night. :( [20:43] mneptok: sounds painful, where was the conf? [21:02] cr3: Santa Clara [21:02] (Silicon Valley) [21:21] ipoo [21:21] suck my dick [21:21] -.- [21:50] I figured it out, subpixel rendering is the demon in this... [21:51] no, your video driver is [21:51] subpixel smoothing works just fine on my Intel chipset [22:00] Turning off subpixel rendering on the text solved it though [22:30] LordMetroid: VBGR? [22:31] VBGR? [22:31] RGB [22:31] Ok then, something is wrong with your driver [22:32] Although even a CPU from the time that GPU came out should be able to do subpixel rendering without using 40% CPU [22:32] hm [22:32] Unless you have that GPU but a 500Mhz CPU or something :P [22:32] I got a 2.4GHz P5 [22:32] *p4 [22:33] mobile variant(if there is such) [22:33] You should be able to do full software rendering without using 40% CPU then as long as you aren't trying to use compiz :P [22:34] LordMetroid: Try with the vesa driver [22:34] how do I change? [22:34] LordMetroid: edit your xorg.conf [22:34] xorf.conf is empty [22:34] LordMetroid: If it doesn't chew CPU with vesa the bug is in the radeon driver [22:34] LordMetroid: dexconf then [22:35] LordMetroid: Did you file a bug, btw? [22:36] Yes, I have, https://bugs.launchpad.net/ubuntu/+source/xorg/+bug/366224 [22:36] Ubuntu bug 366224 in xorg "Rendering artifact on text in firefox and gedit" [Undecided,New] [22:36] yeah, that bug is going to be the driver [22:39] restarting the X server to see if it take effects [23:25] could someone explain why ubuntu does not ship static versions of libgcj? I couldn't find info on that... [23:27] bug 448789 [23:27] Error: Launchpad bug 448789 could not be found [23:27] debian 448789 [23:27] Debian bug 448789 in gcj-4.2 "Static libgcj is not supplied for gcj-4.2" [Important,Closed] http://bugs.debian.org/448789 [23:27] thanks [23:33] afflux, why is static linking desired? debbuntu generally relies on a "include things only once" rule, which is pretty dynamic-linking-reliant [23:34] directhex: I thought it was common to include .a files in -dev packages. In this case I need them for a work project, so I'll go rebuilding gcj. [23:37] directhex@desire:~$ apt-file search -x \\.a$ | cut -f1 -d':' | sort | uniq | wc -l [23:37] 2354 [23:37] some packages do [23:38] I agree