/srv/irclogs.ubuntu.com/2008/11/26/#ubuntu-devel.txt

shayashould bluez stuff be able to handle more than one bluetooth connection at a time?  it seems to have issue with a2dp and a pand network going at the same time00:11
calcwhy isn't tar smart enough to autodetect compression type for uncompressing files? it has -a for compressing but it doesn't work for decompressing00:13
directhexcalc, it isn't? i use "tar xvf foo.tar.gz" all the time00:14
directhexalso, *ngh* @ OOo build system00:14
calcdoesn't work for lzma00:14
calctar xvf ../ubuntu-8.10-desktop-i386.tar.lzmatar: This does not look like a tar archive00:14
calctar: Skipping to next header00:15
directhexah, then it's just old & lame00:15
calcso i end up using tar --lzma -xvf foo00:15
calcbut tar -acvf iirc works with lzma00:15
azeemmaybe it that codepatch just hasn't been tought about lzma yet00:15
azeem-it00:15
calcdirecthex: yep OOo build system is crap, can't be improved until Sun does the right thing00:15
calcazeem: ah maybe so00:16
azeemcodepath*00:16
azeemI should stop drinking00:16
azeemand/or stop listening to Prince00:16
calcazeem: heh even when sober i often drop entire words, etc ;-)00:16
directhexcalc, i think the configure script they use in ooo-build is hideously broken00:17
directhexchecking for mcs... /usr/bin/csc00:17
directhexCannot open assembly '/usr/lib/mono/1.0/mcs.exe': No such file or directory.00:17
directhexwhy is it even looking there in the GAC? i forced a given compiler for a reason!00:17
calcheh, probably00:17
calcOOo is horrid in general :\00:18
directhexswitch to abiword! imagine the space saved on the livecd!00:18
calci am somewhat surprised the mono stuff is broken like that considering Novell wrote that code (afaik)00:18
calcdirecthex: heh yea00:18
directhexcalc, they're always cocking up their configure scripts, but that's extra broken00:19
directhexcalc, if a grep doesn't show anything, then surely it must be in a ooo-build/src tarball?00:19
calcdirecthex: hmm maybe it is after all00:20
calcdirecthex: if so they probably got it merged upstream00:20
directhexgo go gadget bzgrep00:20
directhextime for anap -_-00:20
calcdirecthex: i'm sure Sun didn't write code for Mono though00:20
* calc thinks java and .net should both die, go C go C ;-)00:21
walterswriting entire apps in unmanaged code in 2008 just doesn't make sense00:22
calcjava is just beginning to become run anywhere and last i checked mono isn't very close either00:23
azeemwalters: good thing it's almost 2009!00:23
calcrun anywhere a working RE exists maybe00:23
directhexwalters, how else do we run them on our ARM systems?00:23
waltersdirecthex: what do you mean?  there are VMs for ARM00:24
directhexwalters, i was more thinking JIT footprint00:24
calcthe java vm for ppc is so bad i can't even get it to compile OOo00:24
azeemprobably mono FTBFS on arm, *duck*00:24
directhexazeem, fixed in 2.0.1-0ubuntu200:24
calcand ppc is not nearly as obscure as many other platforms00:24
directhexand the exciting new 2,0.1-1 in debian Incoming00:25
TheMusocalc: How can ppc hardware owners help fix that one up?00:25
waltersdirecthex: nothing prevents one from having an ahead-of-time cache of managed code -> native00:25
calcTheMuso: determine why the build fails in java on ppc00:25
directhexcalc, which JRE?00:26
waltersdirecthex: that's what GCJ does and I think various embedders do use it00:26
calcdirecthex: iirc gcj and openjdk both die00:26
calcdirecthex: or rather hang forever, the logs are available for review, i had to disable java support which is why it builds now, but without a lot of features on ppc00:26
directhexhuh? i can't find the source of this configure weirdness anywhere!00:26
calcdirecthex: should either be in the tarballs or in the configure or in a patch in patches/dev300 (iirc)00:27
calcdirecthex: unless it is magic :)00:27
directhexdefinitely smells like a bad check in a configure00:28
directhexehm..... this patch is sick00:30
TheMusocalc: When I'm off work in the next couple of days, I'll take a peak if you'd like.00:31
calcTheMuso: ok00:31
calcTheMuso: it might just work for you again like before, but if it does that is also a good data point00:31
walterscalc: is debian's openjdk tracking the latest icedtea?  Red Hat has an engineer working on complete zero-assembler for OpenJDK which should remove architecture dependency00:31
calcwalters: not certain, doko works on it00:31
directhexoh poot00:32
directhexthis is... vile00:32
calcdirecthex: any program that has its own version of make i think qualifies as vile ;-) which OOo does00:32
directhexcalc, srsly, whoever thought this was the right way to make a patch needs a slap00:32
calcalong with KDE, X (or used to?), etc00:32
directhexcalc, most sun apps do. ever heard of gridengine?00:32
directhex+                       AC_PATH_PROG(MCS, mcs)00:33
directhex+                       test -z "$MCS" || CSC_PATH=`dirname $MCS`00:33
directhex+               MCS="$CSC_PATH/mcs"00:33
directhexspot the problem00:33
calcoops :)00:33
calcshould be MCS="$CSC_PATH/$MCS"00:33
calci think?00:33
calcor basename $MCS00:34
=== ember_ is now known as ember
azeemdirecthex: I thought Sun bought gridengine from somewhere, though00:34
azeemmabye not00:34
directhexcalc, or just $MCS"00:35
directhexcalc, it should not override it! why would it use an overridable AC_PATH_PROG and then undo any changes you make, unless it's *mad*00:36
calcdirname $MCS would indicate that the full path is in $MCS00:36
directhexazeem, nah, not in recent memory. we've run SGE for dinkeys years00:36
directhexdonkeys00:36
directhexdinkeys too if you like00:36
calcdirecthex: yea it sounds crazy, feel free to send me a patch and i will get it into ooo-build00:37
directhexcalc, has this file changes in 3.0? i'm working against 2.4.1 from current jaunty00:37
calcdirecthex: oh which file is it, i can check00:37
TheMuso8/c00:38
directhexcalc, well i'm looking at patches/mono/mono-build.diff - but i'm assumig that's the one being applied00:38
directhexthere's also -m16 and -m1700:38
calcdirecthex: i think it is still the same (at least more or less)00:39
calcits a pretty long bit of code but i still see:00:39
calc+               if test "z$CSC_PATH" = "z"; then00:39
calc+                       AC_PATH_PROG(MCS, mcs)00:39
calc+                       test -z "$MCS" || CSC_PATH=`dirname $MCS`00:39
directhexin theory this just needs the line removing:00:39
directhex+               MCS="$CSC_PATH/mcs"00:39
directhexlet me check further down00:40
directhexnah, mono-climaker-config seems less foolish00:41
calcwell it looks like it is rewriting the path for MCS there00:42
calcso it probably needs to be MCS="$CSC_PATH/`basename $MCS`"00:42
directhexcalc, it rewrites the path for MCS to `dirname $MCS`00:42
calcunless they are redoing it entirely for crack reasons00:42
calchmm yea00:43
calcoh its using the path to setup AL properly i think00:43
directhexwhich is fine00:43
directhexi approve of setting al00:43
calcso yea MCS doesn't need to be done at all00:43
directhexwell... assuming 'al' is a 2.0 al on mono 2... i think it is...00:44
directhexal:exec /usr/bin/mono $MONO_OPTIONS /usr/lib/mono/2.0/al.exe "$@"00:45
directhexyes. so, believe it or not, OOo would break when building on mono 2.0, if we DIDN'T do this transition00:45
directhexas it uses 'al' (for compiling CLI assembly) from CLI 2.0, but the 'mcs' compiler from CLI 1.000:46
directhexyou don't need to know this stuff, mind. raof dos, but he's special00:46
raofYay!00:47
=== raof is now known as RAOF
* RAOF is special00:47
calcdirecthex: heh00:48
directhexcalc, i'm trying another test build, with that one line removed (and the diff indexes lowered a little accordingly, assuming i remember my diff-fu)00:49
directhexeven just 'debuild' takes ages with ooo, let alone the pbuild00:49
calcyep00:50
calcdirecthex: i use ccache to make it somewhat reasonable00:50
RAOFThat'd be substantially faster if you didn't clean the tree, right?00:50
calctakes it down to 1hr build after cached00:50
RAOF(And assumed that they've got their makefile dependencies correct)00:51
directhexbroken makefiles is the problem!00:53
directhex:p00:53
directhexNeed to get 0B/261MB of archives. After unpacking 916MB will be used.00:53
directhexobscene :x00:53
directhexand people call mono bloated00:53
Riddellcalc: KDE has its own make?00:54
directhexyeah. it's one of those damned apps with a "k" in the name. "make" i think!00:55
calcRiddell: isn't it called cmake or something like that?00:57
calcRiddell: i might be misremembering as i don't work with it anymore00:57
directhexdpkg-source: info: applying openoffice.org_2.4.1-11ubuntu4.diff.gz00:57
directhexhere it goes again00:57
calcRiddell: perhaps they were just threatening to write their own make/build system?01:00
directhexhere it goes. configuring01:00
directhexconfiguring some more01:00
directhexwait... it configures twice before even unpacking _src_core ?01:01
directhexo_o01:01
calcmaybe, don't remember what all it does, heh01:01
directhexmono might win on 'number of binaries from 1 source package', but OOo wins for 'batguano insane'01:01
Riddellcalc: cmake is a build system, it converts files into make, it's not kde specific.  automake does the same but is way more ugly especially on !gnu01:02
directhexi backported cmake once01:02
directhexi think wodim uses it?01:02
Riddellyes01:02
directhexmy memory rocks, or i'm really sad.01:03
calcRiddell: ah ok01:03
directhexpatched without error. configuring...01:04
directhexchecking for mcs... /usr/bin/csc01:04
directhexchecking for gmcs... /usr/bin/csc01:04
directhexchecking for mkbundle2... /usr/bin/mkbundle01:04
directhexconfigure: mono is up-to-date enough - building mono climaker01:04
directhexthank $DEITY01:04
calccool! :)01:05
directhexi'd better let this build finish as a sanity-check, but i think that was the missing link01:05
directhexa little slice of wtf from novell01:05
calcheh01:06
calci'm surprised they haven't fixed it for mono 2 already themselves01:06
directhexconsistency has been an issue with that upstream for a while01:06
directhexhad a nice one today from iulian - his app "giver" has been correctly patched for the mono transition, yet somehow is still not using the right compiler..... seems configure ignores the compiler you give it, makefiles have 'gmcs' hardcoded whatever you do01:08
calcfun01:09
=== fta_ is now known as fta
hyperairpitti: regarding bug 292256, there is an upstream bug filed already (see bfo 18461), but i can't seem to find anywhere to link it back to launchpad.net03:19
ubottuLaunchpad bug 292256 in pm-utils "backlight failed after suspend to ram on IBM X40" [Undecided,In progress] https://launchpad.net/bugs/29225603:19
kirklandbryce: sorry about not using a patch system ...  there wasn't one in place, so i just applied directly03:23
brycekirkland: yeah adding a patch system is a bit more involved ;-)03:24
brycekirkland: for reference, here's the snippets from my packaging notes on adding patch systems:  http://pastebin.ubuntu.com/77006/03:26
brycein this case I followed the notes on quilt, and it worked without issue03:26
kirklandbryce: right03:41
psusipitti: was wondering if you had any thoughts on my response to your message on dev-discuss about permissions and removable media, or if you just tuned it out due to low s/n ratio03:51
ScottKpsusi: Please note that pitti work on European time, so it's the middle of the night for him right now.03:58
ScottKwork/works03:58
=== hyperair1 is now known as hyperair
psusiScottK: ahh that's right... well, maybe he'll see it in the morning and I'll talk to him tomorrow04:07
psusiwould be nice to get rid of those annoying permission problems with removable media without telling people to use fat so as to avoid the issue.04:09
jdongoh god this is such a hack....04:16
psusilol04:16
jdongtrying to make a pseudo-magical 32-bit and 64-bit system04:16
jdongmainly through the use of schroot and bindfs (FUSE) :)04:16
psusidon't we have one of those by now?  I thought someone has been working on that since like fiesty?04:16
jdongpassing through 32-bit or 64-bit executables based on a whitelist and whether the requester is 32-bit or 64-bit04:17
psusidamn... what do you need all that for?  I thought there was a spec with a sane way to do it04:17
jdongthere is?04:17
jdongI sure haven't found it.04:18
psusisomething like having ld.so choose either a 32 or 64 bit search path depending on the executable04:18
jdongit's for things like if I want to run 32-bit firefox but have 32-bit firefox start up 64-bit evince04:18
jdongcurrently I just use dpkg selections to mirror the 32-bit and 64-bit environments to be pretty similar04:19
psusiyea, I swear someone was working on implementing a spec to just let you install normal 32 bit packages on a 64 bit system and run them04:19
jdonghmm I've heard nothing about that for at least 3 release cycles now.04:19
jdongI've seen plenty of ugly hacks like getlibs.04:19
psusiand this was like, back in feisty I remember talking to someone about it04:19
jdongwhich seeminly just hack apart the deb and sed installpaths into lib32....04:19
jdongI tried to critize it for being ridiculously hackish but ended up being flamed to death by 64-bit junkies.04:20
jdonggrumble.04:20
psusihehe04:20
psusiwell I thought the idea was to patch dpkg to fix the install paths to use the *32 versions, and handle the depends properly so it installs the 32 bit version of the depends, then patch ld.so to notice when it is being asked to load a 32 bit image and link against the *32 libs instead04:22
viviersfcjwatson, ping04:25
psusijdong: so you're debootstrapping a 32 bit chroot and instaling the 32 bit versions of packages there?04:27
psusinow who was it that was working on that before?  tolef?04:27
psusiyea, I think it was tfheen04:28
TheMusoviviersf: He is not around now. Just make your statement/ask your question and he will get back to you when he is around.04:28
jdongpsusi: yeah that's what I'm doing.04:28
psusijdong: yikes.... that sounds horrid ;)04:29
jdongpsusi: meh mostly just a waste of space and administration hassle04:29
jdongit's primarily to give skype a more 32-bit like home, and also give firefox a better chance of not crashing every other flash page.04:29
viviersfTheMuso, cool04:30
TheMusojdong: well if the 64-bit flash plugin improves, that shouldn't be a problem re firefox.04:30
jdongTheMuso: that's a big if.04:31
viviersfcjwatson, do you know why an ubuntu install with a live cd and an ubuntu install via kickstart would be different, if the kickstart does do the @ubuntu-desktop in the %packages section04:31
psusijdong: ask Mithrandir about it some time, I think he's the one who was working on that iirc and it sounded like he had a plan that wasn't a total kludge ;)04:31
jdongpsusi: hey! you haven't seen my kvm + X forwarding plan yet!04:32
jdongversion 2.0 is a lot less kludgy than version 1.0 ;-)04:32
psusijdong: x forwarding?  why should X need any special handling?  it shouldn't make any assumptions about the word size on the other end of the connection04:33
jdong32-bit apps are run in a virtual machine with a separate copy of X :)04:34
jdonga bit of session manager glue brings them together04:34
psusioh my god, why?04:36
jdongbefore I figured out schroot and dchroot existed :)04:36
* psusi recoils physically from that04:36
psusiso how is this better than having sed replace paths in the 32 bit package? ;)04:38
jdonglol it's not :)04:38
psusilol....04:38
jdongwell you can argue it's better since it's technically two unsoddomized Ubuntu installs.04:38
jdongjust bridged together via a remote-desktopping solution :)04:38
jdongbut I won't try to argue that :)04:39
psusiI'd argue that's an incestuous relationship and shouldn't be allowed ;)04:39
jdongI'd be inclined to agree :)04:39
* psusi wonders why schroot suggests lvm204:40
jdongprobably because of its COW modes?04:41
psusioh weird04:41
jdongapparently it with sbuild results in some mighty-fast builds :)04:42
psusicould be nicer than the tar approach pbuilder uses though04:42
jdongabsolutely04:43
psusifabbione: is ornellas.apanela.com/dokuwiki your wiki?  I'm trying to read an article there linked to from the ubuntu wiki and it looks like it's been spammed all to hell04:58
fabbionepsusi: nope.. not mine05:02
psusiI swear there was a multiarch spec and now I can't find it05:09
psusis/spec/blueprint05:09
fabbionei am pretty sure there is one05:09
fabbionenot sure how up-to-date it is tho05:09
fabbionebut it should be on the main wiki05:10
fabbioneit was done by Mithrandir afair05:10
fabbioneand doko05:10
fabbionebut I might be mistaken.. it's been too long since I looked at that stuff05:10
psusifabbione: that's what I thought05:10
psusiand I thought their idea did not involve a chroot05:11
psusithe only stuff I can find on the subject now is all about having a whole 32 bit system in a chroot05:11
fabbionei don't think it does involve chrooting at all05:11
fabbionethat's the whole point of multiarch :)05:12
ScottKDoesn't Debian have mutliarch?05:12
ScottKmutli/multi05:12
psusiI dunno.... I assumed that since jdong was mucking around with a way of doing it that Mithrandir never finished it05:13
fabbionenot that I know off (debian multiarch)05:14
fabbionepsusi: i can only find a few references here and there05:20
fabbionepsusi: i think Mithrandir was working with Debian for that goal..05:20
fabbionehttp://wiki.debian.org/multiarch05:21
fabbionethat's the closest thing I can find05:21
psusifabbione: that's what I thought too, but that was a long time ago... looks like it was forgotten about05:27
psusiwow, yea, according to that Mithrandir and sjr were working on that back in '04 and05:29
psusi'0505:29
psusijdong: fabbione found this, which has some good info for you in the morning: http://wiki.debian.org/multiarch05:31
fabbionepsusi: i suggest you talk to Mithrandir in a few hours :)05:39
fabbionei am really not into multiarch05:39
=== jamesh__ is now known as jamesh
dholbachgood morning06:28
dholbachcan you guys please update https://wiki.ubuntu.com/TeamReports/November2008 for your teams? thanks :)06:32
mathiazdholbach: I had a question related to the TeamReports - what's done with them?06:32
dholbachmathiaz: the most obvious thing: they're available and this time Nick Ali is going to highlight them on the fridge/planet/ubuntu weekly news06:34
dholbachmathiaz: I think it's great to show the world (and another teams) what kind of stuff you've done06:34
mathiazdholbach: good. How does that tie into the Ubuntu News suggestion from james_w ?06:34
dholbachalso does it show that the team is healthy06:34
mathiazdholbach: oh yes - I agree with you. I think it's a good thing to do.06:34
mathiazdholbach: I just found that there weren't really advertised much.06:35
dholbachmathiaz: you're right06:35
dholbachmathiaz: interesting question regarding UbuntuDevelopment/News06:35
dholbachmathiaz: I'll talk to James about it06:36
mathiazdholbach: to me it looks very similar06:36
mathiazdholbach: the same processes to gather input could be the same06:37
dholbachmathiaz: I think there's different audiences, the monthly report is also not useful if you want to give people a headsup of a transition or something06:37
dholbachmathiaz: but I agree, it'd be nice if there's not too many wiki pages and too many different processes for this06:38
dholbachbrb06:38
Hobbseedholbach: there's going to be a topic at UDS about it, FYI06:39
=== tkamppeter_ is now known as tkamppeter
Hobbseeor at least, part fo it will be allocated to that.06:45
Hobbseejames_w: I wonder if it's worth having a separate session for the news06:47
* Hobbsee is avoiding studying prolog again, incidently, but only for a bit06:48
ScottKHobbsee: I think the News is something someone ought to just DO.  I don't think it needs a bunch of meetings an paperwork.06:49
HobbseeScottK: that's a point - but how's it best done?06:51
ScottKI'd say whoever actually does some work gets to decide.06:51
pittiGood morning06:51
Hobbseehey pitti!06:52
pittiTheMuso: eww, they are using *tdb* for sample caching? sigh, what's wrong with a simple hashtable...06:52
pittiTheMuso: losing sample caching would really be bad, let's not do that06:53
pittiTheMuso: I just wonder why instead of hashtable, sqlite, or bdb, they had to pick tdb of all things...06:54
StevenKMaybe they're trying to impress Tridge06:54
pittihyperair: right, because there is no pm-utils project in LP; I'm going to create it06:55
TheMusopitti: Take it up with Lennart. :)06:55
pittihyperair: done06:56
dholbachHobbsee: great06:57
* pitti hugs Hobbsee06:57
* Hobbsee hugs pitti back :)06:58
=== Pfiffer|sleep is now known as Pfiffer
hyperairpitti: okay i'll go link the upstream bug07:05
pittihyperair: already done07:05
hyperairoh07:05
hyperairokay thanks =)07:05
pittihyperair: if you click on "also affects project" and the project doens't exist yet, you can click on "register", give it a name and short description, and then it will be there07:05
hyperairi see07:05
hyperairi never knew07:06
dholbachasac, fta: what about the thunderbird task of bug 190688? will you merge the fix in bzr and upload it to jaunty?07:24
ubottuLaunchpad bug 190688 in thunderbird "Use of explicit suffix in 'Icon' field of application launcher" [Low,In progress] https://launchpad.net/bugs/19068807:24
dholbachit's in the sponsoring queue for a while now :)07:25
tjaaltonare there any known security bugs that can't be exploited due to hardened default compiler settings in Ubuntu?07:32
keestjaalton: probably :)07:32
tjaaltonkees: hehe, you should know right?-)07:33
keestjaalton: well, I'd like to spend the time to research every one, but I haven't had time.  but basically, there have been a few lately that we've lowered in priority because they're "just" a DoS instead of a working overflow, for example.07:33
keestjaalton: are you looking for anything in particular?07:34
tjaaltonkees: if you noticed my blog post on the planet, the reason why I'm asking is that I need to know how big a risk it is to open a (linux) server with 20000 user accounts to the world07:35
tjaaltonand how to diminish it07:35
keestjaalton: I'm about 8 hours behind on planet reading (it's almost midnight for me)07:35
keestjaalton: 20k shell accounts?07:35
keestjaalton: or service accounts (imap, etc)?07:35
tjaaltonkees: yes, or maybe 18k but anyway07:36
* kees whistles07:36
tjaalton:)07:36
tkamppeterpitti, good news for you, see bug 292690, you will soon be able to print with SpliX 207:36
ubottuLaunchpad bug 292690 in cups "Garbage bitmaps printed on left margin in ubuntu testpage on A4 on Samsung printers" [Undecided,Fix released] https://launchpad.net/bugs/29269007:36
keesI would say you're brave to have 18k shell accounts on one machine, regardless of the security risks.  :P07:36
keestjaalton: well, my basic "simple" advice is to run 64bit and make sure you've got a CPU with a functioning "nx" bit.07:37
tjaaltonkees: well, we've had them for years, and on Real unices they've been local, in this case behind ldap07:37
tjaaltonkees: yep, that's the plan07:37
keeslocal users can DoS the crap out of eachother unless there are sane resource limits and quotas.07:38
keesif you're running a webserver on it, you may be in for a world of hurt.07:38
tjaaltonno, only ssh07:38
tjaaltonand the current servers do have some limits on them, so that'd be ported for sure07:39
keestjaalton: I'll never say you're 100% safe, but running Intrepid on 64bit with nx puts you in a very good place.07:40
tjaaltonkees: how about hardy then? we got some SELinux expertise, so that could be used as well07:41
keestjaalton: if you go hardy and you can tune up a workable SELinux policy, you should be quite happy too.07:42
keestjaalton: the SELinux infrastructure should work just fine in Hardy.07:42
keestjaalton: intrepid's selinux seems to have a few packaging bugs.07:43
tjaaltonkees: ok, thanks.. I think this makes me feel cozy for the time being :)07:43
keestjaalton: :) \o/07:43
tjaaltonfwiw, we have ~180 workstations that are open to the campus network07:44
tjaaltonwould be a nice chance to harden them as well07:44
loolhi there07:57
=== dholbach_ is now known as dholbach
tkamppeterpitti, in some seconds splix_2.0.0~rc2-0ubuntu6 will appear (I have uploaded it now). Please test it on your printer to see whether the garbage of bug 292690 goes away. Thanks.08:33
ubottuLaunchpad bug 292690 in cups "Garbage bitmaps printed on left margin in ubuntu testpage on A4 on Samsung printers" [Undecided,Fix released] https://launchpad.net/bugs/29269008:33
pittitkamppeter: cool, will do!08:46
tkamppeterpitti, I can also load the same fix onto intrepid-proposed, it is already prepared, I only need to upload it.08:49
pittitkamppeter: if you are unsure, I can test it first, otherwise please go ahead08:50
tkamppeterpitti, here we go, it is uploaded.08:51
tkamppeterpitti, the fixes are identical, either both work or both fail.08:52
tkamppeterpitti, splix has arrived in the -proposed queue now.09:06
viviersfMithrandir, is it possible to boot casper over the network ?09:35
Mithrandirviviersf: it doesn't know how to find the ISO over the network, iirc09:35
viviersfMithrandir, :( ok, that would be really cool tho09:40
Mithrandiryeah, somebody should implement that09:40
viviersfthen one does not need to use kickstart09:40
viviersfcos the install time would be much faster09:40
cjwatsonviviersf: those are two radically different installation methods so there's lots of room for differences. What exactly is different for you?10:00
cjwatsonyou should be able to boot casper over the network nowadays, I think, but it's a very lightly-tested installation path so it might only work in 8.10 not 8.0410:01
=== gouki_ is now known as gouki
ogratjaalton, * Make libdrm-dev depend on libdrm-intel1.11:31
ograwhere is libdrm-intel1 supposed to come from ?11:31
* ogra doesnt see it anywhere11:31
Hobbseedebian experimental, apparently11:33
ograah11:35
HobbseeSource package is drm-snapshot, but that doesn't seem to be in ubuntu either11:36
tjaaltonogra: uh, libdrm should build it :)11:36
StevenKIt didn't11:36
tjaaltonright11:36
tjaaltonsince it's not mentioned on the control file11:37
ograyeah, doesnt look like its n the binary list11:37
tjaaltonI'll fix it shortly11:38
tkamppeterpitti, any news with your printer?12:06
pittitkamppeter: not tested yet, just finished conf call and before that a presentation; I'll report back to the bug once I tested it12:06
tkamppeterpitti, can you upload the curtrent BZR state of CUPS into Debian and Jaunty and then SRU bug 271350?12:52
ubottuLaunchpad bug 271350 in cups "pdftopdf filter on PowerPC corrupts data" [Medium,Fix committed] https://launchpad.net/bugs/27135012:52
tkamppeterCurrently all PowerPC users cannot print at all with Intrepid and Jaunty.12:53
tjaaltoncould an archive admin push the new libdrm through so that xserver 1.5.3 could build?13:29
=== LucidFox_ is now known as LucidFox
=== maxb is now known as Guest74207
asacdoko: can you please bump build score for tbird 2.0.0.18 ... so we see if armel build is fixed?13:36
dokoasac: was already building13:37
asacoh13:39
ograasac, https://launchpad.net/+builds/kahikatea13:42
asacthx ... found it13:43
asacsigh ... finished "inbox" processing ... next bugmail: 8000 unread :(13:45
* asac pushes that back till tomorrow :/13:45
pittitkamppeter: can do; this morning I moved the previous cups SRU to -updates14:13
=== maxb_ is now known as maxb
=== njpatel is now known as njpatel_away
=== The_Company is now known as Company
cjwatsonpitti: is the Schedule section on http://summit.ubuntu.com/uds-jaunty/ supposed to be invisible to those without admin access?16:07
cjwatson(brought over from #ubuntu-meeting)16:07
pitticjwatson: ATM yes, according to Scott, since it's still preliminary16:07
cjwatsonpitti: is there a timeline for making it public, perhaps with a note that it's still a draft?16:08
cjwatsonpitti: it would be useful for team members to be able to see at least what's been scheduled, even if the times are still in flux16:09
pitticjwatson: I don't know whether it's possible to add such a note, but we can publish it without any problem, should we wish to do so16:09
pittiright, I agree16:09
pitticjwatson: better now?16:10
pittiI can't test it myself, since I always get auto-logged in as pitti16:11
cjwatsonme neither, asked others16:11
mvonothing new for me16:11
cjwatsonmight need to log out and back in?16:11
pittiI set it to "published", which is what Scott told me16:12
cjwatsonapparently it works with URL hacking; good enough16:13
pittihow so?16:13
=== Kmos_ is now known as Kmos
=== mdomsch is now known as mdomschbb
=== mdomschbb is now known as mdomsch
MacSlowhow can I revert/undo a bzr commit?16:36
MacSlowis this possible at all?16:36
james_wMacSlow: "bzr uncommit" if it was the last16:37
james_wif you've already pushed then you it will complain next time you try to push16:37
MacSlowjames_w, just found it thanks16:37
MacSlowalready everything fixed again *phew*16:38
tkamppeterpitti, great to haer that your printer finally works with SpliX 2.16:39
pittitkamppeter: the foomatic GDI driver always did, but now splix does again, too16:39
cjwatsonmvo: oh, chdist (in devscripts) might be able to do this16:59
=== njpatel_away is now known as njpatel
james_wmvo: something else I've been meaning to bring up in a meeting, but forgot to add to the agenda. When I was in Lex Cody asked whether it would be possible to install the apt https transport by default?17:01
mvojames_w: sure, I think we just did not do it because of the libcurl-gnutls dependency17:02
mvobut it should be no problem17:02
james_wlibcurl3-gnutls says "Task: minimal"17:03
cjwatsondoko,mvo: yeah, chdist is exactly the tool we need here17:04
cjwatsonchdist -a armel create jaunty-armel17:04
cjwatson# edit .chdist/jaunty-armel/etc/apt/sources.list17:04
cjwatsonchdist apt-get jaunty-armel update17:04
cjwatsonchdist apt-get jaunty-armel install <package>17:04
cjwatsonobviously you won't *actually* be able to install it17:04
mvocjwatson: nice17:04
cjwatsonI'm fixing initramfs-tools uninstallability on armel17:09
=== Zic_ is now known as Zic
=== beuno_ is now known as beuno
=== ssweeny_ is now known as ssweeny
psusipitti: ping18:14
pittipsusi: quick pong18:25
psusipitti: hey... was wondering if you had any thoughts on my reply to your post on devel-discuss about the permissions on external drives, or if you just tuned it out due to low s/n18:28
psusithe thread appears to have died out18:28
pittipsusi: the bindfs option sounds quite promising, but I haven't ever tried that myself18:29
psusipitti: I think we can just do it with the uid= and gid= options18:29
pittihowever, being FUSE it can potentially decrease performance quite severely?18:30
psusiyea... seems overkill for this18:30
pittipsusi: we do already for vfat and ntfs, but there aren't such options for ext3 and hfs18:30
jdongpitti: I use bindfs here to run my 32-bit chroot for my 64-bit machine, performance seems acceptable18:30
pittijdong: do you have some numbers?18:30
jdongpitti: I don't have any on hand18:31
psusipitti: right.... did you see my reply to you?  about 2 years ago I modified the arguments to the udf filesystem to make more sense and allow more flexability... maybe we just need to make ext3 and hfs behave the same way, and modify the mount options a bit, possibly adding a nice gui checkbox to toggle them18:31
jdongI even use it to pass through pbuilder and I haven't really noticed much of a performance impact18:31
jdongI guess I can try to get some solid numbers :)18:31
pittijdong: maybe with dbench or postmark or so18:32
psusijdong: did you see the message I sent you last night after you went to bed?  there was a good page of info on Mithrandir's old efforts for multiarch not using chroots and all that jazz18:32
pittidoesn't need to be rock-solid science, but it'd be intersting whether it's 1% or 50%18:32
jdongpsusi: yeah, I was reading that.18:32
pittipsusi: yes, as soon as the kernel actually supports those options for ext3 (with an "override" semantics instead of a "fallback", as in vfat), we can easily adapt userspace to cope18:32
pittianyway, I need to leave for Taekwondo18:33
pitticu tomorrow!18:33
psusio/18:33
psusilooks like I need to get kernel hacking then18:33
slangasekso after all the trouble that went into getting gnome-panel to do weather by default, why has it recently stopped working for me?  GNOME upstream changed their mind, or did the NOAA decide they couldn't handle the load from all the Ubuntu users? :)19:10
slangasekI guess the former, since it seems to correlate with the last time I added a location, and the location chooser no longer semes to integrate with airport codes19:10
ograor portland did stop having weather19:12
loolslangasek: I guess you mean in 2.25?  I saw vuntz mention some patches weren't integrated in trunk and were only in 2.24, but I thought these were only translations19:12
slangasekogra: all the cities I have listed appear to have stopped having weather19:13
loolHmm looks like the same upstream version in intrepid and jaunty19:13
ograwell, thats positive, at least no bad weather anywhere then :)19:13
slangaseklool: still in 2.24 on intrepid, not 2.2519:13
lool(no 2.25 anyway)19:14
loolslangasek: You upgraded to 1:2.24.1-0ubuntu2.1?19:14
slangasekmeh, there's been plenty of bad weather here, and now I've become accustomed to being able to check my panel for the temp :)19:14
slangaseklool: 2.24.1-0ubuntu1; where's that other version from?19:14
loolgnome-panel | 1:2.24.1-0ubuntu2.1 | intrepid-updates | source, amd64, i38619:14
slangasekoh, gnome-panel19:14
slangasekI was looking at libgweather, hang on19:14
loolCould as well have been libgweather19:15
ogra-pane or -applets :)19:15
ogra*panel19:15
slangasekyeah, that's the version of gnome-panel I have19:15
loolI get this problem too in intrepid19:18
loolWith gp 1:2.24.1-0ubuntu2.119:18
loolBut not in Debian sid, so it's not the weather service19:18
loolThere's this in my .xsession-errors:19:19
lool(gnome-panel:8726): Gdk-WARNING **: /build/buildd/gtk+2.0-2.14.4/gdk/x11/gdkdrawable-x11.c:878 drawable is not a pixmap or window19:19
loolNot sure whether it's new19:19
loolslangasek: Downgrading to intrepid's gp fixes the issue for me19:19
* lool reupgrades19:19
loolNo, probably just a delay, after an upgrade it works19:20
loolslangasek: So it was blank upon login; I had the time to downgrade and kill gnome-panel, got weather almost immediately, then upgraded again, killed gnome-panel, got weather info19:20
loolThat said, a) I have a slow dsl (really slow right now) b) I have issues with squid on boot, so the first panel start might have been unconclusive19:21
slangasekhrm, so should I try downgrading & restarting?19:21
* lool logs out19:21
loolnow confirms that the xorg bug isn't due to his .gnomerc and reboots19:22
lool(I have this odd Xorg bug where Xorg seems to come up fine, I see the mouse on the screen, but then the session never starts...)19:23
loolohhh19:24
lool+(WW) intel(0): ESR is 0x00000001, instruction error19:24
lool+(WW) intel(0): Existing errors found in hardware state.19:24
loolslangasek: So it was definitely my local squid; restarting it and gnome-panel and I have weather19:27
loolslangasek: Does killall gnome-panel help in any way?  You see any net traffic?19:27
loolslangasek: BTW I have portland configured19:27
ograso he could simply ask you then19:27
slangaseklool: well, killing gnome-panel brought up all the weather, doh19:31
slangaseklool: possible that if gnome-panel doesn't find the server on start-up, it caches the negative result?19:31
slangasek(i.e., no network when the panel started)19:32
loolslangasek: That was my case, but I don't know how often it retries getting the weather19:36
slangaseklool: somewhere on the order of 5-15 minutes, I don't remember exactly which; my session has definitely been up for much longer than that :)19:36
loolslangasek: Might be a bug that there's no retry if it can't fetch on startup19:37
loolI don't know the code at all19:37
loolYou do!  :)19:37
slangasekI don't know the code /that/ well :)19:38
slangasekbut yeah, I'll look into it 'n' stuff19:38
loolslangasek: BTW didn't reproduce that hang since I disable bluetooth after startup19:42
loolI fear it's going to be really hard to hunt19:42
slangaseklool: hmm, ok19:42
loolIt's clearly kernel stuff, and it looks so suspiciously similar to the cmpc bug19:43
slangaseklool: fwiw, Friday night I was getting reports of the same sort of hang from Thinkpad users with non-Intel wireless, too19:43
slangasekin fact, bluetooth seems to be one of the few components the systems have in common19:43
loolSo I actually have two bugs; one which is just plain hang, and another one which is a hang with the panic leds blinking19:43
loolThe former looks like it's bluetooth related19:43
slangasek(mako's work Thinkpad)19:43
loolOne thing I could try is leaving the bluetooth dongle plugged on my desktop intrepid system19:44
ogralool, the cmpc has no BT19:44
loologra: The model hanging had bt19:44
ograoh19:45
* ogra is out of cmpc dev for to long apparently19:45
slangasekjames_w`: why did you subscribe u-m-s to bug #207150?20:01
ubottuLaunchpad bug 207150 in python-central "pycentral crashed with UnboundLocalError in read_pyfiles()" [Low,Triaged] https://launchpad.net/bugs/20715020:01
slangasekjames_w`: the people who've followed up to the bug all appear to be core-dev :)20:02
mitsuhikofriend of mine just managed to trash grub by resizing partitions from inside windows (grub error 17)20:54
james_w`slangasek: I was just working through the list of bugs with patches, and it seemed like the quickest way to handle it I guess20:54
mitsuhikoare there ways / ideas how to prevent grub from totally crashing in that situation?20:54
mitsuhikofor dual boot users that's pretty annoying i could imagine20:54
slangasekasac: I thought there was a bzr branch somewhere for firefox-3.0?  was looking at bug #270477 in the sponsorship queue, and I don't see any Vcs fields on the source package but I'm hesitant to upload such a small change on its own directly20:55
ubottuLaunchpad bug 270477 in firefox-3.0 "firefox-3.0-branding has errors in it's package description" [Low,Triaged] https://launchpad.net/bugs/27047720:55
slangasekjames_w`: ah - well, I've unsubbed u-m-s now, on the grounds that there doesn't appear to be anyone needing a sponsor... perhaps ScottK or doko just need a prod, instead?20:55
james_w`likely20:56
=== james_w` is now known as james_w
asacslangasek: fix committed to 3.0 and 3.1 development branches21:03
=== mthaddon_ is now known as mthaddon
=== thunderstruck is now known as gnomefreak
ScottKjames_w or slangasek: I'm suffering from a deep lack of context here.  Need prodding on what?21:15
james_wScottK: I just commented in the bug21:15
james_wScottK: it's an old debdiff you posted that hasn't been commented on21:15
james_wbug 20715021:15
ubottuLaunchpad bug 207150 in python-central "pycentral crashed with UnboundLocalError in read_pyfiles()" [Low,Triaged] https://launchpad.net/bugs/20715021:15
ScottKjames_w: Yeah.  I wasn't core-dev at the time I submitted that.  I'd rather leave it to doko as it's a rather small thing for an upload.21:17
ScottKAnd I'm reasonably sure I don't care to be TIL for python-central.21:17
* slangasek grins21:21
ftawhat happened to the "multiarch" initiative ?22:25
Mithrandirfta: people got bored22:26
ftareally?22:26
slangasekwhere "people" == "Mithrandir"? :)22:26
slangasekI never got bored, just distracted ;)22:26
Mithrandirslangasek: well, or that22:27
ftathat would have helped me for chromium :(22:29
slangasekit would help for lots of things22:29
slangasekcurrently, the toolchain status is in a weird state of "binutils says the patch doesn't belong here and nobody has figured out where to shove it in gcc"22:30
slangasekand package management patches are all over the place22:30
ftaworth a talk at uds ?22:31
slangasekfta: I would be interested; I suspect that we've already filled our plates for jaunty, so it would probably be a blue-sky session22:33
NCommanderhola world22:36
* RainCT glances at NCommander 22:37
NCommanderdid something self-destruct while I've been offline?22:37
ftaslangasek, i'm going through the blueprints.. i would love to see multiarch tracked, so it doesn't slip through more cycles than necessary22:38
RainCTNCommander: should it? :)22:38
slangasekkirkland: I see you followed up to bug #277517 suggesting testing via ppa; I think for lpia that if the package builds with the lpia-targeted gcc then it ought to be fine, and persia represents that kvm is supported on ia64 - maybe this would be an ok change to dump straight to the archive?22:38
ubottuLaunchpad bug 277517 in kvm "Please enable lpia and ia64 builds" [Wishlist,Confirmed] https://launchpad.net/bugs/27751722:38
slangasekNCommander: nope, all the destruction has been externally imposed22:38
slangasekfta: if you have the energy to chase this up and keep it on the radar, I would definitely welcome that - what multiarch lacks most of all is someone who can act as a driver for it22:39
NCommanderslangasek, what can I do to fix it?22:40
NCommanderI can do test building of kvm on ia6422:40
slangasekNCommander: fix what?22:40
NCommandersladen, the destruction that was externally imposed22:40
NCommandere,r slangasek22:40
slangasekoh, does it need fixing?22:41
slangasekI thought it was all the good kind of destruction22:41
james_wI prefer a certain level of unfixed descruction22:41
james_wand a certain level of poor spelling22:41
slangasekanyway, one thing you can do is act on my follow-up to the linux-atm sponsorship request :)22:41
NCommanderslangasek, yes my lord and master :-)22:41
* NCommander sees if he can debootstrap a sid chroot22:41
NCommanderI'm sorta stuck in a dialup wonderland22:42
directhexOOo3 Mono transition patch written *properly* and tested22:52
directhexthat's 10 hours i won't ever get back again22:53
directhex(and that's only compile time, doesn't count failed test builds)22:53
TheMusodirecthex: ouch.22:54
directhexTheMuso, it's an I/O intensive build... on a laptop...22:55
TheMusodirecthex: Even more ouch.22:56
directhexa fast new laptop, but still only a little spinny disk22:57
directhexpig to build, tbh22:57
pochudirecthex: hey, you should try PPAs ;)23:06
pochuthey're really fast23:06
directhexpochu, i think people would swoop in and stab me through the brain if i tried repeatedly to do a 10-hour, 15-gig build like OOo23:06
directhexpochu, that and LP is lame and has no debian support23:06
pochuoh, if it's for debian...23:07
pochubut I've heard PPAs are idle most of the time, so probably noone would complain ;)23:07
NCommanderHuh?23:08
* NCommander returns23:08
rlaagerpochu: Plus, PPAs are limited to 1 GB, aren't they?23:08
pochurlaager: for source and binaries, not for space during the build23:09
directhexthe patch is un-intrusive outside my domain, so it should apply fine to 3.0.0-4ubuntu123:09
directhexbut honestly, i cannot be arsed to test it on there. 3.0.0-4 is enough23:09
* cjwatson hugs his new 'debi --upgrade'23:12
TheMuso8/c23:13
ion_What’s it do, as opposed to e.g. aptitude safe-/full-upgrade?23:13
cjwatsonion_: nothing to do with that23:13
slangasekwow, I've never seen this before23:13
ion_Heh, alright.23:13
cjwatsonhttp://bugs.debian.org/cgi-bin/bugreport.cgi?bug=50666323:13
ubottuDebian bug 506663 in devscripts "[debi] option to only upgrade already-installed packages" [Wishlist,Open]23:13
slangasek"debpkg -i"?  this is like, alternate-universe package management?23:14
cjwatsonthe redirection through debpkg is silly, but it's otherwise useful23:14
directhexThanks. Will commit to ooo-build.23:14
directhexGrüße/Regards,23:14
directhexRené23:14
directhexokay, OOo3 improves :)23:15
ScriptRipper NCommander Hi23:25
=== veloc1ty_ is now known as veloc1ty

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