=== dendro-afk is now known as dendrobates | ||
soreau | Hello. | 00:50 |
---|---|---|
soreau | I am trying to detect whether or not a certain package is installed via bash script. Right now I have pkg --get-selections|grep <pkg-name>| grep -v "deinstall" which works to a certain extent but it also is true if packages with <pkg-name> is in another package name is installed. For instance if compiz is not installed and compiz-core is, it will still detect compiz is installed. How should I do this to get an exact match? | 00:52 |
soreau | In short, I want a reliable way to tell if package X is installed from a bash script | 00:53 |
TheMuso | soreau: dpkg -l could be what you are looking after. | 01:11 |
TheMuso | looking for even | 01:11 |
ion | dpkg --get-selections rather | 01:11 |
ion | That is better for a machine to parse. | 01:12 |
TheMuso | ion: right | 01:12 |
soreau | ion: If you see in my first post, that's what I am currently using and it almost works.. just matches any package containing the string | 01:12 |
soreau | TheMuso: As does grepping dpkg -l | 01:12 |
ion | dpkg --get-selections compiz | 01:13 |
soreau | huh | 01:14 |
soreau | ion: That may be exactly what I need | 01:14 |
soreau | ion: What are all the values that the 'other' field can be? purge, install, deinstall etc | 01:15 |
soreau | If I grep for 'install' it will also match deinstall for instance | 01:15 |
soreau | Is there anywhere I can see all the possible values for that column? | 01:16 |
ion | if dpkg --get-selections compiz 2>/dev/null | grep -qE '\<install$'; then ... | 01:16 |
soreau | ion: I'll bet that's exactly what I wanted. Thanks a lot! | 01:18 |
soreau | ion: Thanks for your help. Still trying to figure out what that grep magic does but whatever it is, it is working :) | 01:29 |
ion | \< stands for word boundary and $ stands for the end of a line. | 01:33 |
lifeless | lambda! | 01:36 |
ion | λ | 01:36 |
lifeless | \ is haskell declares an anonymous function ;) | 01:37 |
ion | Indeed. | 01:38 |
ion | Unfortunately, one can’t use λ in its place even with Haskell’s Unicode support (the last time i checked). | 01:39 |
soreau | thanks again | 01:58 |
=== Hellow_ is now known as Hellow | ||
jdong | who can I thank for the policykit-ification of update-manager? | 04:08 |
ScottK | Depends on if you like it or not. | 04:11 |
pwnguin | who wouldn't like an upgrade button that asks for your password then looks at you confused | 04:14 |
=== Whoopie_ is now known as Whoopie | ||
ScottK | In that case, it was definitely not me. | 04:15 |
=== robbiew is now known as robbiew-afk | ||
* cwillu_at_work pokes doko__ | 06:19 | |
cwillu_at_work | doko__, bug #418962 is apparently fixed in bash-4.1; is there any chance of getting a patch into karmic, or updating to that version? | 06:19 |
ubottu | Launchpad bug 418962 in bash "[karmic] [regression] menu-complete in bash no longer completes filenames." [Undecided,Confirmed] https://launchpad.net/bugs/418962 | 06:19 |
Hellow | Well, I'm glad that got fixed. | 06:21 |
cwillu_at_work | s/got/may get/ :p | 06:26 |
Hellow | :P | 06:26 |
cwillu_at_work | looks like debian doesn't have 4.1 packaged yet | 06:26 |
cwillu_at_work | hell, 4.1 hasn't been released at all yet :( | 06:28 |
cwillu_at_work | hmm; 33 patches to 4.0 are on the maintainers site, none of which are obviously the fix to my eyes | 06:39 |
dholbach | good morning | 06:43 |
manitoba98 | Excuse me; can anyone explain why it's necessary to send kernel log messages through dd, rather than letting klogd pick them up directly? Debian doesn't do it that way. | 06:43 |
manitoba98 | Other than "it lets us run it as user klogd", which isn't a reason in and of itself. | 06:44 |
pitti | Good morning | 07:11 |
pitti | pochu: oh, great! | 07:11 |
dholbach | pitti: I uploaded didrocks' new gdl - it will need binary NEWing | 07:17 |
dholbach | didrocks: will you taking care of uploading the 4-5 rdepends? | 07:17 |
didrocks | hey pitti, hi dholbach | 07:18 |
dholbach | didrocks: uploaded sabayon too | 07:19 |
didrocks | dholbach: thanks. You want to handle the transition now and not dep on the old (with previous soname) bin library? | 07:20 |
didrocks | I can do it, if you think it worthes before beta :) | 07:20 |
dholbach | didrocks: it's sitting in the queue now - I'll defer that to the release team / archive admins | 07:21 |
dholbach | I think it'd be a quick job and build times shouldn't be long either | 07:21 |
dholbach | didrocks: just let me know what they think and I can help out if necessary | 07:22 |
didrocks | dholbach: no pb, I'll ping them later today and handle it :) | 07:22 |
pitti | hey dholbach; ok, thanks | 07:22 |
dholbach | super | 07:22 |
didrocks | pitti: so, what do you think about handling those rebuilds before beta? | 07:27 |
pitti | dholbach: should be fine, better for testing | 07:27 |
pitti | pochu: uploaded | 07:28 |
didrocks | dholbach: ok, I put that on my schedule, so :) | 07:28 |
pitti | dholbach, didrocks: please upload them ASAP with proper versioned build-depends, then we can already accept them and let the buildds figure it out | 07:36 |
manitoba98 | Excuse me; can anyone explain why it's necessary to send kernel log messages through dd, rather than letting klogd pick them up directly? Debian doesn't do it that way. | 07:37 |
dholbach | didrocks: can you just put the debdiffs somewhere and I have a look at them? | 07:37 |
dholbach | manitoba98: maybe the people who can answer the question are not around yet... why don | 07:37 |
dholbach | 't you mail ubuntu-devel@lists.u.c? | 07:37 |
manitoba98 | OK, thanks. | 07:38 |
didrocks | dholbach: ok. I'm working on that this evening as soon as I get back home | 07:45 |
pitti | didrocks, dholbach: gdl NEWed | 07:58 |
dholbach | pitti: I'll take the dog for a walk and take care of the rdepends afterwards | 07:59 |
* pitti hugs dholbach | 08:00 | |
didrocks | pitti: Thanks :) | 08:02 |
pitti | ArneGoetje: argh | 08:04 |
pitti | ArneGoetje: may it be that you forgot to remove the langpacks before the rebuild? The delta langpacks are very big | 08:04 |
pitti | they should be empty.. | 08:05 |
pitti | well, a bit too late now, so we just have to cope and kick out some of them | 08:05 |
ArneGoetje | pitti: no. LP 'forgot' my full-request... :( | 08:07 |
pitti | ah, too bad | 08:07 |
pitti | ArneGoetje: did that happen before? | 08:07 |
pitti | we have to get it right for the final | 08:07 |
ArneGoetje | pitti: I have done a merge with the last full-export and int import in lp-o-matic is currently running... do you still want to have it? | 08:08 |
ArneGoetje | pitti: first time it happened... maybe LP had some hickup | 08:08 |
pitti | ArneGoetje: there were some LP rollouts on Friday and apparently over the weekend, too | 08:08 |
pitti | perhaps it forgot that state when it was restarted | 08:08 |
ArneGoetje | pitti: that might be an explanation | 08:08 |
pitti | ArneGoetje: it gets too late, I'm afraid; we need to have CDs tomorrow around European noon | 08:09 |
ArneGoetje | pitti: hum... ok | 08:09 |
ArneGoetje | pitti: the next export will be a full one again... (if LP doesn't forget it again), so users will need to upgrade | 08:10 |
happyaron | mvo: ping | 08:11 |
mvo | hey happyaron | 08:12 |
happyaron | mvo: I have some question on software-center's pot | 08:13 |
mvo | happyaron: sure | 08:13 |
pitti | ArneGoetje: we don't really need a full export after beta, but I guess we should still do it to test that it still works | 08:13 |
ArneGoetje | pitti: yep | 08:13 |
pitti | hey mvo | 08:13 |
happyaron | mvo: there is a po/software-store.pot, po/po/software-store.pot in your 0.4 version tarball | 08:13 |
happyaron | mvo: and now you've uploaded a po/software-center.pot, which template should be in launchpad, and how to name it? | 08:14 |
mvo | happyaron: software-center should be the one, let me fix that in my source. sorry for the confusion | 08:20 |
happyaron | mvo: only po/software-center.pot should be imported, is that right? I have renamed the original template in lp to software-center, and is that po/po/software-store.pot not needed? | 08:21 |
mvo | happyaron: correct | 08:22 |
happyaron | ok | 08:22 |
mvo | happyaron: the old name should go away everywhere | 08:22 |
happyaron | only one template should be fine | 08:22 |
krabador | NetworkManager Applet 0.7.996 is a massacre | 08:23 |
krabador | all right with dhcp, but totally don't works form manual settings | 08:23 |
cjwatson | krabador: please file bugs in Launchpad rather than on IRC; it'll usually get a better result that way | 08:30 |
ArneGoetje | pitti: I could selectively upload a few language-packs it that helps... which languages are we going to ship on the CD? | 08:51 |
pitti | ArneGoetje: that would help indeed | 08:51 |
pitti | ArneGoetje: en es xh pt de fr bn | 08:52 |
ArneGoetje | pitti: which languages do you want? | 08:52 |
ArneGoetje | pitti: ok | 08:52 |
pitti | ArneGoetje: that won't help the DVD, but at least we can have good CDs | 08:52 |
pitti | ArneGoetje: thanks! | 08:53 |
ion | What, not fi? I’m appalled. | 08:53 |
* pitti hugs ion | 08:54 | |
ion | :-) | 08:54 |
=== dholbach_ is now known as dholbach | ||
didrocks | pitti, dholbach: all libgdl universe rdeps are uploaded. Do you want that I give a hand on main ones? | 09:11 |
pitti | didrocks: that would be nice, if you have some time to test them? | 09:12 |
pitti | I hope most will "just build"? | 09:12 |
didrocks | pitti: universe ones have built like a charm :) I do some minimal test (at work, it's more difficult, spawing a slow VM on my windows machine :/) | 09:13 |
dholbach | uploaded | 09:13 |
dholbach | gnome-python-extras | 09:13 |
dholbach | and committed to the brancht oo | 09:13 |
dholbach | too | 09:13 |
dholbach | should be all set now | 09:13 |
didrocks | great :) | 09:14 |
=== asac_ is now known as asac | ||
pitti | mbiebl: any particular reason why you made libpolkit-gtk-1-dev arch: any? (it's arch: all in Ubuntu) | 09:24 |
\sh | moins | 09:24 |
tgpraveen | any chance we will get theora 1.1 in karmic? | 09:25 |
seb128 | tgpraveen, any chance you will stop asking every day if we will get new major version of software or features weeks after feature freeze now? ;-) | 09:26 |
tgpraveen | seb128: i just wanted a yes/no answer :) | 09:32 |
seb128 | tgpraveen, which is try for my question too ;-) | 09:33 |
seb128 | tgpraveen, dunno, we didn't consider it it's way after feature freeze, if you have excellent recent open a bug asking for an exception | 09:33 |
seb128 | recent -> reason | 09:33 |
asd123123 | http://17fc5c36.linkbucks.com | 09:44 |
=== yofel_ is now known as yofel | ||
pitti | tseliot: does nvidia 96 actually work on karmic? | 09:53 |
tseliot | pitti: yes, it should. Why? | 10:06 |
pitti | tseliot: ok, thanks | 10:07 |
pitti | just cleaning the DVD seeds, and they had some nonexisting versions any more | 10:07 |
tseliot | pitti: aah, ok | 10:07 |
* directhex has downgraded his nvidia - 185 is crashy crashy crashy | 10:09 | |
dholbach | is there a bug open for "gdm only comes up after pressing alt-f7"? | 10:43 |
pitti | dholbach: I don't see one | 10:54 |
dholbach | ok | 10:54 |
dholbach | I just have this on one machine | 10:54 |
dholbach | booting will only show me a black screen until I press alt-f7 | 10:54 |
cjwatson | it happens on the live CD for me | 10:55 |
seb128 | dholbach, there is one | 10:59 |
seb128 | dholbach, bug #434361 | 10:59 |
ubottu | Launchpad bug 434361 in gdm "gdm sometimes fails to start after new upstart merge" [Medium,Incomplete] https://launchpad.net/bugs/434361 | 10:59 |
dholbach | thanks seb128 | 10:59 |
seb128 | it doesn't seem a gdm bug | 11:00 |
pitti | seb128: I retitled the bug, so that it's a bit easier to find | 11:00 |
dholbach | hum - I don't need to log in and restart gdm for gdm to work | 11:00 |
cjwatson | seb128: certainly in my case, gdm starts fine, it just doesn't switch vt | 11:01 |
dholbach | ah ok, comment 7 says the same on amd64 | 11:01 |
seb128 | dholbach, right, read comments on the bugs, just people got confused before finding that ctrl-atl-f7 was working | 11:01 |
cjwatson | ah | 11:01 |
seb128 | cjwatson, well some people have the issue with kdm too | 11:02 |
seb128 | and workaround are weird there | 11:02 |
seb128 | ie booting with splash workaround the bug | 11:03 |
seb128 | or reinstalling linux for some users | 11:03 |
seb128 | if somebody has a clue how to debug that help is welcome there | 11:03 |
seb128 | it's probably a beta blocker too | 11:04 |
cjwatson | yeah | 11:04 |
cjwatson | I wonder if the usplash init script needs to be migrated | 11:04 |
cjwatson | we used to shut down usplash before starting gdm; now we don't | 11:05 |
cjwatson | that seems very likely to be the cause | 11:05 |
seb128 | "without splash" before | 11:06 |
seb128 | but right, that makes sense | 11:06 |
joaopinto | mvo, sorry for the duplicate bug repos on software-center I didn't noticed they were reported on the previous project name :\ | 11:12 |
joaopinto | reports | 11:12 |
mvo | joaopinto: no problem, thanks for reporting the stuff, the rename caused some confusion unfortuatnely | 11:13 |
cjwatson | hmm, this is slightly tricky, I *really* don't want to reproduce the busy-wait in the usplash init script in upstart-wordl | 11:14 |
cjwatson | world | 11:14 |
pitti | does upstart send a signal after a process has terminated? | 11:16 |
cjwatson | well, even that wouldn't be sufficient, since the init script waits for a while and kills it if it doesn't die by itself | 11:17 |
cjwatson | really, I think we want to have upstart take over supervision of the running usplash process when it starts, and pretend that it's a job, so that we can use the 'kill timeout' stuff | 11:19 |
dholbach | seb128: do you know why bug 424311 is set to "low"? :) | 11:40 |
ubottu | Launchpad bug 424311 in gnome-screensaver "gnome-screensaver does not put display to sleep" [Low,New] https://launchpad.net/bugs/424311 | 11:40 |
seb128 | dholbach, why not? | 11:41 |
dholbach | I find it a bit annoying that the screen does not go to black | 11:41 |
dholbach | it just flickers every few minutes | 11:41 |
seb128 | dholbach, because it has not been confirmed and happens to one user only | 11:41 |
dholbach | ok | 11:41 |
* dholbach can confirm | 11:41 | |
Laney | and me | 11:42 |
dholbach | Laney: can you add your details to the bug too? | 11:43 |
seb128 | dholbach, did it start recently? | 11:43 |
dholbach | seb128: no, a week I'd say | 11:44 |
seb128 | dholbach, it's probably not the same bug then | 11:45 |
Laney | done | 11:45 |
seb128 | dholbach, yours probably started with the 2.28 update | 11:45 |
dholbach | seb128: you think? | 11:45 |
seb128 | well that one has been opened 3 weeks ago | 11:45 |
dholbach | maybe a bit longer already - I really dunno | 11:45 |
seb128 | well, you are the third one to mention screensaver today | 11:45 |
seb128 | I really think it's new since 2.28 | 11:45 |
seb128 | I would prefer a new bug if that's the case rather than abusing an old bug about something different | 11:46 |
Laney | no I had it before then | 11:46 |
seb128 | ok, cool, let's see if somebody is interested to work on that | 11:46 |
seb128 | we don't have anybody working on screensavers that I know | 11:46 |
seb128 | but maybe somebody will be annoyed enough to look at it ;-) | 11:47 |
seb128 | I put it on my list but I've like one hundred bugs already on my karmic list so low chance I look at that | 11:47 |
Laney | can you view the sleep/wake events? | 11:48 |
seb128 | Laney, gnome-screensaver --no-daemon --debug | 11:50 |
Laney | cool | 11:50 |
Laney | i'll do it when i get home | 11:50 |
seb128 | but the power management tab is a gnome-power-manager thing | 11:51 |
seb128 | not a gnome-screensaver one | 11:51 |
Laney | yeah | 11:52 |
TomaszD | hey asac, would you care to upload a translation update for ubufox? https://bugs.launchpad.net/ubuntu/+source/ubufox/+bug/437604 | 12:34 |
ubottu | Launchpad bug 437604 in ubufox "[pl-PL] Polish translation update for ubufox [attached]" [Undecided,New] | 12:34 |
asac | assigned myself now. thx. TomaszD | 12:41 |
TomaszD | asac, cool, thanks! | 12:41 |
asac | TomaszD: if you request a merge into lp:ubufox it would be even quicker ;) | 12:41 |
asac | but not needed | 12:41 |
TomaszD | asac, I know, but I don't have time to learn another DVCS just to upload a few files :P | 12:42 |
asac | ll | 12:44 |
asac | kk | 12:44 |
cjwatson | seb128,pitti: FYI I'm experimenting with http://paste.ubuntu.com/280312/ - feel free to comment while I break my system with it :) | 12:47 |
pitti | cjwatson: "start on stopped gdm" sounds a bit overzealous; shouldn't this rather mean "start on (rc0 or rc6)" or so? | 12:49 |
pitti | nice, this looks so much cleaner than the current init script with its looping | 12:51 |
=== Tonio__ is now known as Tonio_ | ||
amitk_ | is it a bad time to upgrade (dist-upgrade fails with lots of python errors) | 12:57 |
AnAnt | can someone review LP 438092 ? | 13:04 |
ubottu | Launchpad bug 438092 in texlive-bin "Candidate revision texlive-bin_2007.dfsg.2-7ubuntu1" [Undecided,Confirmed] https://launchpad.net/bugs/438092 | 13:04 |
sistpoty|work | james_w: bug #404003: source package is actually opencore-amr, sorry for the inconvenience | 13:04 |
ubottu | Launchpad bug 404003 in ubuntu "please sync libopencore-amr for karmic from unstable" [Wishlist,Confirmed] https://launchpad.net/bugs/404003 | 13:04 |
AnAnt | I dropped this Ubuntu change: build with g++ -fno-tree-ter on armel, since debian now uses g++-4.4 | 13:05 |
james_w | thanks | 13:05 |
AnAnt | james_w: btw, I've done devscripts merge | 13:06 |
james_w | thanks AnAnt | 13:06 |
AnAnt | james_w: LP 414298 | 13:07 |
ubottu | Launchpad bug 414298 in devscripts "Please merge devscripts_2.10.53 from Debian unstable" [Undecided,Confirmed] https://launchpad.net/bugs/414298 | 13:07 |
AnAnt | james_w: that's actually devscripts_2.10.55 | 13:07 |
cjwatson | pitti: I believe it has to start after gdm, or the chvt ordering goes wrong | 13:08 |
cjwatson | pitti: at least it certainly used to fail if you tried to start usplash_down before gdm | 13:09 |
cjwatson | it does mean that it won't run if you don't have a *dm, I concede | 13:09 |
=== doko__ is now known as doko | ||
kalla | some audio sync problem with videoplayer in ubuntu... :( | 13:48 |
Laney | http://launchpadlibrarian.net/32596463/buildlog_ubuntu-karmic-armel.sound-juicer_2.28.0-1_FAILEDTOBUILD.txt.gz - check out the UCF installation. Does that look transient? | 13:50 |
pitti | re | 13:51 |
pitti | cjwatson: ah, perhaps it'd need to be (stop gdm && (rc0 || rc6))? | 13:51 |
=== ember_ is now known as ember | ||
cjwatson | pitti: mm, maybe | 13:55 |
* cjwatson is wrestling with getting the initramfs->upstart import stuff Keybuk hacked in to work properly | 13:55 | |
pitti | ogra, lool: so bug 417009 is believed to be fixed now, right? You just keep it open until it's verified, or does it need more work? | 13:57 |
ubottu | Launchpad bug 417009 in openoffice "all openoffice apps die in 'com::sun::star::ucb::InteractiveAugmentedIOException' on armel in karmic" [Unknown,Confirmed] https://launchpad.net/bugs/417009 | 13:57 |
amitk_ | mvo, doko: do you deal with python/apt? I have a failed dist-upgrade and here is the result of trying to fix it. http://pastebin.ubuntu.com/280364/ | 13:58 |
mvo | amitk_: hm, that smeels like disk corruption "EOFError: EOF read where object expected | 14:00 |
mvo | " | 14:00 |
lool | pitti: Not sure why it isn't closed; I know we're actualy hiding an issue in some binaries instead of fixing it | 14:00 |
lool | doko: Did you intend to keep it open? | 14:01 |
amitk_ | mvo: should I force reinstall some packages? | 14:01 |
asac | ogasawara: plz check the patch in 334413 ... thx | 14:03 |
ogasawara | asac: will do | 14:04 |
mvo | amitk_: try reinstalling python2.6 | 14:05 |
doko | lool: yes, we have a work around, not a fix. just take it off the release radar once the build is in the archive | 14:07 |
lool | doko: Ok thanks | 14:07 |
lool | pitti: ^ | 14:07 |
amitk_ | mvo: 'sudo dpkg-reconfigure python2.6' seems to have done the trick | 14:07 |
AnAnt | can someone look at LP 416949 | 14:07 |
ubottu | Launchpad bug 416949 in console-setup "Keyboard layout toggle does not work anymore in karmic" [Undecided,New] https://launchpad.net/bugs/416949 | 14:07 |
doko | amitk: bad files, zero lenght? | 14:08 |
mvo | amitk_: hm, that is a bit scary | 14:08 |
amitk_ | doko: no, I checked that the files existed | 14:08 |
=== marjomercado is now known as marjo | ||
amitk_ | liw: moi! does Computer Janitor track how often a package is used? I am wondering why it is trying to remove deboostrap, kvm, qemu, etc. | 14:19 |
cjwatson | oh usplash, why do you mock me | 14:34 |
directhex | you were naughty in a previous life? | 14:38 |
zul | archive-admins: i think linux-ec2 is sitting in binary-new can we get that in the archive? please k thx bye | 14:54 |
pitti | zul: nothing like *ec2* in the NEW queue | 15:01 |
zul | pitti: hmm.. | 15:01 |
=== robbiew-afk is now known as robbiew | ||
pitti | zul: see https://edge.launchpad.net/ubuntu/+source/linux-ec2/2.6.31-300.3, nothing NEW there | 15:04 |
=== dendrobates is now known as dendro-afk | ||
zul | pitti: but it said it built? | 15:09 |
pitti | right | 15:09 |
=== bjf-afk is now known as bjf | ||
fbond | mvo: Have a second to discuss possible bugs in python-apt? | 15:14 |
fbond | Just want to confirm that the bugs exist as I see them before filing a report. | 15:15 |
liw | amitk, computer janitor does not track how much packages are used; if you click on the package in the UI what does it tell you as the reason for the removal? | 15:16 |
fbond | mvo: Nevermind, just confused by the various semi-dict-like interfaces. | 15:18 |
mdz | ttx, good afternoon, how is the test going? | 15:20 |
ttx | mdz: there is an issue with eucalyptus-cc upstartification. httpd child processes are crashing. That prevent sthe node from being correctly polled | 15:22 |
ttx | (I'm testing your branch) | 15:22 |
ttx | I'm slowly closing in on the issue, but axis2c is not easy to debug | 15:22 |
mdz | ttx, the command line and config file used to start apache2 are identical. I did drop the LD_LIBRARY_PATH change because it seemed superfluous, but we can put it back in if needed. did you test if that is the problem? | 15:24 |
mdz | (command line -> except for "-D foreground") | 15:24 |
ttx | mdz: I just managed to get a /etc/init/eucalyptus-cc.conf that doesn't crash. Now I need to find what in the combo of things I added fixed it. | 15:24 |
mdz | ttx, send me the diff? | 15:24 |
mdz | pitti, zul, I believe slangasek took care of it already: | 15:26 |
mdz | linux-ec2 | 2.6.31.300.0 | http://archive.ubuntu.com karmic/main Packages | 15:26 |
mdz | pitti, zul was talking about the linux-ec2 binary package from linux-meta | 15:26 |
zul | mdz: thanks | 15:26 |
pitti | mdz: ah; well, I didn't see that in NEW either; | 15:27 |
pitti | if it's done now, fine | 15:27 |
mdz | pitti, zul, I can confirm that it's installed in the images at http://uec-images.ubuntu.com/karmic/current/ubuntu-uec-karmic-i386.manifest | 15:27 |
ttx | mdz: http://pastebin.ubuntu.com/280428/ | 15:28 |
ttx | mdz: I'll reduce the patch to see what actually fixed the problem | 15:28 |
ttx | mdz: http://pastebin.ubuntu.com/280430/ is better | 15:29 |
=== dendro-afk is now known as dendrobates | ||
mdz | ttx, /etc/eucalyptus/axis2/lib is a symlink to /usr/lib/axis2/lib | 15:30 |
giskard | dude how you write stats (cpu , load) in the motd (before call pam_lastlog i mean) | 15:31 |
mdz | ttx, it still segfaults when I use your LD_LIBRARY_PATH change | 15:32 |
giskard | btw, hi all | 15:32 |
ttx | mdz: yes, EUCALYPTUS=/ and LD_LIBRARY_PATH do not help... | 15:32 |
mdz | ttx, have you debugged the segfault? | 15:33 |
ttx | mdz: no | 15:33 |
mdz | I have a stack trace | 15:34 |
ttx | ah | 15:34 |
ttx | env AXIS2C_HOME=/etc/eucalyptus/axis2 | 15:34 |
mdz | ttx, http://people.canonical.com/~mdz/temp/eucalyptus-cc-apache2-crash.png | 15:34 |
ttx | seems to be the critical one. Let me doublecheck that. | 15:34 |
=== fddfoo is now known as fdd | ||
AnAnt | fabrice_sp__: thanks | 15:35 |
fabrice_sp__ | AnAnt, for? | 15:35 |
=== fabrice_sp__ is now known as fabrice_sp | ||
AnAnt | fabrice_sp: the tablelist bug | 15:36 |
AnAnt | bugs | 15:36 |
mdz | it's doing this: | 15:36 |
mdz | service = adb_getKeysType_get_serviceTag(request, env); | 15:36 |
mdz | 15:36 | |
mdz | response = adb_getKeysResponseType_create(env); | 15:36 |
mdz | status = AXIS2_TRUE; | 15:36 |
mdz | rc = doGetKeys(service, &outCCCert, &outNCCert); | 15:36 |
mdz | and crashing in DoGetKeys | 15:36 |
fabrice_sp | AnAnt, ahhh, you're aelmahmoudy. You're welcome :-) | 15:37 |
mdz | ttx, I still get the segfaults after setting AXIS2C_HOME | 15:37 |
ttx | grmbl | 15:37 |
AnAnt | fabrice_sp: yup | 15:37 |
=== nixternal_ is now known as nixternal | ||
ttx | mdz: there is a combination of factors, probably two bugs. | 15:39 |
TomaszD | mvo, hi, have you perhaps noticed that there is something wrong again with update-manager's template? Even the one waiting in the queue is not up-to-date, and the strings that have not been changed are now displayed in English | 15:39 |
TomaszD | mvo, could you take a look at this issue? | 15:39 |
cjwatson | pitti: I think this usplash upstartification is now working - I just needed to crank the initial timeout up a little bit | 15:39 |
cjwatson | pitti: since native upstart jobs don't talk to usplash (at the moment?), so the timeout has a habit of expiring | 15:40 |
ttx | oooooh | 15:40 |
pitti | you rock! | 15:40 |
mvo | TomaszD: I have a look | 15:41 |
mvo | TomaszD: is ths about bug #438077 ? | 15:41 |
ubottu | Launchpad bug 438077 in update-manager "update-manager's translation template is very likely not up to date again" [Undecided,New] https://launchpad.net/bugs/438077 | 15:41 |
cjwatson | hmm, it's not quite stopping properly if gdm isn't in use though | 15:41 |
TomaszD | mvo, indeed it is | 15:42 |
TomaszD | mvo, I'm especially worried about the policykit dialog window, where there's a new string about the system policy preventing updates or something close to that | 15:42 |
ttx | mdz: can you crash it with the full patch ? | 15:42 |
mdz | ttx, segfault is at: home = strdup(getenv("EUCALYPTUS")); | 15:42 |
mvo | TomaszD: that is comes from aptdaemon | 15:43 |
ttx | mdz: yes... but if you set that alone, the polling to the node doesn't really occur. | 15:43 |
* ttx rechecks | 15:43 | |
TomaszD | mvo, ok, but the rest is still worrying | 15:45 |
cjwatson | pitti: actually - I wonder if the right answer would be for the gdm job to emit a stop-splash event | 15:45 |
cjwatson | or even a dm-starting event | 15:45 |
mvo | TomaszD: it is, I targeted it | 15:45 |
mdz | ttx, adding export EUCALYPTUS=/ to the upstart job doesn't seem to get it into the apache2 enivronment for some reason | 15:45 |
mdz | or env | 15:45 |
cjwatson | pitti: that would save having to keep track of whether gdm actually gets started (e.g. 'text' on command line) | 15:45 |
TomaszD | mvo, thanks | 15:45 |
ttx | mdz: you need AXIS2C_HOME | 15:45 |
ttx | with env EUCLYPTUS and AXIS2C_HOME set it works | 15:46 |
cjwatson | mdz: the init(5) manual page documents env KEY=VALUE and export KEY, but not export KEY=VALUE | 15:46 |
ttx | all the other changes in the patch are superfluous. | 15:46 |
mdz | cjwatson, env KEY=VALUE is what I'm using | 15:47 |
ttx | mdz: http://pastebin.ubuntu.com/280457/ works for me | 15:48 |
mdz | it works for HTTPD_CONF=/var/run/eucalyptus/httpd-cc.conf | 15:48 |
mdz | but not for EUCALYPTUS=/ | 15:48 |
apachelogger | asac: ping | 15:49 |
mdz | ttx, if you look at /proc/xxx/environ do you see EUCALYPTUS=/ there? | 15:49 |
ttx | mdz: yes | 15:50 |
mdz | ttx, I can't get that to work for some reason | 15:51 |
mdz | ttx, ararrggghh | 15:52 |
mdz | using "restart" doesn't seem to process the new env stanzas | 15:52 |
mdz | but 'stop' and 'start' did | 15:52 |
mdz | keeeeyyyybbuuuuuuukkk.... | 15:52 |
ttx | mdz: uh :) | 15:52 |
* dholbach hugs mdz | 15:52 | |
ttx | there were two bugs, the segfault was masking the other. | 15:52 |
* ttx has been banging his head against that one for a few hours :) | 15:53 | |
cjwatson | hmm, now I have usplash stuck on | 15:53 |
cjwatson | whoopsie | 15:53 |
highvoltage | heh, so now you have to press alt+F1 AND alt+7? | 15:54 |
mdz | ttx, gdb made quick work of the segfault, it was clear that it was missing $EUCALYPTUS | 15:54 |
mdz | if my changes to the upstart job had actually worked, the second one would have been quick too ;-) | 15:54 |
mdz | ttx, is the cc registration working? | 15:56 |
ttx | mdz: yes. | 15:57 |
cjwatson | highvoltage: no, it's just hosed. undoubtedly my fault | 15:57 |
ttx | mdz: I tested from a daily + upgrade though, which is a slightly different test path | 15:57 |
mdz | ttx, it doesn't seem to re-run the registration when I restart the cc, though I thought it would | 15:57 |
mdz | ttx, I've pushed a new auto-registration branch | 15:58 |
mdz | with the env fixes | 15:58 |
mdz | ttx, are we ready to merge it into ubuntu and upload? | 15:58 |
ttx | mdz: I think so. I was planning to rebuild and test, but the sooner it lands on the daily, the better | 15:58 |
asac | apachelogger: ? | 15:59 |
apachelogger | asac: any news on trademark in kubuntu-firefox-installer? also, I really really think that making the firefox packages conflict and replace kubuntu-firefox-installer is the best option at hand :) | 16:00 |
apachelogger | asac: btw, is firefox-gnome-support a recommends? | 16:01 |
mdz | ttx, yes, dialing | 16:02 |
ogra | could $archive_admin_of_the_day get linux-fsl-imx51 and linux-mvl-dove out of NEW for me ? | 16:03 |
mdz | ttx, uploaded | 16:03 |
mdz | cjwatson, can you review and approve? | 16:04 |
ogra | james_w, ^^^ ? | 16:04 |
mdz | and spin new server ISOs? | 16:04 |
asac | apachelogger: can you get me a mail with screenshots etc. ... would be easier for me to check with mozilla if i had something like that. | 16:04 |
ttx | mdz: I'll spend some cycles this evening validating that iso | 16:04 |
apachelogger | asac: sure | 16:05 |
mdz | ttx, what about the public IPs issue? | 16:05 |
james_w | ogra: already in the middle of the queue | 16:05 |
ogra | merci :) | 16:05 |
=== bjf is now known as bjf-afk | ||
mok0 | Any hints on how to easiest solve the getline() problem with karmic's libc= | 16:20 |
mok0 | ? | 16:20 |
cjwatson | mok0: rename the local function | 16:25 |
mok0 | cjwatson: I've been doing that but in this particular instance I don't want to | 16:25 |
mok0 | cjwatson: not if I can avoid it. | 16:26 |
mok0 | cjwatson: The error comes from a 4-line change in stdio.h | 16:26 |
cjwatson | that's correct, but it was an addition to POSIX | 16:26 |
mok0 | cjwatson: the addition is to include getline as a std. function? | 16:27 |
ttx | mdz: did you commit the -0ubuntu9 changes to lp:~ubuntu-core-dev/eucalyptus/ubuntu ? | 16:27 |
ttx | mdz: about the public IP issue, maybe we can discuss it on the call. | 16:28 |
mok0 | cjwatson: earlier you had to define __GNU to get it | 16:28 |
mok0 | cjwatson: now it's controlled by __USE_XOPEN2K8 which is apparently defined somewhere. | 16:28 |
cjwatson | it used to be a glibc extension, but POSIX incorporated it | 16:29 |
cjwatson | I'm sure you can use feature test macros to declare that you want an older version of POSIX | 16:29 |
cjwatson | info libc, search for feature test | 16:29 |
mok0 | cjwatson: I'll look into it, thanks | 16:29 |
cjwatson | that sort of declaration might be a stopgap; I wouldn't recommend it long-term | 16:30 |
mok0 | cjwatson: Well, we have an FTBFS list a mile long, ultimately upstream should fix it | 16:30 |
sistpoty|work | mok0: if it uses autotools, you could also add a configure check for it | 16:30 |
cjwatson | sistpoty|work: shouldn't need that ... | 16:31 |
mok0 | sistpoty|work: it's not a missing function, it's a dual declaration | 16:31 |
mdz | ttx, yes I did | 16:32 |
sistpoty|work | mok0: yep, in the sense of #ifndef HAVE_GETLINE local definition #endif, and let HAVE_GETLINE set by autotools, but you could also just remove the local definition though ;) | 16:32 |
mok0 | sistpoty|work: AFAICT it's not a standard getline() | 16:33 |
sistpoty|work | mok0: then you'll need to rename it, like cjwatson first wrote | 16:33 |
mok0 | sistpoty|work: I have to edit a dozen files then | 16:34 |
mok0 | sistpoty|work: which is why I'd rather unset __USE_XOPEN2K8 | 16:34 |
ttx | mdz: ok, got it. launchpad is late. | 16:34 |
cjwatson | do not touch that macro directly | 16:34 |
cjwatson | it is internal | 16:34 |
cjwatson | there are public macros available for your use | 16:35 |
mok0 | cjwatson: I am looking at features.h to find it | 16:35 |
cjwatson | specifically, I should think, #define _POSIX_C_SOURCE 200112L | 16:35 |
mok0 | cjwatson: that doesn't sound like a public macro either | 16:37 |
cjwatson | well, it is. | 16:37 |
cjwatson | it's in info libc. | 16:37 |
mok0 | cjwatson: ah, thanks. | 16:37 |
lolufail | hi | 16:38 |
lolufail | I'm experiencing kernel oops in with my fresh raid5->dmcrylt->lvm2->ext4 configuration | 16:39 |
lolufail | I cannot mkfs.ext4 | 16:39 |
lolufail | http://pastebin.com/m56628b69 | 16:39 |
lolufail | should I file a bugreport? | 16:40 |
sistpoty|work | mok0: or you could use seddery: for i in *c; do sed -e 's/getline(/getline_local(/' $i | sponge $i; done | 16:43 |
mok0 | sistpoty|work: I could, yeah | 16:43 |
sistpoty|work | (and review the debdiff, that only good things matched) | 16:44 |
mok0 | sistpoty|work: or an appropriate #define getline anothergetline | 16:44 |
nh2 | does it make any difference if I autoreconf before PPA upload? | 16:47 |
mok0 | Gotta go see you later | 16:48 |
=== nxvl_ is now known as nxvl | ||
slangasek | ttx: question regarding the instructions in http://testcases.qa.ubuntu.com/Install/ServerEConfig | 17:21 |
slangasek | ttx: will euca-bundle-image do the right thing if you pass it /boot/vmlinuz instead of /boot/vmlinuz-2.6.31-11-generic? (I.e., will it follow symlinks?) | 17:22 |
slangasek | davmor2: ^^ | 17:22 |
ttx | slangasek: no clue. I'd need to try | 17:22 |
slangasek | ttx: low priority, but if it doesn't work I think we need to fix it so it does; otherwise we have to embed kernel ABI versions in our test case | 17:23 |
ttx | would work with /boot/vmlinuz-$(uname -r) though | 17:23 |
slangasek | true | 17:23 |
slangasek | davmor2: can you switch the testcase to use the above rune? | 17:23 |
davmor2 | slangasek: np's | 17:23 |
ttx | slangasek: in the end we would rather test the UEC kernel and ramdisk provided with the images | 17:24 |
ttx | slangasek: see http://uec-images.ubuntu.com/karmic/current/ | 17:25 |
ttx | though I'd like to retest those on my now-working UEC | 17:26 |
* ttx disappears again | 17:26 | |
slangasek | ttx: oh? so we're now publishing kernels / initrds alongside the images? | 17:26 |
slangasek | that's a rather late change to the release publishing | 17:27 |
ttx | slangasek: bug 429106, see smoser | 17:27 |
ubottu | Launchpad bug 429106 in vm-builder "kernel and initramfs should be available for uec" [Medium,Fix committed] https://launchpad.net/bugs/429106 | 17:27 |
slangasek | ok | 17:27 |
ttx | slangasek: this can be reverted easily. | 17:28 |
ttx | :) | 17:28 |
slangasek | well, one of the consequences of distributing kernels in this fashion is that they aren't accompanied by source code... or even a clear indicator of what version of the source they are | 17:29 |
slangasek | (s/clear/unambiguous/; obviously the ABI version narrows it down, but it's not unique) | 17:29 |
slangasek | smoser: see above; I don't think we should be publishing kernels / initramfs in this fashion, this has GPL compliance implications | 17:30 |
=== Tonio__ is now known as Tonio_ | ||
smoser | slangasek, i'm not 100% bent on publishing kernels in that fashion | 17:43 |
smoser | i have 2 thoughts though: | 17:43 |
smoser | 1. we're essentially doing this with ec2 kernels on ec2 (we make them available to run, but similarly to above, there is no absolute link of source -> binary) | 17:44 |
smoser | 2. if a user of UEC wants to create a kernel/initrd/guest image and they're running on their local system anything other than the release for which they want to create the guest, they're kind of SOL | 17:46 |
smoser | we need to provide them with a kernel and initrd that they can easily pair with a.) our uec images b.) ones they make themselves | 17:46 |
slangasek | smoser: 1) is also a bug we need to solve, then; it just wasn't as obvious to me in that case as it is when I'm being pointed at the uec-images directory listing :) | 17:46 |
slangasek | smoser: 2) I guess that's fair - we should still get the GPL reqs addressed before we start publishing in this fashion, though | 17:47 |
smoser | i agree that that directory listing is far from perfect . and in that bug, i mentioned that.. | 17:48 |
mvo | TomaszD: I think the u-m problem is fixed, it turns out something is meddling with gettext | 17:49 |
smoser | slangasek, i really would like your insight on how to make that directory output easily usable and "correct" in other ways also. | 17:49 |
smoser | the other thing that has come to my mind is the deletion of non-current kernel images (as i belive is done in the archives). that should then apply here too i think | 17:49 |
TomaszD | mvo, great :) | 17:50 |
slangasek | smoser: generally, by integrating anything you need into the scripts from the 'cdimage' branch that I copied to nectarine, and reusing those | 17:53 |
smoser | slangasek, i guess i dont follow that. but i'll look more into it. | 17:54 |
* hyperair wonders if anyone's looking at sreadahead | 17:55 | |
slangasek | smoser: cf. http://cdimage.ubuntu.com/daily/current/, which always includes an autogenerated header... we should be reusing the existing toolset for UEC publishing | 17:55 |
slangasek | smoser: e.g., fixing the daily builds to call the checksum-directory tool as I had pointed out | 17:56 |
davmor2 | slangasek: what do you want me to do then leave the docs for now and wait till after beta to alter them? So we at least have a working testcase for beta? | 17:58 |
slangasek | davmor2: no, change them to use the $(uname -r) rune | 17:59 |
cjwatson | hyperair: an upload went in last night to speed it up a lot | 17:59 |
cjwatson | hyperair: if that's not what you mean, be more specific ... | 17:59 |
smoser | slangasek, i can/will get the directory checksumming in. i was asking more about the naming of files and such. I'd like to name the kernels such that they can easily be referred to with, as you suggested, a static identifier (ie, like /vmlinuz). so that any scripts based on directory output wouldn't have to fish around for versions | 18:01 |
davmor2 | slangasek: how's that now? http://testcases.qa.ubuntu.com/Install/ServerEConfig | 18:03 |
slangasek | smoser: server-side scripts, or some sort of user-side scripts? | 18:03 |
slangasek | smoser: for server-side, I don't think it should be much of a problem, since we *have* to publish information about what source package this came from | 18:04 |
smoser | slangasek, i was meaning client side. | 18:04 |
slangasek | davmor2: looks good to me, thanks | 18:05 |
smoser | ie, when you document "heres how to install our UEC image on your cloud" | 18:05 |
davmor2 | np's | 18:05 |
smoser | best if you don't have to say: download http://uec-images.u.c/releases/<release>/fish-for-kernel-name.img.gz | 18:05 |
smoser | or what nt | 18:06 |
smoser | not | 18:06 |
slangasek | smoser: ok. can be done; shouldn't be done until we first have the GPL issue addressed prominently | 18:06 |
smoser | but rather http://uec-images.u.c/releases/i386/kernel.gz | 18:06 |
smoser | slangasek, ok. so what do i need to do on the gpl issue. | 18:06 |
slangasek | well, I would prefer ubuntu-uec-karmic-i386-vmlinuz-generic-pae | 18:06 |
smoser | slangasek, thats fine | 18:06 |
smoser | but a static name. rather than with a version in it | 18:07 |
slangasek | smoser: we need a header that gives the user a link to the corresponding source code | 18:07 |
smoser | but ew dont want to lose the source of that, and ideally would like to make the version info available | 18:07 |
slangasek | that can be included in the header | 18:07 |
hyperair | cjwatson: hmm that's what i meant. (it was no-op-ing) | 18:07 |
hyperair | but i still don't see a pack file anywhere =\ | 18:07 |
hyperair | i guess i'll have to reboot twice to see | 18:08 |
slangasek | we can't just point at the archive, because the matching kernel version isn't guaranteed to be there - so we'll want to link to the launchpad UI, specifying an exact version number | 18:08 |
slangasek | smoser: and the kernel / initrd are being extracted from vm-builder, right? So the information in the manifest is guaranteed to always be correct? | 18:08 |
smoser | slangasek, right. they pull from vmbuilder image | 18:09 |
slangasek | ok | 18:10 |
slangasek | in that case, perhaps we don't have any more GPL problem here than for anything else in the image, it just wasn't obvious that this is the case | 18:10 |
slangasek | (on EC2, OTOH, we may still have a GPL issue) | 18:10 |
mdz | slangasek, can you notify me, kirkland, ttx and mathiaz when there is a server build with eucalyptus -0ubuntu9 available? | 18:14 |
mdz | kirkland, and will you take responsibility for notifying eucalyptus and giving them a download URL? | 18:15 |
emgent | k | 18:18 |
davmor2 | mdz: most of the testcase is written now I'm just finishing off a couple of bit then I'm going to do a fresh install follow the testcase to make sue I didn't miss any steps but if there is going to be a respin I might wait till after that in case anything changes. | 18:19 |
mdz | davmor2, the main change in the next build is that the registration happens automatically, so that it actually works when it comes up | 18:19 |
mdz | davmor2, there will be at least one more respin after that, where there will be one additional question asked during installation | 18:20 |
davmor2 | mdz: I'll wait then and make the mods once everything is in place. | 18:20 |
mdz | davmor2, it's definitely worth trying with the next build (with eucalyptus -0ubuntu9) | 18:21 |
=== bjf-afk is now known as bjf | ||
davmor2 | mdz: I've taken the steps to split out the current post install config to minimise damage to the cluster and node install pages so there are 3 pages currently. So the changes shouldn't cause too much interruption. I did actually mean I would wait for the updated version before I removed and text, I just didn't say it very well :) | 18:24 |
davmor2 | s/and/any text | 18:24 |
slangasek | mdz: ack | 18:25 |
pitti | james_w: still here by chance? | 18:26 |
pitti | james_w: just going through couchdb update with kenvandine | 18:26 |
james_w | I am | 18:26 |
pitti | james_w: I'm curious why you recommended using Breaks: for a package split with moving files, instead of hte standard Conflicts:? | 18:26 |
mdz | davmor2, if you send me a link, I can give you a heads up as to the expected changes | 18:27 |
james_w | pitti: that was Colin's advice | 18:27 |
davmor2 | mdz: http://testcases.qa.ubuntu.com/Install/ServerEConfig | 18:27 |
pitti | james_w: because it's confiles? | 18:27 |
pitti | conffiles | 18:27 |
pitti | since for non-conffiles, breaks: isn't strong enough | 18:27 |
james_w | I'm not sure | 18:28 |
davmor2 | mdz: I'm assuming steps 1-7 can be dropped | 18:28 |
slangasek | pitti: because the Conflicts: part of Conflicts/Replaces was never strictly necessary when one package takes over files from another | 18:28 |
mdz | davmor2, exactly | 18:28 |
pitti | slangasek: so using Breaks:/Replaces: sometimes makes apt's job easier? | 18:29 |
slangasek | yep | 18:29 |
pitti | ah, ok; thank you | 18:29 |
mdz | davmor2, I think the euca-bundle-image call should use the kernel from the downloaded image, not the kernel from the local system (in /boot) | 18:29 |
mdz | davmor2, http://uec-images.ubuntu.com/karmic/current/ provides the kernel and ramdisk which should be used | 18:30 |
slangasek | pitti: in fact, Conflicts+Replaces has a special meaning that people usually forget about: it means "this package should be preferred as a replacement for the other package on upgrade" | 18:30 |
slangasek | mdz: though not under names that are going to stay as they are currently | 18:30 |
davmor2 | mdz: slangasek thought that would be a cleaner way around kernel updates. I just bowed to the greater knowledge | 18:31 |
cjwatson | once upon a time, people used to use Conflicts in addition to Replaces because Replaces didn't work when the packages were installed the wrong way round | 18:32 |
cjwatson | that was fixed ages ago, but old habits die hard | 18:32 |
slangasek | davmor2: no, I only said we shouldn't have hard-coded version numbers in the test case; mdz is talking about grabbing the kernel from somewhere else entirely | 18:32 |
cjwatson | but if there really is a reason to avoid installing the old package simultaneously, then Breaks is independently a reasonable thing to do; it isn't really Breaks+Replaces | 18:32 |
mdz | slangasek, davmor2, those two kernels are not the same | 18:33 |
mdz | the one installed on the cluster controller is a -server kernel | 18:34 |
mdz | the downloadable one should be the -virtual kernel | 18:34 |
mdz | smoser, confirm? | 18:34 |
cjwatson | the -virtual vmlinuz is the same as the -server one, no? | 18:34 |
smoser | -virtual vmlinuz is subset of modules of -server | 18:34 |
cjwatson | last I looked, -virtual was constructed by stripping down -server at the file level | 18:34 |
slangasek | mdz: well, previously the instructions said 'generic', and the kernel available for download on uec-images is -generic-pae | 18:35 |
davmor2 | mdz: on my cluster install it reads 2.6.31-11-generic | 18:35 |
smoser | for i386, the -virtual is subset of -generic-pae | 18:35 |
smoser | for x86_64, subset of -server | 18:35 |
cjwatson | huh, interesting | 18:35 |
slangasek | ah | 18:35 |
mdz | cjwatson, it is | 18:35 |
mdz | more or less | 18:35 |
mdz | the point being that we want to install a lighter kernel on virtual machines | 18:35 |
slangasek | it's not lighter | 18:36 |
slangasek | the vmlinuz is supposed to be the same | 18:36 |
mdz | slangasek, lighter as in fewer modules | 18:36 |
slangasek | ah - the initrd will be lighter, sure | 18:36 |
mdz | and a smaller disk footprint | 18:36 |
mdz | which matters when it will be used as the basis for hundreds of variations | 18:37 |
davmor2 | mdz: so should the lines relating to kernels/initrd's be specific to the download then? | 18:37 |
smoser | lines related to ?... | 18:38 |
smoser | one thing you *could* do is something like what pygrub does (or at least what i thought it does). mount the image, poke around for kernels | 18:38 |
davmor2 | smoser: http://testcases.qa.ubuntu.com/Install/ServerEConfig | 18:39 |
smoser | ah | 18:39 |
smoser | i think we want to have the test download the correct paired kernel. | 18:40 |
jdstrand | kees, mdeslaur: can you look at a potential fix for bug #437854: http://paste.ubuntu.com/280574/ | 18:45 |
ubottu | Launchpad bug 437854 in libvirt "apparmor profile denies access to alsa" [Medium,In progress] https://launchpad.net/bugs/437854 | 18:45 |
* kees looks | 18:45 | |
kees | jdstrand: erm, fix is where? | 18:46 |
jdstrand | actually, /tmp/pulse-*/ should be /tmp/pulse-*/* | 18:46 |
jdstrand | I'm not completely keen on it, cause abstractions/audio pulls in quite a bit | 18:47 |
jdstrand | I'm not sure how to get away without it though | 18:47 |
kees | I didn't see a patch on that bug -- you were thinking of just adding the audio abstraction plus the /tmp/* paths? | 18:48 |
jdstrand | kees: that whole stanza (no, I don't have a patch) | 18:48 |
jdstrand | kees: I need to add ipc_lock, deny kill, add /dev/shm/pulse-shm*, all of it | 18:49 |
kees | oh! I totally missed the pastebin URL, reading | 18:49 |
kees | jdstrand: I use this for pulse: http://paste.ubuntu.com/280576/ | 18:50 |
mdeslaur | jdstrand: why does it need /tmp/pulse-*/ w,? | 18:50 |
jdstrand | mdeslaur: it needs to create that dir | 18:51 |
jdstrand | kees: I can play with it more, but it didn't like when I used 'owner' | 18:51 |
kees | hrm, weord | 18:51 |
kees | weird too | 18:51 |
jdstrand | kees: I think because kvm is running as root, but it is trying to access my uid's pulse stuff | 18:51 |
kees | aah, yeah, good point | 18:51 |
mdeslaur | jdstrand: hmm...why is it trying to access _your_ pulse? | 18:52 |
mdeslaur | also, why isn't /tmp/pulse-* in abstractions? | 18:52 |
kees | mdeslaur: that's a separate bug that needs to be fixed | 18:52 |
jdstrand | mdeslaur: pulse has been a moving targer | 18:52 |
jdstrand | target | 18:52 |
jdstrand | we add stuff as it comes up | 18:52 |
jdstrand | mdeslaur: I think you are right regarding root and my pulse | 18:53 |
jdstrand | it may have failed cause of the /tmp stuff | 18:54 |
jdstrand | kees, mdeslaur: anyhoo, pulse abstraction tinkering aside, I wasn't totally pleased with the access to the files in /dev or ipc_lock | 18:55 |
jdstrand | it did flat refuse to allow 'kill' atm | 18:56 |
jdstrand | also, if I use @{HOME} as is, that expands to all in /home/*/... | 18:57 |
jdstrand | as this process is root, DAC won't do anything, so I want to be extra careful with the access | 18:57 |
* jdstrand -> afk | 18:58 | |
davmor2 | mdz: How is that I've added the 64 version after the /!\ hopefully it's right now. | 18:58 |
mdz | davmor2, I don't see a step where you actually download the kernel and initrd from uec-images | 19:01 |
mdeslaur | jdong: I probably wouldn't use the abstraction in order to not have the @{HOME} | 19:01 |
mdeslaur | jdong: sorry, wrong auto-completion | 19:01 |
mdeslaur | jdstrand: ^ | 19:01 |
mdeslaur | jdstrand: it shouldn't need anything besides the root-owned pulse files...nothing in /dev should be needed I think, and the esd stuff shouldn't be used anymore | 19:03 |
davmor2 | mdz: is that not what the wget and gunzip lines do? It's still set for release beta but that's because that's what these notes are specific for? | 19:03 |
mdeslaur | uhmm...that, of course, assumes libvirt is compiled with pulse support | 19:03 |
ttx | mdz: nurmi just confirmed the mailhost patch, I can commit it, do a quick build/run test and release | 19:14 |
ttx | mdz: just checking you aren't already committing that. | 19:15 |
jdstrand | mdeslaur: ok, those are excellent points. I'll play with it. thanks! | 19:18 |
jdstrand | mdeslaur: it does need access to /dev/snd/* | 19:25 |
mdeslaur | huh | 19:26 |
jdstrand | mdeslaur: it is kvm confinement that I'm adjusting | 19:26 |
jdstrand | mdeslaur: it uses alsa | 19:26 |
mdeslaur | oh :( | 19:27 |
=== fabrice_sp__ is now known as fabrice_sp | ||
=== rickspencer3 is now known as rickspencer3-afk | ||
jono | folks, I am still getting a stack of boot messages before xsplash, should I file bugs against them | 19:32 |
robbiew | jono: what are they? | 19:33 |
robbiew | jono: if they start with some decimal number in brackets: [ 1198.829291] foo: blah blah...then file against kernel | 19:34 |
jono | robbiew, I get that, but also some kind of pci message, an apparmour message about firefox | 19:35 |
jono | and I also see some services starting, like couchdb | 19:35 |
jono | are these messages logged? dmesg and messages doesnt show them | 19:36 |
doko | hmm, firewire disks are not mounted when they are already connected on start | 19:36 |
=== porthose is now known as porthose|afk | ||
robbiew | jono: I know Keybuk is aware of the firefox/apparmour stuff...I think that's handled by kees and the gang | 19:36 |
jono | also, it seems KMS starts a little while into the boot | 19:37 |
robbiew | jono: /var/log/syslog | 19:37 |
jono | as opposed to right away | 19:37 |
mathiaz | doko: right - I've seen the same behavior with usb disks | 19:37 |
cjwatson | jono: apparently starting KMS early loses us 2-3 seconds, so Keybuk is trying to avoid it | 19:37 |
cjwatson | which is why it was taken out of the initramfs | 19:37 |
jono | syslog doesnt show anything | 19:37 |
jono | cjwatson, ahh | 19:37 |
kees | jono: if you have firefox/apparmor issues, please open a bug for it. jdstrand has been doing that profile's development. | 19:38 |
doko | mathiaz: hmm, the connected usb disk is always mounted on boot | 19:38 |
jono | is there a way I can pause the boot to note down these messages? | 19:38 |
jono | I don't seem to have them logged anywhere | 19:38 |
mathiaz | doko: well - I connected my encrypted usb disk before booting my karmic laptop and I wasn't prompted for the usualy password | 19:38 |
cjwatson | ctrl-s might do it | 19:38 |
cjwatson | (ctrl-q to unpause) | 19:38 |
kees | jono: oh, do you mean "Skipped: ...firefox?" that's known and will be uploaded shortly. | 19:38 |
robbiew | jono: you can also check if they are still on vt1 | 19:39 |
jono | kees, thats the one, it skips the profile | 19:39 |
doko | mathiaz: is there already a bug report? | 19:39 |
mathiaz | doko: hm - I don't know - I haven't had time to investigate that yet | 19:39 |
jono | robbiew, they are on vt1 | 19:39 |
jdong | kees: a higher-level AA question; does the apparmor/lsm framework provide a feasible way to do some Windows-UAC-like interactive denials? | 19:40 |
jono | I wonder if I can pipe them into a file somehow | 19:40 |
kees | jono: if you want to subscribe: bug 435285 | 19:40 |
ubottu | Launchpad bug 435285 in apparmor "apparmor log message on booting" [Low,Fix committed] https://launchpad.net/bugs/435285 | 19:40 |
jono | thanks kees | 19:40 |
cjwatson | jono: doubt it | 19:40 |
jdong | kees: in some cases it'd be a cool feature to have for things like Firefox and an occasional, say, thumbdrive access | 19:40 |
jono | ok, I will copy it into a file by hand in vt | 19:40 |
jono | 2 | 19:40 |
jono | brb | 19:40 |
* cjwatson admires jono's optimism in believing that there's a shell on vt2 at that point | 19:40 | |
kees | jdong: not presently (it would require user-space call-back and blocking the syscalls, etc) | 19:40 |
* robbiew admires jono's optimism that we will fix his bugs | 19:41 | |
robbiew | lol | 19:41 |
jdong | kees: *nods* would be pretty neat if implemented though, but yeah I was afraid it'd be nontrivial in how the userspace<->syscall interaction would be | 19:41 |
mdz | ttx, see mail, I don't think we need it | 19:43 |
jdstrand | kees, mdeslaur: I think this is as good as it gets: http://paste.ubuntu.com/280609/ | 19:43 |
mdz | ttx, so the only further change needed at this point is to fix up the question text to explain the allowed format | 19:43 |
ttx | mdz: that was my guess as well. | 19:43 |
jono | cjwatson, I am flicking into vt2 right now and copying over content from vt1 by hand into a text editior | 19:44 |
mdz | ttx, I'm working on that now | 19:44 |
jono | so I can get the boot messages | 19:44 |
cjwatson | I'd use a digicam myself ... | 19:44 |
kees | jdstrand: how about making that a "pulse" abstraction, and then adding that to the "audio" abstraction? | 19:44 |
jono | cjwatson, I am most of the way through it now | 19:44 |
ttx | mdz: ok. If you don't need me I'll call it a day. I'll test ISOs first thing tomorrow morning, and do some UEC image validation as well. | 19:44 |
jdstrand | kees: well, I mentioned the @{HOME} issue with mdeslaur | 19:45 |
kees | jdstrand: hrm, yeah, good point. | 19:45 |
jdstrand | kees: @{HOME} will expand to /root and /home | 19:45 |
cjwatson | sigh, this is confusing. is it really taking 30 seconds to get out of the initramfs, or am I doing something to block it by mistake? | 19:45 |
jono | brb rebooting | 19:45 |
jdstrand | kees: I am adding /tmp/pulse-*/ stuff to the audio abstractino though | 19:45 |
cjwatson | oh, hah - & on that strace might be an idea | 19:45 |
jdstrand | we never saw it cause our gui apps use user-tmp | 19:46 |
kees | jdstrand: aaah, yeah | 19:46 |
jdstrand | I'm really not keen on the /dev access, but there isn't anything I can do but add it and comment on it | 19:47 |
jdstrand | people can fine-tune it if needed | 19:48 |
mdz | ttx, I've pushed new question text to the Ubuntu branch | 19:48 |
mdz | ttx, it should be ready to test and upload now | 19:48 |
ttx | mdz: ok, building and quicktesting now. | 19:48 |
jono | ok | 19:50 |
jono | the error I get on boot shortly after it says grub is loading is: | 19:50 |
jono | e1000e 0000:00:19.0: pci_enable_pcie_error_reporting failed 0xfffffffb | 19:50 |
jono | this is just before the AppArmour profile | 19:50 |
jono | I assume this is a kernel bug | 19:51 |
jono | pgraner, ^ | 19:51 |
robbiew | I get a similar message | 19:52 |
robbiew | so it may be a thinkpad thing | 19:52 |
jdstrand | oh, I do too | 19:52 |
robbiew | worth reporting...could simply be a nag message | 19:52 |
robbiew | for apw to help clean up | 19:52 |
jdstrand | I do not have a thinkpad, but an intel mobo | 19:52 |
jono | ok, I will file it | 19:52 |
pgraner | jono, robbiew: I'm not seeing it on my non-stinkpad | 19:52 |
robbiew | driver related then? | 19:52 |
jono | pgraner, should I file against the kernel? | 19:53 |
pgraner | robbiew: most likely we updated to .31.1 in the last upload | 19:53 |
jono | I also have the messages I see before xsplash, will file a bug against those too | 19:53 |
pgraner | jono: yea and assign it to rtg and make it high | 19:53 |
jono | pgraner, will do | 19:54 |
robbiew | jono: that pci_enable message should be in your dmesg output as well | 19:54 |
jono | pgraner, what package name do I pass to ubuntu-bug? | 19:54 |
jono | robbiew, thats where I dug it out from | 19:54 |
mdeslaur | jderose: looks good to me | 19:54 |
robbiew | linux | 19:54 |
mdeslaur | jderose: sorry, not you | 19:54 |
pgraner | jono: -p linux | 19:54 |
mdeslaur | jdstrand: looks good to me | 19:54 |
jono | pgraner, ta | 19:55 |
jono | pgraner, might be sensible to have an alias for 'kernel' that points to 'linux' | 19:55 |
pgraner | jono: give me the bug number when your done | 19:56 |
jono | pgraner, will do | 19:56 |
pgraner | jono: there you go thinking like a normal person... | 19:57 |
robbiew | lol | 19:57 |
robbiew | pgraner: I'm now deeply offended | 19:57 |
robbiew | lol | 19:57 |
maco | -p sound aliasing to linux but doing the "apport-collect -p alsa-base" thing might be nice too | 19:58 |
maco | since right now thats two steps | 19:58 |
jono | pgraner, it was already reported: | 19:59 |
jono | https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bug/436370 | 19:59 |
ubottu | Launchpad bug 436370 in linux ""pci_enable_pcie_error_reporting failed" error when booting" [Undecided,New] | 19:59 |
pgraner | jono: ack, I'll get rtg on it | 19:59 |
jono | pgraner, I haven't changed the priority on this yet - you may want to do this | 20:00 |
ttx | mdz: just updated the branch (got rev631), it's not the text with the "delimited by spaces" example. | 20:00 |
jono | robbiew, where do you think I should file these boot messages before xsplash? is that an upstart thing? | 20:00 |
pgraner | jono: ack | 20:00 |
jono | pgraner, cheers buddy | 20:00 |
* ttx pushes the revised version. | 20:01 | |
=== bjf is now known as bjf-afk | ||
robbiew | jono: just add ubuntu-boot in the tag and subscribe me | 20:02 |
jono | robbiew, will do | 20:02 |
robbiew | then I can triage | 20:02 |
jono | thanks pal | 20:03 |
jono | robbiew, done: https://bugs.edge.launchpad.net/ubuntu/+source/upstart/+bug/438335 | 20:07 |
ubottu | Launchpad bug 438335 in upstart "Boot messages show before xsplash kicks in" [Undecided,New] | 20:07 |
robbiew | ack | 20:08 |
slangasek | mdz, kirkland, ttx, mathiaz: server ISO build is up | 20:10 |
* mathiaz syncs | 20:10 | |
ttx | slangasek: ok | 20:10 |
jono | kirkland, do you still need some community testing? | 20:10 |
jono | or are you doing fine with the level of testing going on now? | 20:10 |
davmor2 | slangasek: is this the uec image with update? | 20:11 |
slangasek | davmor2: no, this is a server ISO with updated eucalyptus packages | 20:12 |
davmor2 | slangasek: :) thanks | 20:12 |
mdz | slangasek, thanks | 20:15 |
mdz | mathiaz, davmor2's work in progress test plan is at http://testcases.qa.ubuntu.com/Install/ServerEConfig | 20:16 |
mdz | kirkland, ttx, ^^ | 20:16 |
siganderson | a me su karmic è sparito sistema->amministrazione->servizi ... è normale? | 20:20 |
siganderson | oops... sorry | 20:22 |
pgraner | jono, robbiew: can you try booting with apparmor=0 on your grub line and see if that message goes away | 20:22 |
robbiew | I could...if I cared :P | 20:22 |
robbiew | sure one sec | 20:23 |
pgraner | robbiew: dooh | 20:23 |
jono | siganderson, sorry this is an english speaking channel | 20:24 |
siganderson | I have nomore system->amministration->services is it normal? If i type services-admin in the terminal it says that I must install gnome-system-tools, but I have it already installed | 20:24 |
jono | pgraner, glad Robbie cares :-) | 20:25 |
pgraner | jono: yea tell me about it | 20:26 |
jono | haha | 20:26 |
=== bittin_ is now known as bittina | ||
slangasek | doko: is this openjdk nss build-dep going to change the package size and impact CD sizes? | 20:28 |
doko | slangasek: no, libnss3 is already on the CD, but it's only useful if ca-certificates-java is accepted as well, and ca-certificates is synced to import the new certificates from the start | 20:31 |
robbiew | pgraner: the message is gone...but I also got "* AppArmor not available as kernel LSM. [fail]" | 20:31 |
doko | slangasek: ahh, wait ... please reject it | 20:31 |
slangasek | doko: both? or just openjdk? | 20:32 |
doko | slangasek: no openjdk-6 only, the dependency is libnss3-1d ... damn dlopened code ... | 20:33 |
slangasek | ok, done | 20:33 |
slangasek | (neither appears to be something I'm going to accept during the beta freeze, anyway) | 20:33 |
pgraner | robbiew: ok thanks | 20:34 |
smoser | slangasek, is it allowed to un-mark something for beta ? i think that bug 414997 is functioning as it should, with no chnages needed other than documentation. so i dont think it needs to be beta block, and i'd like to wait on soren before doing anything rash. | 20:41 |
ubottu | Launchpad bug 414997 in ec2-init "ec2-set-defaults should be 'run_once_per_ami'" [Medium,In progress] https://launchpad.net/bugs/414997 | 20:41 |
slangasek | smoser: that bug is /not/ marked as a beta blocker | 20:42 |
slangasek | smoser: only bugs that are targeted to the release are blockers | 20:42 |
smoser | oh. i didn't realize that. so that bug is set to 9.10 beta milestone, but it needs a karmic track that is milestoned that in order to be blocker ? | 20:43 |
slangasek | smoser: correct. so if there are other bugs that you think are marked as beta blockers, you may want to double-check that this is really the case | 20:43 |
smoser | agreed! | 20:43 |
rickspencer3-afk | slangasek, any chance you could do a review of desktopouch and couchdb? | 20:46 |
=== rickspencer3-afk is now known as rickspencer3 | ||
slangasek | rickspencer3: reviewing of what? | 20:47 |
rickspencer3 | the packaging, I suppose | 20:47 |
rickspencer3 | before it gets uploaded | 20:47 |
slangasek | rickspencer3: better to have someone else do it today | 20:47 |
rickspencer3 | slangasek, ok | 20:47 |
slangasek | rickspencer3: btw, someone needs to attend to erlang's build-dependency on wxwidgets2.8 | 20:48 |
ia | hello. could you tell me, please, where (in system) i can find grub boot picture (which can be seen after boot menu, but before gdm appears) - white ubuntu logo on black background. | 20:48 |
slangasek | rickspencer3: wxwidgets2.8 is in universe, either it needs an MIR or erlang needs to not b-d on it (I think the latter is much saner) | 20:48 |
rickspencer3 | oops | 20:49 |
kenvandine | it should definately not b-d on wxwidgets | 20:50 |
slangasek | -rw-r--r-- 1 lp_publish lp_publish 36710522 Nov 5 2008 /home/lp_archive/ubuntu/pool/universe/w/wxwidgets2.8/wxwidgets2.8_2.8.9.1.orig.tar.gz | 20:50 |
slangasek | (much, /much/ saner) | 20:50 |
ttx | slangasek: new eucalyptus just uploaded per mdz latest changes, please review and if approved, trigger a CD build with it. | 20:50 |
* ttx is calling it a day | 20:50 | |
slangasek | ttx: how urgently is that wanted? there are several other CD-affecting changes that I need to review and accept for beta | 20:50 |
slangasek | (d-i) | 20:50 |
slangasek | if it affects your testing timeline I'll do eucalyptus first, otherwise I'll process them in order of the number of CDs they affect | 20:51 |
ttx | slangasek: it would be great if I could get the new server ISO toasted for tomorrow's breakfast (European time) :) | 20:51 |
slangasek | ttx: right, that's not a challenge then :) | 20:51 |
slangasek | ogra: redboot-tools> needed for beta? | 20:52 |
ttx | slangasek: others might raise priority, though. | 20:52 |
slangasek | ogra: (not clear from the changelog, no bug references) | 20:52 |
slangasek | kenvandine: well, it builds the erlang-wx binary, which rather distinctly involves wxwidgets. I'd be fine if we dropped it, but it has reverse-deps... | 20:54 |
slangasek | (erlang-{common-test,debugger,megaco,reltool,x11}) | 20:54 |
kenvandine | i know nothing about the erlang package | 20:55 |
slangasek | but you have an opinion on whether it should b-d on wxwidgets? :) | 20:55 |
kenvandine | slangasek, looking | 20:58 |
kenvandine | slangasek, it looks like this was added in debian back in april, have we just not synced with them in a while? | 21:00 |
mathiaz | kirkland: mdz: why is postfix installed with a cluster controller? | 21:00 |
slangasek | kenvandine: the present erlang package in Ubuntu dates to Jun 18, this has probably been an outstanding problem since erlang was promoted | 21:03 |
rickspencer3 | slangasek, is that erlang-wx dependency blocking beta? | 21:03 |
mathiaz | kirkland: mdz: well I've found why - eucalyptus-cloud needs it to be able to send email when accounts are created | 21:04 |
slangasek | rickspencer3: no, but we should be making progress on it now so it can be resolved immediately post-beta | 21:04 |
rickspencer3 | slangasek, ack | 21:04 |
rickspencer3 | is there a bug #? | 21:04 |
* rickspencer3 checks | 21:04 | |
slangasek | rickspencer3: no, I haven't filed one yet | 21:05 |
rickspencer3 | ok | 21:05 |
rickspencer3 | slangasek, I'll take care of logging the bug | 21:05 |
slangasek | rickspencer3: thanks | 21:05 |
rickspencer3 | I'll assign to kenvandine, and you should see progress in the next day or so | 21:05 |
slangasek | rickspencer3: best to file the bug w/ tasks against both erlang and wxwidgets2.8 | 21:05 |
rickspencer3 | k | 21:06 |
=== Guest22498 is now known as YDdraigGoch | ||
=== bjf-afk is now known as bjf | ||
rickspencer3 | slangasek, I wasn't smart enough to figure out what wxwidgets project to assign it to ... so there's only the one task | 21:21 |
slangasek | rickspencer3: 'also affects distribution' -> wxwidgets2.8 + ubuntu | 21:22 |
slangasek | (bug #?) | 21:22 |
rickspencer3 | https://bugs.edge.launchpad.net/ubuntu/+source/erlang/+bug/438365 | 21:22 |
ubottu | Launchpad bug 438365 in erlang "erlang should not depend on wxwidgets2.8" [High,New] | 21:22 |
rickspencer3 | there's only wxwidgets2.6 (that I could find) | 21:22 |
slangasek | rickspencer3: I think you're probably searching in the wrong place, then; note that you have to use "also affects distribution", not "also affects project" | 21:23 |
rickspencer3 | aah | 21:23 |
rickspencer3 | d'oh | 21:24 |
slangasek | "project" is for things that aren't packages :) | 21:24 |
jdstrand | soren: hi! I've been meaning to tell you that I have a little bash script you can use/extend/steal/whatever. it's not the prettiest thing you'll ever see, but I found it helpful | 21:30 |
jdstrand | soren: http://bazaar.launchpad.net/~ubuntu-bugcontrol/qa-regression-testing/master/annotate/head%3A/scripts/apparmor/libvirt-apparmor.sh | 21:30 |
slacker_nl | hello | 21:45 |
slacker_nl | how can one change the something on the following page: https://wiki.ubuntu.com/UbuntuDevelopment/UsingDevelopmentReleases | 21:45 |
seb128 | slangasek, I've just uploaded a nautilus update which fix a bug in the xsplash code, it's a one word change to use the correct software name | 21:55 |
seb128 | slangasek, without that xsplash waits for a "nautilus" which never registers | 21:56 |
slangasek | seb128: ok - does that mean with this fix, we now get proper xsplash -> desktop transitions? (timeout + watch are both correct?) | 21:59 |
seb128 | slangasek, yes | 22:00 |
slangasek | rockin' | 22:00 |
seb128 | slangasek, oh question for you since you are around | 22:01 |
seb128 | slangasek, how would you recommend running a command with an another user in a postinst? | 22:01 |
slangasek | seb128: su -c? | 22:02 |
seb128 | slangasek, ok thanks | 22:05 |
slangasek | mathiaz, soren, kirkland, smoser: should linux-headers-ec2 be seeded somewhere? (components-mismatches wants to drop them to universe) | 22:07 |
slangasek | tedg: did I read correctly that you're working on bug #436181 today? | 22:13 |
ubottu | Launchpad bug 436181 in indicator-session "Indicator-applet-session crashed" [Critical,In progress] https://launchpad.net/bugs/436181 | 22:13 |
tedg | slangasek: Nah, we decided that it wasn't a bad enough user experience ;) | 22:13 |
tedg | slangasek: Yeah, I'm not sure of the root cause. But I have a branch that I can't get to crash. seb128 has volunteered to test on his setup. | 22:14 |
slangasek | ok | 22:14 |
seb128 | I'm just doing this gdm change first | 22:14 |
seb128 | slangasek, "su -c 'gconftool --get /desktop/gnome/interface/gtk_theme' gdm" | 22:14 |
seb128 | slangasek, is there anything stupid in this command? | 22:15 |
slangasek | looks sane to me | 22:15 |
seb128 | it's not supposed to write on stdout? | 22:15 |
slangasek | you may or may not need a '-i' as well to specify that it's a login shell | 22:15 |
seb128 | because it doesn't | 22:15 |
slangasek | dunno | 22:16 |
slangasek | er, not -i -- -l | 22:16 |
slangasek | but no, that doesn't help either | 22:16 |
slangasek | seb128: oh, -s /bin/sh | 22:17 |
slangasek | instead of trying to execute the command using the /bin/false shell ;) | 22:17 |
seb128 | now it makes sense ;-) | 22:17 |
seb128 | slangasek, thanks | 22:17 |
slangasek | sure | 22:17 |
=== pochu_ is now known as pochu | ||
mathiaz | slangasek: yes - probably | 22:20 |
mathiaz | slangasek: I'd suggest on the dvd seed | 22:21 |
mathiaz | slangasek: to keep it in main | 22:21 |
slangasek | mathiaz: mumble; nothing else from ec2 is on the dvd seed | 22:21 |
mathiaz | slangasek: right - the rest is in the uec seed | 22:21 |
slangasek | the dvd seed corresponds to an actual image, please don't use it "to keep [things] in main" :) | 22:22 |
mathiaz | slangasek: which is what is installed by default in uec | 22:22 |
mathiaz | slangasek: I don't think we wanna have the headers installed by default the uec image | 22:22 |
mathiaz | slangasek: ok - so where else should it go? | 22:22 |
slangasek | that's fine - in that case, the supported seeds would be the right place for it | 22:22 |
mathiaz | slangasek: ok - supported then | 22:22 |
slangasek | supported-kernel, specifically | 22:23 |
mathiaz | slangasek: should I add to this line (in supported-kernel)? http://paste.ubuntu.com/280726/ | 22:25 |
mathiaz | slangasek: I'm not that ec2 should be considered at the same level as -virtual | 22:25 |
slangasek | mathiaz: yep - I see a couple other things to clean up in there though, so I'm happy to take care of it | 22:25 |
slangasek | (e.g., "backports-jaunty" is wrong :) | 22:26 |
mathiaz | slangasek: ok - I'll let you add it during your cleanup | 22:26 |
kirkland | mathiaz: right, email needed to create accounts | 22:38 |
slangasek | siretart: any reason to regard the auctex upload as beta-critical? | 22:38 |
kirkland | slangasek: mdz: i'm syncing the iso's now | 22:38 |
kirkland | jono: some testing would be great, yeah; what else do you need from us to make that happen? | 22:39 |
kirkland | slangasek: regarding the kernel seed, "yes, probably so..." but I'm not the expert on that one | 22:39 |
kirkland | slangasek: do we have an ec2-image seed? | 22:39 |
slangasek | kirkland: there's a uec seed; mathiaz said he didn't think it was needed there; I've seeded it in supported-kernel | 22:40 |
kirkland | slangasek: hmm... maybe jjohansen1 knows? | 22:40 |
slangasek | mm? | 22:40 |
* jjohansen1 goes to read scroll back | 22:40 | |
=== jjohansen1 is now known as jjohansen | ||
slangasek | kirkland: I do agree with mathiaz that I think it's unlikely to be needed by default, so there's no need to include it in our image; so supported-kernel is the sensible place to keep it around | 22:41 |
kirkland | slangasek: coolio | 22:42 |
jono | kirkland, we would still need http://testcases.qa.ubuntu.com/System/Eucalyptus updating will a full set of tests and where to report feedback | 22:42 |
jono | I can push the message if that page is updated, but I would need that first | 22:42 |
kirkland | mdz: i'm just getting the message about notifying eucalyptus about iso's, as I've been flying all day today | 22:43 |
kirkland | mdz: i'll tell them now | 22:43 |
slangasek | smoser: what else is needed to close out bug #431103 for beta? | 22:44 |
ubottu | Launchpad bug 431103 in linux-ec2 "ssh host key fingerprint no longer available in the console log" [High,Fix released] https://launchpad.net/bugs/431103 | 22:44 |
siretart | slangasek: no, auctex is not beta-critical. post-beta it is required for emacs23 | 22:49 |
slangasek | siretart: ok | 22:51 |
slangasek | siretart: (motu-release has approved emacs23?) | 22:51 |
siretart | slangasek: yes, they did. the bug is referenced in the changelog | 22:56 |
siretart | that's bug #433397, BTW | 22:57 |
ubottu | Launchpad bug 433397 in emacs22 "FFe for emacs23" [Wishlist,Confirmed] https://launchpad.net/bugs/433397 | 22:57 |
slangasek | siretart: ok, cool | 22:57 |
siretart | TBH, I even think we should even promote it for main, and demote emacs22 to universe. I've been using pre-release versions of emacs23 for over a year, and I'm totally sold by emacs23 | 23:00 |
siretart | but that's a different story | 23:00 |
seb128_ | slangasek, I've uploaded a new gdm which set a specific theme for the login screen | 23:00 |
=== rickspencer3 is now known as rickspencer3-afk | ||
slangasek | seb128_: thanks, will review | 23:01 |
seb128_ | thanks | 23:01 |
=== seb128_ is now known as seb128 | ||
slangasek | siretart: not something I'm inclined to consider post-beta | 23:01 |
siretart | you aren't an emacs user, are you? ;-) | 23:02 |
siretart | anyway, you're probably right, because a lot of add-on package need to become 'emacs23'-enabled, like I just did for auctex | 23:03 |
slangasek | siretart: heh, my editor usage is secondary to my RM hat. >:) | 23:03 |
siretart | that's rather straight forward for most packages, but it will take some time to get it sorted out, given our bad performance at preparing the emacs23 package :-( | 23:04 |
mathiaz | kirkland: hey - going through http://testcases.qa.ubuntu.com/Install/ServerEConfig | 23:08 |
=== robbiew is now known as robbiew-afk | ||
mathiaz | kirkland: point 3 ran into an error: http://paste.ubuntu.com/280758/ | 23:08 |
mathiaz | kirkland: BTW I don't have br0 interface configured on the CC | 23:08 |
mdz | mathiaz, kirkland, how is it going? | 23:16 |
mathiaz | mdz: I've just installed the CC and the NC on two physical systems | 23:16 |
mdz | mathiaz, steps 1-7 are obsolete with the current build | 23:16 |
mdz | everything should be registered automatically | 23:17 |
kirkland | mdz: sync'd ISO's to home; i'm at the airport now, won't be home for a few hours | 23:17 |
mathiaz | mdz: is the bridge interface supposed to be setup automatically on the CC? | 23:17 |
kirkland | mdz: i just reviewed davmor's test instructions; they look good | 23:17 |
mdz | mathiaz, not afaik; I believe only the node controller gets a bridge set up | 23:17 |
kirkland | mathiaz: i don't think the CC needs a bridge | 23:17 |
mdz | kirkland: you mean the steps after #7, right? | 23:18 |
mathiaz | kirkland: mdz: oh right. | 23:18 |
mdz | actually I lie. #2 and #6 are still valid | 23:18 |
mdz | the node won't be registered automatically yet | 23:18 |
kirkland | mdz: minus the manual registration, you mean? | 23:18 |
mdz | I'll just go ahead and edit it | 23:19 |
mdz | kirkland: the SC, walrus and CC will be registered automatically now | 23:19 |
mdz | but the nodes are still manual | 23:19 |
mathiaz | mdz: step 6# gave me an error: http://paste.ubuntu.com/280769/ | 23:21 |
mdz | mathiaz, I have never done that bit before; ask #eucalyptus? | 23:22 |
mathiaz | mdz: ok | 23:22 |
=== porthose|afk is now known as porthose | ||
NCommander | Can someone confirm a bug for me? Anyone with the en_US lang, and latest karmic, please run apt-get update on the console, and see if you get , vs. . | 23:29 |
NCommander | (separating numbers that is) | 23:29 |
NCommander | er wait, nm | 23:30 |
NCommander | bah, false alarm | 23:30 |
Turl | Hi | 23:35 |
Turl | can you guys tell me why my karmic firefox says "shiretoko" in the title? it has all the firefox branding in the other places | 23:36 |
slangasek | mdz, kirkland, mathiaz: 20090928.2 is now the current server ISO, including eucalyptus ubuntu10 | 23:44 |
slangasek | TheMuso: where have things landed wrt alsa and ia32-libs for karmic? update-manager tells me it doesn't want to upgrade lib32asound2-plugins | 23:49 |
* mathiaz syncs | 23:50 | |
mdz | slangasek, thank you | 23:54 |
TheMuso | slangasek: I put libasound2-plugins back into ia32-libs and reverted the changes made to lib32asound2-plugins... | 23:54 |
slangasek | TheMuso: ok, cool - unfortunately, installing lib32asound2-plugins wants to remove ia32-libs, and nothing else wants to kick lib32asound2-plugins out for me... looks like some tweaking may still be needed here? | 23:57 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!