[00:06] Hello guys! I'm looking for some information on how I could compile a source to be installed just as a version from ubuntu's repo. If I compile it using the default options it will use other folders for binaries, libraries and won't make menu shortcuts [00:07] #ubuntu-motu should help with packaging. [00:08] ok [00:08] will try, thanks [00:10] http://www.klikradio.org/listen.asx <-- the 17 yr old highschooler daughter is live on the air. She even sometimes plays songs I like [00:16] The radio compression is painful. === Chex changed the topic of #ubuntu-devel to: Karmic Alpha-6 released, please help find boot bugs | Launchpad will be down/in read-only from 22:00 UTC until 24:00 UTC for a code update | Archive: FeatureFreeze, UIFreeze | Development of Ubuntu (not support, not app development on Ubuntu) | #ubuntu for support and general discussion for dapper-jaunty | #ubuntu-motu for getting involved in development | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com/Help [00:34] smoser, around? [00:35] here [00:35] but i might have to go at a moments notice. [00:36] smoser, just wanted to show you this [00:37] smoser, I hacked up a first cut of the apport patch for ec2 [00:37] https://bugs.staging.launchpad.net/ubuntu/+source/coreutils/+bug/433845 [00:37] Error: This bug is private [00:37] (no it isn't) [00:37] smoser, is there any other information which might be useful to add for bug reports from ec2 instances? [00:38] everything in the metadata service is useful. [00:38] with the possible explicit exception of ignoring user-data [00:38] (for privacy reasons) [00:38] smoser, is any of it potentially privacy-sensitive? [00:38] i can go through and get a list of what might be private and what not [00:39] at very least, things in there we'd want are image/ramdisk/kernel id [00:39] zone [00:39] stuff like that [00:43] smoser, I'm already getting image/ramdisk/kernel, I can get zone [00:43] smoser, thanks for the feedback === Chex changed the topic of #ubuntu-devel to: Karmic Alpha-6 released, please help find boot bugs | Archive: FeatureFreeze, UIFreeze | Development of Ubuntu (not support, not app development on Ubuntu) | #ubuntu for support and general discussion for dapper-jaunty | #ubuntu-motu for getting involved in development | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com/HelpingWithBugs [02:18] hi, i tested current daily live 64 bit, then i installed htop on my quad core [02:18] and one cpu is permanently at 100% [02:19] X needs it, Intel onboard Q45 btw [02:19] thats definitely not normal [02:21] booted with toram option via network, but should not matter [02:22] mdz, sorry for being dense before. i didn't realize what i was looking at. i thought you were pointing me at code that did the collection, and i couldn't find where it was. [02:26] just tested it on nvidia hardware, same problem, one core running wild of 4 [02:28] also vts are blocked why on earth would somebody do that [02:30] Kano: what do you mean by blocked? [02:31] ctrl-alt+f1 or chvt 1 does not work [02:31] btw. when i install nvidia binary drivers in live mode then cpu usage is 1% [02:31] my guess is the vts are just dead [02:32] which seems to happen not very infrequently when X crashes or is not working properly, in my experience [02:32] on 2 testsystems with different hardware? [02:33] in terms of not getting to the vts I wouldn't be surprised [02:34] smoser, I hadn't noticed you being dense :-) [02:34] I rarely had working vts for a few releases, although Jaunty and Karmic seemed to do good for me [02:42] Enter text here...hello [02:42] smoser, now that you saw it for what it was, do you have further thoughts on how we could make it more useful? [02:43] ok i m having troubles with grub [05:10] * LaserJock stabs apport for freezing up his laptop [05:10] is there a way to force redetection of headphone jack status? [05:20] hyperair: unless there is a mixer control for it, not that I know of. [05:20] hmm [05:20] my headphone jack is acting strangely when resuming from suspend [05:20] as in, my speakers won't mute unless i plug in headphones pull them out and plug them back in [05:21] and if i leave my headphones in before suspending, and pull them out before resuming, the speakers remain muted [05:21] i'd like to write a pm-utils script for it [05:34] hyperair: I think that problem needs solving at the kernel level. [05:36] ah damn =\ [05:36] i should go file a bug then [05:41] argh! gnome-terminal constantly segfaults, yes bug filing has been done. [07:26] Good morning [07:26] mdz: ah, great [07:26] <\sh> good morning pitti :) [07:27] mdke: I just don't know whether we'll get another full langpack export until beta, so that people actually have translated docs; but we will definitively coordinate this for the RC/final [07:27] ArneGoetje: do you still plan new langpacks between now and Monday? [07:27] Hey pitti! [07:27] hey \sh [07:27] hey StevenK! [07:29] * \sh had a nice accident yesterday release-upgrading from a fresh jaunty install to yesterdays karmic .. after do-release-upgrade -d and reboot the machine couldn't find any kernel anymore... [07:30] <\sh> I wonder if it's a machine problem (ibm thinkpad t43 with a new HD) or if something happend during upgrade [07:31] pitti: one update today, and one on sunday... according to crontab... but if it's not wanted, we can drop those [07:31] <\sh> looks like that I have to re-test this way this evening again [07:31] ArneGoetje: oh, that's fine; I actually meant a full export [07:31] ArneGoetje: so that the delta packs get empty, and we don't overflow CDs [07:31] ArneGoetje: actually, hang on [07:31] pitti: I can make one on Sunday [07:31] ArneGoetje: can we skip the one on Sunday [07:32] pitti: ? [07:32] ArneGoetje: ok, full one on Sunday is possible? that would be good [07:32] ArneGoetje: didn't finish typing my original thought yet :) [07:32] pitti: yes, because I juse produced full ones last Sunday [07:32] just [07:32] ArneGoetje: the alternative would be to skip the sunday one and fake a full export today by merging the last full and current delta tarball before building packs [07:33] how long will the export on Sunday take? [07:34] cjwatson: let me check [07:38] cjwatson: the LP export starts at 21:00 UTC on Saturday and takes 9 hours [07:39] ok [07:39] hm, that means we could build final CDs only on Tuesday morning [07:39] sounds doable, but we shoulnd't build them much later [07:39] cjwatson: means, it will be available around Sunday 6:00 UTC. I can then kick of the lp-o-matic run, which takes about 4 hours [07:39] well, if some langpacks aren't built yet, it's not a catastrophe [07:40] pitti: I haven't imported any translations from rosetta yet for ubuntu-docs, the translations included are still those from Jaunty. I tend to import translations after translation freeze because it takes quite a long time to do it and fix the various errors [07:41] good morning [07:41] cjwatson, pitti: BTW: when to you want to have the final langpacks ready? [07:42] three days before RC would be great [07:42] pitti: date? [07:43] ArneGoetje: October 19 [07:44] I'm not sure whether we should do another one for final (we used to in previous releases sometimes), but I'm inclined not to [07:44] pitti: ok, that means the export on 17/18 would be too late? [07:45] the LanguagePackTranslationDeadline is October 22 [07:46] mdke: for that we'd need an upload after RC, since the RC is due on October 22 [07:46] so translators will think (unless corrected) that their translations up until that date will be included in Ubuntu [07:46] mdke: I'm not firmly opposed to do another upload after RC [07:46] pitti: yeah, I'm not saying that the deadline is appropriate, just that this is how it is at the moment [07:46] ArneGoetje: can we move that a day earlier? [07:47] pitti: the export? need to ask the rosetta team to kick off a manual export [07:48] right, the export and the build; so export on oct 16/17, upload on 17, build finishes until oct 19 [07:48] pitti: then we can also make the export on 14/15 the final one... [07:50] pitti: when is release day? 29? [07:50] ArneGoetje: yes [07:51] ArneGoetje: I think we should at least do another full export on Oct 22 [07:51] (on the deadline) [07:51] so that we have the possiblity of another update post-RC for final [07:51] we can still not upload it if the situation doesn't allow us to for some reason, but at least we'll be prepared [07:51] pitti: ok, according to LP cron schedule we have the foolowing export dates: 14, 17, 21, 24 before release. which ones to take and which ones to skip? [07:52] ArneGoetje: that's for deltas, right? [07:52] pitti: that's for full or deltas, depending on how I request it to be [07:52] pitti: probably a good idea to let the translators know the plan so that they don't get bitten if there is no post-RC upload [07:52] ArneGoetje: I'd skip them all, and request a manual full export on 16 and 22 [07:52] mdke: *nod* [07:53] mdke: that's why we are discussing this now. ;) [07:54] pitti: ok, so the last one according to schedule would be on 10. Then two manual ones on 16 and 22. [07:54] ArneGoetje: ok [07:55] pitti: how does the xml export work? Is there a date I need to have the ubuntu-docs translations uploaded by to ensure that they are shipped? [07:55] ArneGoetje: sounds like a plan [07:55] mdke: xml export? [07:56] mdke: for being caught by the Oct 22 export, I think you need to have it uploaded by Oct 19 or so, to give LP some spare time for importing them all [07:56] ArneGoetje: ^ sounds sane? [07:56] pitti: ok, I will announce this to the translators [07:56] pitti: the thing where xml from /usr/share/gnome/help is moved from ubuntu-docs to langpacks [07:57] obviously when I upload ubuntu-docs, it gets stripped out during the build, and then where does it go back in? [07:57] pitti: yes, I guess so [07:57] mdke: oh, ignore me for the import [07:57] it's not really "imported" [07:57] mdke: langpack-o-matic grabs the stripped files produced on build, and merges them into the langpacks [07:57] right, I didn't use the right word there I guess [07:58] mdke: no, I was just confused; I thought it used po files [07:58] * pitti still needs to wake up [07:58] me too :) [07:58] so the langpack-o-matic trick happens on every upload of an affected package? [07:58] mdke: so, from a pure import POV, it needs to be built a about two hours before I start langpack-o-matic [07:58] aha [07:59] still needs to be earlier due to RC freeze and CD building, of course [07:59] mdke: so I think "oct 19" still stands [07:59] ok. that sounds ok, we have the translation deadline on Oct 15 so I can do it over that weekend [08:00] thanks pitti [08:00] great [08:22] pitti: have you seen my post to ubuntu-devel mailing list? [08:34] ArneGoetje: yes, I did, will followup today [08:35] pitti: on my jaunty system the help button for jockey is broken because it points at "ghelp:hardware#restricted-manager" instead of "ghelp:hardware#jockey" (fixed in karmic). Have you had many bug reports about that? if not I presume it's not worth an SRU? [08:36] tseliot: is screen-resolutions-extra still required? We don't need to set VirtualSize any more, and dontzap has gone from xorg.conf as well, no? [08:37] tseliot: it's one of the last two packages which still uses the old policykit API (without depending on it, so it won't even work) [08:37] pitti: screen-resolutions-extra is still required for -ati (which AFAIK doesn't have dynamic front buffer reallocation) [08:38] dontzap (the package) was removed from the archive [08:38] tseliot: does it check that it doesn't fiddle xorg.conf for intel? [08:38] tseliot: s-r-e also has dontzap code AFAICS [08:38] ah, it calls /usr/bin/dontzap [08:38] pitti: yes, it has dontzap code but doesn't use it [08:39] tseliot: ok, so I file a bug for getting it ported to polkit-1 (and the missing dependencies) [08:39] pitti: and no, it won't be used for intel cards as the framebuffer size should be bug enough now that -intel uses UXA [08:39] tseliot: I know; I wondered whether that's actually checked somewhere [08:39] Does anyone know how to run a panel applet in the foreground? Or look at it's stdout/stderr? [08:40] robert_ancell, run it on a command line before adding it to gnome-panel [08:40] robert_ancell, it will sit there until added and the adding will use the running instance [08:40] pitti: the gnome randr applet checks the framebuffer size and calls s-r-e if necessary [08:41] ah, I see [08:41] thanks [08:41] pitti: I'll file a bug report about polkit-1 [08:41] you? ok, thanks [08:41] in s-r-e [08:41] seb128, I did that for gnome-pilot but gdm-user-switch-applet quits immediately [08:41] tseliot: https://wiki.ubuntu.com/DesktopTeam/PolicyKitOneMigration has some links to API and porting guide [08:42] pitti: aah, you said you would do it. Sorry, I misread your message [08:42] pitti: please do it and I'll have a look at it later [08:42] tseliot: *nod* [08:42] robert_ancell, do you already have the applet on the panel? [08:42] seb128, no [08:43] robert_ancell, running /usr/lib/gdm/gdm-user-switch-applet sits on a command line waiting there [08:43] $ /usr/lib/gdm/gdm-user-switch-applet [08:43] ** (gdm-user-switch-applet:3979): DEBUG: GdmUserManager: Found current seat: /org/freedesktop/ConsoleKit/Seat1 [08:43] ** (gdm-user-switch-applet:3979): DEBUG: Creating new user [08:43] etc [08:43] when adding the applet to the gnome-panel config [08:44] robert_ancell, you are sure there is no gdm-user-switch-applet already running? [08:44] seb128, ah, thanks. It must be my local build that is quitting early [08:44] robert_ancell, does it work for the system one? [08:45] seb128, yes, thanks [08:45] you're welcome [08:49] robert_ancell, note that we don't use this applet in ubuntu [08:49] robert_ancell, btw do you know why gdm lists the sabayon user on the login banner? [08:51] seb128, yeah, but I need to patch all of GDM for them to accept the UserManager patch. This is the easiest bit to fix first [08:51] mvo, cjwatson: could you please take a look at my post on ubuntu-devel@ ("Getting translation packages installed automagically") and follow up? Thanks a lot. [08:51] seb128, what uid is sabayon? [08:51] robert_ancell, > 1000 I guess it's a sabayon bug [08:52] seb128, if the UID is greater than 500 and it has a valid shell it gets shown. There is a blacklist of ~20 users that are filtered out (e.g. mysql) [08:52] seb128, the good news is the blacklist is now a gconf key so much easier to fix these sorts of problems [08:53] ok, thanks [08:53] tseliot: done, bug 435709 [08:53] Launchpad bug 435709 in screen-resolution-extra "Needs porting to policykit-1" [Undecided,New] https://launchpad.net/bugs/435709 [08:53] pitti: thanks [08:54] tseliot: do you think you'll have time for this after beta? [08:54] tseliot: if not, please tell me, and I'll try to make some time for it [08:55] robert_ancell, do you plan to land gdm changes before your end of day? [08:55] ArneGoetje: thanks, I just read over it, I think we can do something in u-m for karmic, but not add support to software-store at this point [08:55] robert_ancell, did you put the failsafe session part there too? [08:55] pitti: I doubt I will. There's a new mesa, a new intel driver and some other X stuff that need my attention [08:55] tseliot: ok, keeping on my radar then [08:55] pitti: thanks [08:55] ArneGoetje: did the language-pack names change, i.e. if someone has a full translated jaunty and does a dist-upgrade, will he miss some translations this way in karmic? [08:55] seb128, no changes today [08:56] robert_ancell, ok, not even the failsafe gdm counterpart? ;-) [08:56] robert_ancell, I will test the patch on the bug then and see if I can get that uploaded before the freeze [08:57] seb128, Amaranth has done all the work. I just updated one patch to the current version [08:57] robert_ancell, right I was checking in case you had a bzr ready to upload for this one like gnome-session [08:57] seb128, ok. Nope, I'm winding up now so don't have anything pending [08:58] ok, good [08:58] thanks robert_ancell [09:03] mvo: no. the language-pack names stayed. it's just that language-support-translations and language-support-extra metapackages went away. Those dependencies are now installed by language selector (and can be checked with the script I mentioned), based on which applications (thunderbird, oo.o, etc.) the user has already installed or is going to install [09:03] cjwatson_, hi [09:03] cjwatson_, about bug #434769 [09:03] Launchpad bug 434769 in gdm "Login Prompt on Live CD Text Mode" [Undecided,New] https://launchpad.net/bugs/434769 [09:03] mvo: so, an update will most likely remove those extra translations from the system, since the metapackages are gone which pulled them. [09:04] cjwatson_, /etc/init/gdm.conf has logic for the text argument, isn't that enough? [09:04] " for ARG in $(cat /proc/cmdline) [09:04] do [09:04] case "${ARG}" in [09:04] text|-s|s|S|single) [09:04] exit 0 [09:04] " [09:05] ArneGoetje: urgh, ok. that is definitely something we need to add to the release notes (if it has not been added already) [09:06] ArneGoetje: I've just read your e-mail as well, I'll forward it to ubuntu-translators, since they might be interested (especially Timo Jyrinki, who started the related discussion) [09:07] cjwatson_, oh, the change has been uploaded later that I though the cd images probably didn't have it yet [09:07] ArneGoetje: I do a test run with the upgraer now, it has some cleverness when it comes to translations already, I check if it keeps its hands off them or if it needs extra tweaking [09:10] mvo: ok. Please also have a look at my branch and commit it, so that we have the script ready to be used. :) [09:11] ArneGoetje: I put it on my todo list [09:11] mvo: thanks [09:12] mvo: FYI a script run on my machine with multiple language-packs installed takes about 4 seconds to run and produces accurate output in terms of missing packages. === asac_ is now known as asac [09:26] doko_: checkout http://paste.ubuntu.com/276885/ ... feel free to suggest a fix in xulrunner. the other instruction thing is supposed to be fixed. [09:26] doko_: also i tried -O2 on powerpc [09:27] doesnt help much ... posted the error in the bug. seems a assembler regression somewhere [09:27] doko_: patches welcome :-P (435560 has the new assembler problem when building with -O2) [09:31] ohh joy [09:32] asac: but the powerpc package currently is installable ... so it would help to upload to get an installable one on sparc [09:33] hmm, this is the sparc build ... [09:34] doko_: i worked around the solaris symbol for nanojut by disabling "jit" ... but it doesnt work either. .. seems like gcc header bug on sparc: http://paste.ubuntu.com/276891/ [09:35] yes [09:42] asac: no, xulrunner is soooooooo wrong, redefining 'inline' [09:43] namespace std [09:43] { [09:43] inline namespace __gnu_cxx_ldbl128 { } [09:43] } [09:46] doko_: redefining inline? [09:46] doko_: thats from the /usr/include/c++/4.4/sparc-linux-gnu/bits/c++config.h:214: error: expected unqualified-id before 'namespace' [09:46] asac: yes, please look at the preprocessed source [09:48] doko_: did you already look at the preprocessed source or are you speculating that its a xulrunner bug just on sparc? [09:49] asac: *you* speculate that it is a sparc bug, but apparently it only happens when including xulrunner's headers [09:50] doko_: look on faure in /tmp/ns*.ii [09:50] /tmp/nsWifiScannerUnix.ii [09:50] doko_: ^^ [09:52] asac: yes, as I said: this is expanded to: [09:52] # 212 "/usr/include/c++/4.4/sparc-linux-gnu/bits/c++config.h" 3 [09:52] namespace std [09:52] { [09:52] inline __attribute__((always_inline)) namespace __gnu_cxx_ldbl128 { } [09:52] } [09:54] asac: maybe try to reproduce on x86 with -D_GLIBCXX_LONG_DOUBLE_COMPAT -D__LONG_DOUBLE_128__ [09:55] has anybody experienced a breakage in karmic netwrok inst [09:56] i tried last night and again this morning (CEST timezone), but base installation fails [10:07] ArneGoetje: I just checked, u-m will not auto-remove the translations on upgrade, it shoud still mark them manual instlal I think, so we should add a quirk to u-m (and add it to the release notes) [10:08] * ogra hopes the new usplash artwork was tested on 16pbb displays and doesnt look as awful as xsplash === tkamppeter_ is now known as tkamppeter === dholbach_ is now known as dholbach [10:25] pitti: aren't you afraid that you'll need to blacklist more and more stuff? ^^ [10:25] sebner: well, there's just one left now :) [10:25] it only supports four jmicron bridges [10:26] heh, lucky you === johe|work is now known as johe [10:40] mvo: could you have a look at bug 378075 when you've got some time? There is a patch for it as well [10:40] Launchpad bug 378075 in ddtp-ubuntu "Typos in package summaries" [High,Fix committed] https://launchpad.net/bugs/378075 [10:41] dpm: that is for jaunty? or karmic? [10:42] dpm: nevermind [10:42] yep, jaunty :) [10:51] is there supposed to be an upstart "reload" command? [10:52] ah, mterry is already on it :-) === doko_ is now known as doko [11:08] asac: filed bug #435793 [11:08] Launchpad bug 435793 in xulrunner-1.9.1 "redefinition of 'inline namespace' in installed headers" [High,New] https://launchpad.net/bugs/435793 [11:12] doko: do you see the "definition" in th .ii file? [11:12] both definitions only refer to the linux headers [11:14] shouldnt .ii files give hints where the defines used for expansion come from? [11:14] * asac logs on faure again [11:16] asac: you me want to run -E -dD [11:17] doko: are those compiler flags? [11:18] asac: yes, see gcc(1) [11:20] ok one sec [11:22] somewhat that doesnt dump anything to stdout [11:22] asac: for powerpc, there's now a gcc patch fixing wrong code gen with -Os -fstack-protector. will ask slangasek for a freeze exception (432222) [11:22] which is what i understand it should do on -E [11:22] doko: great. any idea about the assembler parsing (?) bug? [11:23] doko: -E -dD doesnt spit out anything (no clue) ... is -save-temps with -dD doing the same? [11:24] if so i put it to /tmp/nsWifiScannerUnix.ii.dD now [11:24] but doesnt show any xulrunner headers surrounding that inline thing [11:25] so i assume that flag didnt work :/ [11:26] asac: have a look at the other -d* options [11:26] dont see anything that would show more headers in the .ii [11:26] asac: -dM ? [11:28] doesnt work with g++ [11:28] asac: powerpc assembler: again, the preprocessed source would help [11:29] doko: i can get that [11:30] doko: would be precious if you can tell me exactly the command i should run to get _all_ include/expand info into the ii ... the command (with -save-temps) is here: http://paste.ubuntu.com/276972/plain/ [11:31] asac: working dir? [11:31] asac: run the preprocessor only [11:32] karmic network inst: debootstrap program exited with error 127 [11:32] doko: on davis: xptcstubs_ppc64_linux.ii [11:32] in my home [11:33] the compile rerror is in the bug [11:33] also put the original .cpp file there so you can directly fix it if its an official real assembler syntax you understand [11:34] asac: same command line as in the bug report? [11:34] pitti: Hi, you know that you broke tuxonice-userui package with latest usplash-theme-ubuntu change? ;) [11:34] doko: yes if oyu want to be sure i can paste you the command [11:35] Whoopie: uh? [11:35] doko: http://paste.ubuntu.com/276978/plain/ thats the command line (with -save-temps) [11:36] asac: attaching the files to the bug report would help [11:36] doko: cant i just tell g++ to just run preprocessor rather than tweaking that command line until cpp is happy? [11:36] asac: there's no nsWifiScannerUnix.cpp in your home [11:36] doko: well. you are here ... i am here [11:36] pitti: tuxoniceui_usplash uses the progressbar functions to draw a progressbar during hibernate? [11:36] pitti: s/?/. [11:36] doko: if you say you cannot find it i will attach them === cjwatson_ is now known as cjwatson [11:37] hoerup: need full logs [11:37] Whoopie: how does that break hibernation, though? [11:37] pitti: no, the progressbar is not shown anymore. [11:37] right [11:37] doko: i talked about powerpc ... [11:37] the new theme doesn't have a progress bar, by design team [11:38] pitti: that's one benefit of tuxonice instead of vanilla-hibernate. [11:38] doko: now it is on faure [11:39] hi, does anybody else have got 100% load on Xorg with (h)top? [11:40] at least with nv driver (not nvidia binary) with daily iso [11:42] asac: nsWifiScannerUnix.cpp:44:23: error: nsAutoPtr.h: No such file or directory [11:42] nsWifiScannerUnix.cpp:45:27: error: nsWifiMonitor.h: No such file or directory [11:42] nsWifiScannerUnix.cpp:46:31: error: nsWifiAccessPoint.h: No such file or directory [11:43] doko: are you in karmic chroot? [11:43] I don't want the files in your home, but just know the *working directory* (pwd) for calling the compiler [11:43] yes [11:43] doko: sure isnt that even in the paste? [11:43] one second [11:43] on powerpc its: ~/build-area/xulrunner-1.9.1-1.9.1.3+build1+nobinonly/build-tree/mozilla/xpcom/reflect/xptcall/src/md/unix [11:43] no [11:43] thanks [11:43] alls builddeps are installed in karmic chroot [11:44] doko: the hot spot on sparc is in: ~/build/build-area/xulrunner-1.9.1-1.9.1.3+build1+nobinonly/build-tree/mozilla/netwerk/wifi/src [11:44] asac: there is no ~asac/build-area on faure [11:44] ahh [11:45] doko: faure == sparc ... davis == powerpc ;) [11:45] maybe we should focus on one issue to avoid confusion ;) [11:48] hoerup: (ideally, the installer syslog) [11:48] cjwatson: already working on it [11:48] asac: it does work. just use -E -dD as I suggested. so it looks to be a problem in iwlib: [11:49] doko: cool ... can you paste the full -E -dD command you used so i can "learn" [11:49] # 72 "/usr/include/iwlib.h" 3 4 [11:49] #undef IW_GCC_HAS_BROKEN_INLINE [11:49] # 83 "/usr/include/iwlib.h" 3 4 [11:49] #define IW_GCC_HAS_BROKEN_INLINE 1 [11:49] #define inline inline __attribute__((always_inline)) [11:49] g++ -E -dD -fvisibility=hidden -DMOZILLA_INTERNAL_API -D_IMPL_NS_COM -DEXPORT_XPT_API -DEXPORT_XPTC_API -D_IMPL_NS_COM_OBSOLETE -D_IMPL_NS_GFX -D_IMPL_NS_WIDGET -DIMPL_XREAPI -DIMPL_NS_NET -DIMPL_THEBES -DOSTYPE=\"Linux2.6\" -DOSARCH=Linux -DIMPL_NS_NET -I. -I. -I../../../dist/include/xpcom -I../../../dist/include/string -I../../../dist/include/pref -I../../../dist/include/storage -I../../../dist/include -I../../../dist/incl [11:49] ude/necko -I/usr/include/nspr -I/usr/include -I/home/asac/build/build-area/xulrunner-1.9.1-1.9.1.3+build1+nobinonly/build-tree/mozilla/dist/sdk/include -fPIC -fno-rtti -fno-exceptions -Wall -Wpointer-arith -Woverloaded-virtual -Wsynth -Wno-ctor-dtor-privacy -Wno-non-virtual-dtor -Wno-invalid-offsetof -Wno-long-long -pedantic -g -fno-strict-aliasing -fshort-wchar -pthread -pipe -DNDEBUG -DTRIMMED -Os -freorder-blocks -fno [11:49] -reorder-functions -DMOZILLA_CLIENT -include ../../../mozilla-config.h nsWifiScannerUnix.cpp -save-temps [11:49] ok makes sense ... didnt find any redefinition of inline in xul code [11:50] odd ... i used the same approach (just putting to the front) [11:50] IW_GCC_HAS_BROKEN_INLINE ... let me check [11:51] doko: http://paste.ubuntu.com/276994/ [11:52] /* However, gcc 4.0 has introduce a new "feature", when compiling with * '-Os', it does not want to inline iw_ether_cmp() and friends. * So, we need to fix inline again ! [11:53] gah ... [11:53] * I really hate when the compiler is trying to be more clever than me, [11:53] * because in this case gcc is not able to figure out functions with a [11:53] * single call site, so not only I have to tag those functions inline [11:53] * by hand, but then it refuse to inline them properly. [11:53] * Total saving for iwevent : 150B = 0.7%. [11:53] * Fortunately, in gcc 3.4, they now automatically inline static functions [11:53] * with a single call site. Hurrah ! [11:53] * Jean II */ [11:53] when people try to be more clever than a compiler [11:54] obviously 150 bytes is more valuable than several man-hours of time several years later trying to figure out what on earth is wrong :P [11:56] 0.7% ;) [11:59] hoerup: could just be the bug fixed in debootstrap 1.0.18 which I just synced, of course ... [11:59] * Only use dpkg from the chroot, as there is no guarantee dpkg is [11:59] available outside of the chroot (d-i installation for example). [12:00] (which was my muppet error) [12:00] ok [12:01] * asac off for a while doing some NM testing [12:01] cjwatson: last line from syslog "dpkg: not found" [12:03] Does anyone know if it's possible to migrate normal users to be system users on a package upgrade? [12:03] * Laney is touting this question around [12:04] hoerup: right. that's fixed in debootstrap 1.0.18 which you should get shortly. Sorry about that [12:04] Laney, "system users"? [12:04] directhex: --system [12:05] chooses a different uid, doesn't make a home directory and other things [12:32] hm, since upstart is required now, sysvinit can't ever be installable; should we just drop it? [12:32] cjwatson_: ^ opinion? [12:32] it keeps nagging us in the uninstallable report [12:34] Riddell: khelpcenter got demoted to universe in jaunty, but now koffice-data reverse-recommends it and wants to pull it back; promote or fix dependency? [12:36] soren: [12:36] o kvm: kvm [12:36] [Reverse-Depends: Ubuntu.Karmic virt-host seed] [12:36] sommer: I think some seed needs to be fixed to qemu-kvm? [12:37] asac: independently of build the file on ppc with -O2 or -Os, I always get these errors [12:38] yes [12:38] asac: is it correct that the file name has an `ppc64' in it? [12:38] i didnt say that assembler is Os [12:38] and not just ppc? [12:39] i wondered about that too, but uname -a showed 64-bit at least [12:39] 2.6.15-54-powerpc64-smp [12:39] asac: please rebuild, prefixing the build with linux32 ... [12:39] opening an rt ticket ... [12:40] doko: so our ppc isnt 64-bit? [12:40] ok [12:40]