[00:00] lamont: Are you tagging the dapper->hardy upgrade bug reports? [00:12] cjwatson_: when would be a good time to look over procps with me? I've put my merge up if you want to see the disappearing conffile trick... http://people.ubuntu.com/~kees/hardy/ [00:14] keescook: I'm supposed to be on holiday now and packing. UDS? [00:14] I'm only still here to try to finish off PS3 stuff and get this dapper installer work tested [00:14] cjwatson_: works for me -- I'm in no rush. :) [00:15] I can't even do the stuff I want to do with sysctl until 2.6.23 lands. ;) [00:20] jdong: you commented on 145123 - I'm trying to recreate it === rob1 is now known as rob [00:51] woo hoo! the gibbon is giving the werewolf a run for his (her?) money! :) [00:57] hi, if this is the wrong place to ask, please point me to the right place: now that Gimp 2.4.0 final is released, is it going to be officially available for gutsy somehow, or is -rc3 where it's going to stay? === cjwatson_ is now known as cjwatson [00:58] khatahn: gutsy will remain at its current version. At most, the final release may become available in backports [00:58] firefox is some kind of exception then i assume? [00:59] khatahn: you mean the security update? [00:59] i guess so [01:00] khatahn: we have to exempt firefox because we can't otherwise reasonably provide security updates for it [01:01] but it is definitely an exception (and one we'd like to cease, if upstream practices were to improve), not the general rule [01:01] ok, thanks for the information [01:01] (FAOD I doubt the way firefox is handled will change in the near future though) [01:03] interesting. ubuntu's ifconfig reports MiB as MB, but debian's reports MiB as MiB. [01:04] so, my system is hanging hard after an upgrade to gutsy. i've filed a bug, attached a sysrq-t dump, dmesg, and lscpi [01:04] anything else useful to include? [02:02] uh oh [02:02] somebody made a mistake [02:03] chowmeined: For what? [02:04] the new thunderbird broke a bunch of stuff [02:05] all the pages are filled with XML parsing errors after upgrading [02:06] ah [02:06] thats my initial sensationalist description of the problem [02:06] ill refine it in a bit [02:21] hmm. fascinating. ubuntu causes laptop harddrives to die before their time [02:21] * desrt wonders if that has anything to do with his laptop harddrive dying before its time [02:23] zul: you should also send your email about Xen In Hardy to ubuntu-devel [02:24] zul: or ubuntu-devel-discuss [02:24] mathiaz: sure [02:24] mathiaz: sounds reasonable though? [02:24] zul: yes. I've just quickly read the email. [02:25] zul: I don't have a lot of experience with xen (for now). [02:25] zul: so I can't really comment on it. [02:25] zul: but from what I've read, it seems reasonable. [03:01] it looks like bug 59695 got dugg [03:01] Launchpad bug 59695 in acpi-support "default value in power.sh potentially kills laptop disks" [Wishlist,Confirmed] https://launchpad.net/bugs/59695 [03:01] just checked, my load count is 580000 after 18 months :/ [03:04] Amaranth: By default, power.sh does nothing [03:05] mjg59: sure but powertop told me to enable laptop mode :/ [03:05] Amaranth: If powertop told you to jump off a cliff, would you? [03:06] if it got me 5 minutes extra battery life... [03:08] If you enable aggressive power management, then yes, your hard drive life will be shorter [03:09] bdmurray: I suspect it's a problem with the way the new fglrx 8.42 works... [03:10] bdmurray: also, mouse-pasting seems to work in gnome-screensaver! [03:10] If your BIOS defaults to programming your drive to have aggressive power management, then that's an issue with your BIOS rather than us [03:10] IMO that's also a security breach [03:10] especially since now we give "Leave MEssage" a cleartext textbox that can echo back clipboard contents [03:13] bdmurray: lol, I just read your comment on the bug... ignore my comments before the last comment :) [03:16] anyone else think bug 146862 shoudl be tagged as a security bug? [03:16] Launchpad bug 146862 in gnome-screensaver "Should not allow to paste in the Leave Message box" [Medium,Triaged] https://launchpad.net/bugs/146862 [03:16] it discloses the contents of the clipboard before locking the screen [03:17] jdong: seems reasonable to me [03:17] I'm not really the person to ask, but if you're just looking for general populace feedback, there you go. [03:19] tonyyarusso: thanks; I'm just wanting to make sure I'm not completely out of line in thinking that's a security problem [03:19] jdong: np :) [03:20] I can't really think of an instance where it would be a problem for me personally, but in a corporate environment it certainly could be. [03:20] tonyyarusso: the problem is more serious with mouse-paste IMO -- anything a user might've accidentically selected while manipulating or reading would be accessible from the clipboard buffer... [03:21] by someone else... [03:21] jdong: yeah [03:21] I mean the point of locking your screen is primarily privacy from people walking by of what you were doing [03:21] jdong: Unless of course you happened to close the application, and that other clipboard bug cleared it for you. :P [03:21] tonyyarusso: rofl :) [03:21] tonyyarusso: two bugs make a security feature. [03:22] * tonyyarusso still gets very very irritated with how Gnome handles copy and paste between apps.... [03:25] tonyyarusso: more of a problem with the X11 clipboard protocols... [03:25] tonyyarusso: KDE's workaround (a competent clipboard manager) seems good... but not without its limitations [03:27] Indeed, i can lock the screen, click “Leave Message” and paste the clipboard’s contents to the textbox. [03:28] ion_: yeah... I am shocked by how many screensaver vulnerabilities there are this release :( [03:29] ion_: there's currently a compiz-related one where you can kill the screensaver [03:30] As a concept, the “Leave Message” thing is really neat. [03:30] ion_: yeah, without a doubt it's awesome [03:30] http://www.linux-hero.com/rant/explanation-ubuntu-hard-drive-wear-and-tear -- please fix this [03:30] File bug reports at Launchpad. [03:30] smallfoot-: we've heard and were discussing this arleady [03:30] (That one’s already there.) [03:31] smallfoot-: the story's on digg, we don't need 10,000 people coming in here one by one asking for a fix... [03:32] okay, but you need to fix it immediatly, and make a patch [03:32] this is an very important issue, i dont want my hard drive be crashed [03:33] smallfoot-: The fix is right in the article [03:33] if it happen in windows, microsoft would release a patch to fix it, you must patch too [03:33] smallfoot-: Does it happen to you? [03:33] c [03:33] ugh [03:33] i dont know [03:33] wrong tab [03:33] yes, but i cant recommend ubuntu to friends, if it will destroy their hard drives [03:33] ... [03:33] smallfoot-: mine's not dead yet, so I'd take issue with "will destroy" [03:34] smallfoot-: The summary is that Ubuntu does not touch the power management settings on your hard drive. [03:34] If your hard drive behaves like that, it's because your BIOS tells it to [03:34] then why this guy made post in digg.com? [03:34] smallfoot-: You'd have to ask him that [03:34] oh ok [03:34] he have hitatchi [03:34] i have samsung [03:34] smallfoot-: the article is incorrect [03:34] ok [03:35] i dont need worry? [03:35] smallfoot-: he must have set ENABLE_LAPTOP_MODE=YES in acpi-support [03:35] smallfoot-: in a greedy attempt to get more battery life [03:35] smallfoot-: laptop-mode might set an aggressive powersaving mode like this... it's not default config [03:35] smallfoot-: also, many articles on "power saving" under Linux recommend similar tactics [03:35] oh [03:35] smallfoot-: we cannot stop users from trying things like this :( [03:35] ok [03:35] (doesn't laptop-mode get automatically enabled by the installer if it thinks the machine is a laptop?) [03:36] tonyyarusso: no [03:36] i went into services in ubuntu, and it had both APM and ACPI enabled, i only need ACPI, why both enabledf? [03:36] tonyyarusso: no, it doesn't.... [03:36] smallfoot-: Because you can pass an argument to the kernel to tell it to use apm instead of acpi [03:36] mjg59: Oh. It is at least installed by default if not enabled...right? (Or I've *completely* lost it) [03:36] tonyyarusso: Yes [03:36] k [03:37] 67:# Switch to laptop-mode on battery power - off by default as it causes odd [03:37] 68-# hangs on some machines [03:37] tonyyarusso: ^^ [03:37] 69-ENABLE_LAPTOP_MODE=false [03:37] * ajmitch wonders what a 'sane' load cycle count is supposed to be [03:37] ajmitch: one that predicts the hard drive having > 2yr life? [03:38] jdong: well my laptop is coming close to 2 years of age, running ubuntu from about the first day I had it :) [03:38] I'm sure that I had laptop mode enabled at some point [03:38] so I see a count of ~240K [03:38] i bought my comptuer in 2007, ACPI is the successor to APM, i dont need APM [03:38] APM is old technolgoy, i have new computer, and must use new technology! [03:38] mine's an acer, so having a broken BIOS is a given [03:39] ”Welcome to the Ubuntu Settings Wizard. How many months do you want your laptop HDD to last? [ 24 ]” [03:39] smallfoot-: Then it's using ACPI. Don't worry. [03:41] ok, but i disabled APM, cuz i dont want to waste resources like Windows Vista [03:41] i want a mean, lean, clean, machine, thats fast! [03:41] FULL POWER! [03:41] use little resources, run fast, quick, responsive [03:41] That's fine. It's not running anything APM related. [03:41] http://mjg59.livejournal.com/77672.html has a description of the issue [03:41] ok, i disabled it anysays [03:42] oh [03:42] set your BIOS to wait a reasonable amount of time [03:42] mjg59! [03:42] depending on how you use your laptop [03:43] smallfoot-: also see http://www.thinkwiki.org/wiki/Problem_with_hard_drive_clicking#Possible_Cause_and_Speculation [03:44] smallfoot-: it seems like a lot of laptop drives are culpable to this behavior, even without help from laptop-mode [03:44] Solarion: Mm? [03:44] mjg59: hi. :) [03:44] smallfoot-: then do it this way... [03:44] i dont have laptop [03:45] laptop are for girls, im real men, i use desktop computer, with 19" screen! [03:45] with real keyboard, and real mouse! [03:45] smallfoot-: dont use a GUI, dont use a hard drive (flash drive instead), and dont use wireless [03:45] and im gonna buy a 22" screen [03:45] smallfoot-: that explains your back pains, then. :) [03:45] FWIW, I have a 3-y-o laptop with load cycle count 877074 [03:45] and it's still alive and doing well [03:45] Solarion, laptop is more ergonomic? [03:45] smallfoot-: than a desktop, yes. [03:45] backpack-wise, anyhow [03:46] chowmeined, flash drive expensive, and i need the GUI, i dont want use 1960 computer, i must have GUI [03:46] lol [03:46] computers in the 1960s didnt have consoles [03:46] Anyway. Unless anyone has any technical issues they want to bring up, could you move this conversation elsewhere? [03:46] they had punch cards and other nonsense [03:46] The level at which it impacts Ubuntu is pretty obvious [03:48] mjg59: on the other hand, would it be sane to do some monitoring of load cycle counts via a periodic cronjob, and if it increments alarmingly fast pop up a notification balloon of some sort? [03:50] jdong: I don't know that that value is guaranteed to be an actual measurement of the number of load cycles [03:50] If it turns out that a large number of BIOSes are programming this oddly, then we ought to be more proactive. But that's still us defending ourselves about poor BIOSes, not a fundamental failure of Ubuntu. [03:51] mjg59: defending the user against poorly designed hardware, when we can, is a definite plus [03:51] I agree strictly speaking it's not the responsibility of the OS [03:51] Yes, I'd agree. But it's a bonus if we do, not a failure if we don't. [03:52] sure logically it is [03:52] now the other question I have.... [03:52] but people will see it as a failure [03:52] if you drop your laptop once, would park-on-idle have paid off? [03:52] especially with sensationalist postings on digg flying around [03:52] At least now I know how to get on Digg.... :P [03:52] I expect the disk head crashing down on data parts of the hard drive would be a bad thing (tm)? [03:53] jdong: backups pay off [03:53] "UBUNTU WILL MAKE YOUR MOUSE EXPLODE!!1!" [03:54] chowmeined: backups will always pay off, but that's not the question I'm raising [03:54] mine is assessing the benefits vs harm of aggressive idle management [03:55] i think thats a different thing however [03:55] at least.. the laptops that sense they are falling and park the head [03:56] chowmeined: yeah, that is different [03:56] i dunno [03:56] i dont drop my laptop [03:56] chowmeined: I don't think anyone intends to drop their laptop [03:56] well if they do, then they don't deserve one and should send it to me :) [03:57] lol [03:57] when will there be builds of hardy available? i want to get a head start fixing bugs, since its LTS.. id like it to work as well as possible [03:57] there were some serious issues in gutsy unfortunately [03:58] chowmeined: awesome, that's the spirit :) [03:58] like ralink wireless drivers being very very flakey [03:58] ug [04:00] chowmeined: that's not really gutsy's fault, unless I have drifted out of date with that project [04:00] chowmeined: serialmonkey's team has produced a driver that's problematic in a SMP setting [04:02] well i think it was considerably more people than that [04:03] chowmeined: the Ubuntu default kernel is SMP... which means it'll affect everyone running the generic kernel [04:03] Can anyone explain to me why the apmd init script never gets run [04:03] jdong: oh.. thats right, sorry i forgot about that [04:03] godlkwrth: It does, so no [04:03] in feisty [04:04] Oh, wait. Maybe it doesn't. How odd. [04:04] Ah, I don't have apmd installed [04:04] chowmeined: I hung around upstream's form a lot 2 years ago, and looked a lot at the project.... My final conclusion was that there was no real hope for unhacking those drivers for SMP support, but they needed a rewrite... [04:05] chowmeined: seems like they're doing a rewrite and today they still have SMP problems :-/ [04:05] ALSO [04:05] godlkwrth: Ok. If you have apmd installed, the init script it provides will be run. [04:05] is it true that /etc/hdparm.conf gets parsed everytime a harddrive is added [04:05] No? [04:06] apmd is installed by default [04:06] godlkwrth: it's run once at bootup by the hdparm init script [04:06] As far as I know, it's only read when /etc/init.d/hdparm is run [04:06] it appears that there is udev rule that runs /lib/udev/hdparm everytime a hd device is added [04:06] godlkwrth: Then /etc/init.d/apmd is run by default shortly after you enter multiuser mode === smallfoot is now known as smallfoot- [04:07] just try running it [04:07] it doesn't run [04:07] there are links in rc4.d and rc5.d for it [04:08] but it doesn't work [04:08] godlkwrth: What do you mean, "it doesn't work"? [04:08] I don't have an APM BIOS, so I can't run it. [04:08] these are serious problems [04:09] godlkwrth: What do you mean, "it doesn't work"? [04:09] and it doesn't matter what BIOS you have [04:09] try running it, it does nothing [04:09] godlkwrth: is ACPI enabled currently? [04:09] godlkwrth: That means you don't have APM [04:09] every PC has APM as well as ACPI these days [04:09] godlkwrth: No [04:09] godlkwrth: totally untrue [04:10] godlkwrth: You can't run APM and ACPI simultaneously [04:10] The ACPI spec explicitly prohibits that [04:10] well, ubuntu would have us [04:10] looking in the runlevel directories [04:10] godlkwrth: Yes. We try to run apmd, in case your system is running in APM mode. [04:11] If your system isn't running in APM mode, then the script does nothing. [04:11] The decision to run your system in ACPI mode rather than APM mode is made right at the start of kernel boot [04:11] So long before any init scripts are run [04:11] there are no acpi and apm "modes" [04:11] they are either supported or not [04:11] godlkwrth: I'm sorry, you're wrong. [04:12] godlkwrth: You are talking to a guy who eats this stuff for breakfast. :) [04:12] lol, no kidding... [04:13] godlkwrth: See section 15.3.1 of the ACPI 3.0 spec, for instalce [04:13] 15.3.1 Placing the System in ACPI Mode [04:13] " When control is passed to the operating system, OSPM [04:13] will check the SCI_EN bit and if it is not set will then enable ACPI mode by first finding the ACPI tables, [04:13] and then by generating a write of the ACPI_ENABLE value to the SMI_CMD port (as described in the [04:13] FADT). The hardware platform will set the SCI_EN bit to indicate to OSPM that the hardware platform is [04:13] now configured for ACPI. [04:13] " [04:13] So yes, the kernel puts the platform into ACPI mode on boot. From then on, you can't make APM calls. [04:14] modern BIOS support both, I can power off my computer with APM (if acpi isn't supported) [04:14] i.e. I start with noacpi kernel arg [04:14] if you start with noacpi, then yeah, acpi is off. [04:14] godlkwrth: If the kernel never puts the system in ACPI mode, you can use APM if the platform supports it. Many modern systems don't support APM at all. [04:14] You can never use APM and ACPI simultaneously. [04:14] I can also poweroff my computer with APM EVEN IF ACPI is on [04:14] No, you can't. [04:14] Whatever you think you're doing, it doesn't involve APM calls. [04:15] You can't make APM calls if the kernel has enabled ACPI. The APM code is disabled. [04:15] you still can't explain why the script doesn't work... [04:15] The script does nothing because you're in ACPI mode, and so can't use APM [04:15] the script doesn't do any detection of ACPI [04:15] Correct. It checks for APM. [04:16] And you have no APM because the kernel has enabled ACPI, so won't let you use APM. [04:16] doesn't explain why the script is in the default runlevel and isn't enabled [04:16] It is enabled [04:16] It's there so that if your computer doesn't support ACPI or if you've disabled ACPI, it'll run apmd [04:17] if (PM_IS_ACTIVE()) { [04:17] printk(KERN_NOTICE "apm: overridden by ACPI.\n"); [04:17] apm_info.disabled = 1; [04:17] return -ENODEV; [04:17] } [04:17] Line 2259 of arch/i386/kernel/apm.c from the kernel [04:17] PM_IS_ACTIVE will evaluate to true if ACPI is enabled, since switching to ACPI mode is done some time before the kernel gets around to initialising the APM code [04:18] mjg59: I think he's saying he actually does not have an apmd script symlink in rc2.d [04:18] which I cannot explain either -- it's default [04:19] jdong: Then he's deleted it [04:19] mjg59: put plainly, yeah. [04:19] Or some other piece of software has. Regardless, it's not our fault :) [04:52] mjg59: can we please set laptop-mode.conf's default -B mode when activated to something other than the most aggressive? [04:52] mjg59: maybe a value near 200 or 150? === mthaddon changed the topic of #ubuntu-devel to: LP going down in 15 mins for approx 1 hour for update - Development of Ubuntu (not support, even with hardy; not application development on Ubuntu) | #ubuntu for support and general discussion for dapper/edgy/feisty/gutsy | #ubuntu-motu for getting involved in development | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com/HelpingWithBugs | Hardy opened, go wild! [07:00] smurf: ping? === mthaddon changed the topic of #ubuntu-devel to: Development of Ubuntu (not support, even with hardy; not application development on Ubuntu) | #ubuntu for support and general discussion for dapper/edgy/feisty/gutsy | #ubuntu-motu for getting involved in development | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com/HelpingWithBugs | Hardy opened, go wild! === carlos_ is now known as carlos [08:39] morning.. does scott james renmant, the guy behind upstart ever hang out in here ? [08:40] gilligan_: keybuk is your man, and yes [08:40] gilligan_: although upstart-devel is a probably a better place to ask questions [08:40] burgundavia: ah, great..thanks ;] [08:41] if only there was sucha channel ;] or where you referring to a ml ? [08:41] no worries [08:41] gilligan_: https://lists.ubuntu.com/mailman/listinfo/upstart-devel [08:41] s/where/were [08:41] ah ok [08:42] figured that was the next question [08:42] gilligan_: if you are anywhere near Boston, there is an Ubuntu Development Summit next week [08:43] burgudavia: i'm afraid i'm stuck in south Germany hehe [08:49] so #ubuntu+1 isnt open yet? === asac_ is now known as asac [09:08] burgundavia: is there some where I could find information about how the ubuntu build system (building the actual distribution/cd-image) is implemented ? there is some info on organizational matters on the devel wiki but no technical details === doko_ is now known as doko === Gman is now known as GmanAFK [11:48] asac: you probably know, but n-m has issues with ipw3945; dbus needs to be restarted after suspend/hibernate [11:49] this on a thinkpad X60 [11:50] shouldn't need to restart dbus; just /etc/dbus-1/event.d/*NetworkManager* restart [11:55] sladen: ah, right.. good to know. Now I can't suspend anymore, though :) [11:56] works once [11:57] sudo pmi action suspend force ? [11:58] I could suspend from the logout dialog [11:58] click cancel, try again [12:00] hmm, now it works [12:00] I'll try reboot [12:11] tepsipakki: sometimes somewhere in asycronous dbus message soup, there'll be a timeout [12:23] server man .. [12:28] tepsipakki: can you please try the ~ppa3 package from my ppa? [12:28] (its technically a downgrade ... but please test it anyways) [12:29] tepsipakki: https://edge.launchpad.net/~asac/+archive [12:35] * asac back to holiday [12:41] asac: sorry, and thanks, will try :) [13:04] it's so quiet today, so I guess pretty much everyone is having a day off ;) [13:08] slackers [13:09] my thoughts exactly :) [13:09] :) [13:11] Or packing, and getting ready to fly out to the states. [13:12] pfft, excuses excuses [13:15] * TheMuso is packed actually, and is working his way through merges. :p [13:15] * ogra_cmpc will pack tonight [13:40] http://archive.ubuntu.com/ubuntu/dists/gutsy-{updates,security,backports}/Contents-i386.gz is missing. is this intended or a (known) bug? === infinity3 is now known as infinity [14:36] probably something simple, but -- what exactly triggers update-manager on say feisty to know that gutsy upgrade is available (for some reason when pointing to a local full mirror, clients don't see this), yet pointed to a archive.ubuntu.com mirror triggers this ? [14:37] is there going to be voip again for uds again? [14:40] zul: i hope so. Spads should know. [14:40] it's going to be a problem if it doesnt. [14:41] bryang: i'd guess it's the published release announcement. [14:41] bryang: (on archive.ubuntu.com, but probably not mirrored to your mirror) [14:41] Hobbsee: what file(s) would that be ? [14:41] i dont remember. i don't know enough about it :) [14:42] fwiw, I mirror (via rsync) archive.ubuntu.com::ubuntu (so pretty much everything) [14:44] mvo would be able to tell you [14:44] but not for the next few days. [14:44] (oh, yes, that's where all the canonical people are) [14:44] okay, thnx Hobbsee [15:15] Heya === mthaddon_ is now known as mthaddon === cprov is now known as cprov-lunch [16:17] Guys, I've fixed a reported bug on launchpad. It's my first time doing this, and have uploaded all resources and information to apply the fix. Who do I notify/what do I do to get it included in Gutsy/Hardy? See: https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.20/+bug/155436 [16:17] Launchpad bug 155436 in linux-source-2.6.20 "iPod won't mount in Gutsy on 2.6.20-16 kernel" [Undecided,Fix committed] [16:18] OldPink: 2.6.20 isn't the Gutsy kernel. Try #ubuntu-kernel. [16:18] ScottK: Thanks, I'm already there. Just excited about getting my first bugfix in :P === cprov-lunch is now known as cprov === mhb_ is now known as mhb === norsetto_ is now known as norsetto === Mez is now known as Mez|Away === Mez|Away is now known as Mez [18:19] Can somebody pretty please approve the nomination for gutsy SRU on bug 149641? Without the Gutsy task it does not appear in ubuntu-sru's list and the noise (users complaining) won't get less when waiting longer. The fix is quite simple and taken from upstream/Debian. [18:19] Launchpad bug 149641 in logcheck "logcheck fails when auth.log.1.gz missing" [Undecided,Fix released] https://launchpad.net/bugs/149641 [18:19] It would be great if someone even would sponsor the upload right away.. :) === rulus_ is now known as rulus [18:29] blueyed: You ought to subscribe ubuntu-main-sponsors to the bug if you haven't. [18:30] ScottK: ok, done. I've thought that it would have to get approved by ubuntu-sru before.. [18:30] ..and the nomincation approved. [18:30] blueyed: Ah. Didn't realise it was an SRU. Nevermind. [18:30] * ScottK should read more carefully. [18:49] which package contains debugging symbols for dolphin? [18:49] I'm experiencing some crashes related to the fish-protocol [18:53] there are usually -dbg packages for all apps [18:55] not dolphin, the default file manager in kubuntu [18:55] so I suspect they've crumpled it together into kdebase-dbg or something [18:56] yeah, kde packqaging is sometimes a bit confusing [18:58] you don't say... [18:59] partly blame on the kde devs, who send out large bundles with apps, which have to be split [18:59] yeah [19:08] !debug [19:08] For help debugging your program, please see https://wiki.ubuntu.com/DebuggingProcedures [19:08] PhinnFort: https://wiki.ubuntu.com/DebuggingProgramCrash [19:12] Amaranth: thanks [20:10] Anyone familiar with the inner workings of gtk+2.0, and object memory allocation (on x86_64) ? [20:12] how "inner"? [20:14] very :) see my last comments on bug #124336 - there appears to be a random issue with the pointer returned by gtk_list_store_new() being invalid and causing a segfault. I'm building a modified gtk+2.0 now to try and pinpoint it, but was hoping to get some suggestions or clues as what to look for. [20:14] Launchpad bug 124336 in network-manager-applet "nm-applet crashed with SIGSEGV in gtk_combo_box_set_model" [Medium,Confirmed] https://launchpad.net/bugs/124336 [20:15] IntuitiveNipple: I expect the problem is not that the pointer is invalid, but that it's being stored in an incorrect type; this is a common error [20:16] IntuitiveNipple: particularly with gtk, which inherited a type system from gtk 1.2 where it was valid to use "int" and "GType" interchangeable [20:16] y [20:19] Thanks... that makes sense based on my instinct, although I'm having a hard time finding an argument as to why it is so random [20:19] I was looking for 32-bit/64-bit size issues but not found any so far in nm-applet [20:21] I've just added a further comment which shows the pointer values returned by gtk_list_store_new() - does that give any clues? [20:22] IntuitiveNipple: for starters, could you try using %p to print the pointer values instead of %lX? [20:23] I prefer lX when I'm not sure of the size of the values, but sure [20:24] well, for one thing, %lX is a signed integer, which gives confusing output for pointers :) [20:25] Here's the output: http://pastebin.intuitivenipple.net/80 [20:26] fun [20:27] IntuitiveNipple: I think all hell would break loose if what you're saying is actually correct [20:28] Chipzz: Well, it's happening for quite a few people using 64-bit, but not for 32-bit... I've been seeing it for several months but assumed someone else had fixed it during development. [20:29] IntuitiveNipple: if what you're saying would be true, just about every gtk app in ubuntu would be crashing [20:29] so, not very likely :P [20:29] I'm not 'saying' anything yet, I'm looking at the symptoms and trying to pinpoint the fault [20:31] The essence of it is, the fault almost always occurs but seems less-frequent when there's debugging code inserted (which could imply a timing issue of some kind). First time I built a debug-version of nm-applet I couldn't get it to fail. [20:34] IntuitiveNipple: i would blame gtk_list_store_set() [20:35] Company: Once I've got the debug libs ready, hopefully it'll shed some light on it === ogra_cmpc_ is now known as ogra_cmpc [20:44] Company: There are some memcpy() calls that depend on a sizeof() in gtk_list_store_set_n_columns() - I'll keep an eye on them [20:45] IntuitiveNipple: that and gtk_list_store_set is a varargs function [20:45] IntuitiveNipple: har. prototype for wso_wpa_create_phase2_type_model is missing [20:45] IntuitiveNipple: if you check the build log, I expect you'll see "warning: conversion makes pointer from integer without a cast" [20:46] you're kidding!? [20:46] ouch [20:46] missing include :( [20:47] IntuitiveNipple: nope. I just stepped through the code, wso_wpa_create_phase2_type_model returns the right value and then it gets corrupted in the assignment [20:47] Company: missing prototype; there's no include in the tree that defines this function [20:47] Gotchya! well spotted [20:48] slangasek: that's even worse [20:48] Hmmm, that doesn't explain why the value reported *inside* wso_wpa_create_phase2_type_model() is incorrect also when the segfault occurs [20:48] AFAIK the segfault doesn't happen inside type_model() though? [20:49] No it doesn't, well not so far! [20:49] I'll add the prototype and see how we go. It doesn't surprise me since that function looks to be a copy/paste of wso_wpa_create_key_type_model() [20:52] I've added it to wso-private.h, now to test [20:54] wso-wpa-eap.c: In function 'wso_wpa_eap_new': [20:54] wso-wpa-eap.c:296: warning: implicit declaration of function 'wso_wpa_create_phase2_type_model' [20:54] yep, there it is. :) [20:54] wso-wpa-eap.c:296: warning: assignment makes pointer from integer without a cast [20:56] I missed that in all the fuzz. I'm so used to so many releases being shipped with warnings I rarely check them anymore [20:56]