* ogra_cmpc sighs about fedora00:12
=== cjwatson_ is now known as cjwatson
sorenogra_cmpc: hm?00:36
ogra_cmpcsoren, ach ... endless discussions with the fedora people about the --arch option in ltsp00:37
ogra_cmpcthey never heard about amd64 :P00:37
ogra_cmpcand somethimes the attitude between the lines "there is no linux apart from fedora" is a bit well ...00:39
ogra_cmpc... tiring ?00:40
slangasekmaddening?  exasperating?  pity-inspiring? :)00:42
sorenGah... Is that really the time?00:43
soren*headdesk* I have work tomorrow you know!00:43
slangasekno, you're reading it upside-down00:43
Hobbseegeser: ah good, i'm not going mad (w.r.t dpkg-gencontrol: failure: cannot read -)00:44
ogra_cmpcsoren, lol00:44
slangasekit's actually hh:0100:44
slangaseksorry, hh:1000:44
sorenOh... good?00:45
slangasekit's always good when it's the hour of the hardy heron00:45
=== emgent is now known as enJoy
=== enJoy is now known as emgent
Hobbseesoren: oh dear.  i told you that you broke it :)00:49
sorenHobbsee: You did.00:52
sorenYou actually did.00:52
sorenI was bound to happen sooner or later.00:52
* jdong reads scrollback00:52
jdongsounds like fun, folks00:52
jdonganyone else notice stuff FTBFSing on Hardy? (kidding :D)00:53
StevenK"I was bound to happen sooner or later." <- sounds like the sage words of someone crazy enough to merge dpkg00:53
slangasekStevenK: he is inevitable00:53
sorenThat's me. Inevitable.00:54
=== emgent is now known as enJoy
sorenThe best part about uploading dpkg? Knowing that your crack will be built *right* *now*.00:56
jdongsoren: and your crack seeds all other crack too :)00:56
sorenYeah, that's the not so best part of it.00:57
=== enJoy is now known as emgent
* soren goes to bed00:58
sorenLS:10 is way too late for me to still be up.00:58
sorenI completely stopped making sense at hE:10.00:59
=== emgent is now known as enjoy
=== greeneggsnospam is now known as jsgotangco
Hobbseesoren: do try not to break it again, OK? :)01:03
=== enjoy is now known as enJoy
Hobbseesoren: Knowing that your crack will be built *right* *now*. <-- sounds like you should apply to be a buildd admin.01:04
Hobbseethere are priorities for a reason!01:04
sorenHobbsee: It's ironic, really. I was trying to be a good boy and even added a new test case to dpkg with this upload. Did it help? Nooooo..01:04
Hobbseesoren: your upload is done and built and didn't break the world now, i tkae it?01:04
sorenI guess it needs to be published before the buildd's pick it up.01:05
sorenAfter that, it's give-back galore.01:05
* Hobbsee looks01:06
infinitysoren: No, it's built already.01:06
Hobbseeyeah, just as i'm going on VAC.  great.01:06
Chipzzsoren: http://imgs.xkcd.com/comics/insomnia.png ;)01:06
infinitysoren: I'll do mass-give-backs once it's published.01:06
Hobbseeinfinity: who really cares about ia64 anyway01:06
* cjwatson hopes he didn't just break all the seeds01:06
Hobbseecjwatson: it's the day for breakage. why not?01:07
* cjwatson HIGHLY recommends not merging the seeds any more01:07
sorencjwatson: Just fix them before FF, otherwise we'll have to live with broken seeds. :(01:07
cjwatsonunless you feel like a really fun resolution pass01:07
Hobbseecjwatson: so, uh, how does one merge seeds, or do that equivalent, now?01:07
* soren decides to stop trying to be funny and *really* goes to bed.01:07
Hobbseeor am i missing something obvious?01:07
Chipzzcjwatson: btw, are you responsible for consolekit integration?01:08
cjwatsonHobbsee: largely, the idea is that we should stop needing to, because common things are in a single common place01:08
Hobbseesoren: you'll probably break dpkg between now and then anyway, so the seed point will be moot.  *g*01:08
cjwatsonChipzz: no01:08
cjwatsonHobbsee: we're not quite at that ideal yet, but a lot closer than we were01:08
Hobbseecjwatson: oh good!  i'd wondered why they weren't previously01:08
Chipzzhrrrm then I must be mistaken; nevermind then. who is though?01:08
cjwatsonChipzz: pitti01:08
cjwatsonerr, and if you just updated your seed branches, you might have to force it a bit - I just made a mistake and uncommitted01:09
Chipzzcjwatson: is it possible that I recall you having a discussion about sudo and ck integration a couple of weeks ago?01:09
cjwatsonChipzz: yes01:10
cjwatsonChipzz: doesn't mean I'm responsible for it though, I was just playing with it and trying to improve a few things01:10
Chipzzcjwatson: ah, because that actually was a reason for concern for me01:10
Chipzzhaven't played with ck yet, so I may be totally mistaken01:11
Chipzzbut integrating ck with sudo and/or login rang a couple of very loud alarmbells here01:12
cjwatsondo you mean ssh rather than sudo?01:12
Chipzzno, sudo01:12
cjwatsonI wasn't integrating it with sudo01:12
cjwatsonall I was trying to do was make policykit applications not break hopelessly when invoked via sudo01:13
Chipzzanyway, I'll be better off voicing my concerns with pitti maybe?01:13
Chipzzbut your input would be nicetoo I guess ;)01:13
cjwatsonwell, I don't know what your concerns are01:14
cjwatsonbut it's also 1:15am01:14
Chipzzbasically my concern is that adding ck as a dependency may complicate recovering a broken system01:14
Chipzz(ie booting in rescue mode)01:14
Hobbseedeath to all evil keyboard bugs11101:14
cjwatsonChipzz: well, nobody AFAIK was talking about making sudo register a CK session01:15
=== enJoy is now known as emgent
cjwatsonChipzz: and rescue mode uses sulogin, not login, and therefore does not create a PAM session, so that would not involve CK either01:16
cjwatsonany sane implementation of CK integration lets the login manager continue if CK isn't available, anyway01:16
cjwatsonfor instance if you use the PAM module you make it optional01:16
Chipzzthat's what I figured01:17
cjwatsonI think there's an argument (though it's still debatable) that sudo should forward CK credentials, but having it pretend to be a new console session seems rather a stretch01:17
Chipzzbut I shall ask pitti about it tomorrow. imho care should be taken not to complicate such scenarios (though one might argue that when your system is broken, well, it's broken...)01:19
cjwatsonI don't really think it would01:19
Chipzzcjwatson: anyway, since this is pitti's stuff anyway, I'll bother you no more and let you get some sleep ;)01:21
cjwatsonthanks ;)01:21
Chipzzgood night (in case you're hitting the sack ;))01:22
* infinity manually shoves dpkg through the publisher...01:27
* lamont waves01:31
Hobbseehi lamont01:33
=== Pici` is now known as Pici
=== lando_ is now known as lando
=== lando_ is now known as lando
=== lando_ is now known as lando
=== asac_ is now known as asac
SeqDoes anybody know how to create their own kernel derivitive using the binary-custom folder? Mine seems to continually fail05:11
=== \sh_away is now known as \sh
=== \sh is now known as \sh_away
warp10Good morning!06:05
pwnguinim assuming the answer's no, but if I trigger a kernel panic, that stack trace wont be present anywhere on disk, right?06:22
pittiGood morning07:22
stgraberhi pitti07:23
pittigeser: ugh; due to the new dpkg, I assume? I'll have a look soon07:23
pittiLaserJock: I just accepted the packs in -proposed07:24
pittiLaserJock: today I'll care for getting dapper to feisty langpacks synced, and then do an announcement on -translators@; if all goes well, I'll move them in a week07:24
pittisoren: compat level 1 in pkg-create-dbgsym> yes, unfortunately we have packages which are *that* old and crappy :/07:26
pittisoren: that sounded like fun; thanks for fixing it, is there anything I still need to do?07:28
LaserJockpitti: k, awesome07:29
pittiChipzz: so, what exactly are your concerns with sudo and CK?07:30
Hobbseemorning pitti!07:31
Chipzzpitti: more like login and CK07:32
Chipzzhrrrm wait :P sudo too actually07:33
Chipzzmy concern is that in case something breaks, that having CK in the mix would make it harder to fix things07:34
Chipzzbasically the way it is now, recovering a broken system (or rather: getting to a state where you can attempt to recover it) is not too complex07:36
ChipzzFSVO "not too complex"07:36
Chipzzbooting with init=/bin/bash07:37
Chipzzlets say the sudo <-> CK integration breaks07:37
Chipzzatm sudo setup is not too complex, and little is required to say, reboot, log in as user, and use sudo to gain root07:38
Chipzzlittle is required -> from a technical pov07:38
Chipzzmy concern (and I don't know if it's a valid concern) is that complicating sudo (and login) config may make it more error-prone, and increase the chances of you being unable to for example gain root access at all when stuff goes haywire07:40
dholbachgood morning08:04
simiradholbach: the "good" part still is missing. But I hope you are having one. :)08:04
dholbachthanks simira08:05
* dholbach hugs simira08:05
=== \sh_away is now known as \sh
simirahi Hobbsee08:13
* simira is at school, with a terribly bad teacher.... (he is french, and not totally comfortable with Norwegian...)08:14
sorenpitti: I think it's under control. infinity said he'd be giving back all the stuff that broke because of it.08:23
pittisoren: you didn't disable NO_PKG_MANGLE yet, right?08:25
sorenpitti: Right. I'll do that later today.08:25
pittisoren: the joys of dpkg :)08:25
* pitti hugs soren08:26
pittiStevenK: oooh, promising changelog on libosso! :)08:26
slytherinWhy is latest daily alternate CD for i386 so small (491MB)?08:27
Hobbseeseeds are probably botched08:28
slytherinHobbsee: looks like that is the case. Many important packages are missing08:29
seb128hey Hobbsee pitti soren08:30
seb128hello mvo08:30
seb128soren: so the fixed dpkg has built?08:30
mvohey seb12808:30
sorenseb128: Yes, around 2AM last night.08:30
pittigood morning mvo, seb12808:30
* soren is sleepy eyed.08:30
Hobbseehey seb12808:30
Hobbseesoren: 2am is a great bedtime!08:31
* soren hugs pitti back08:31
sorenHobbsee: It would have been, yes.08:31
* mvo hugs Hobbsee08:31
mvohey pitti08:31
seb128around the time I stopped working then08:31
* Hobbsee hugs mvo. morning!08:31
* seb128 is tired still and need coffee08:31
* Hobbsee inserts the caffeine drip08:31
* mvo needs tea08:32
seb128hum, tea?08:32
seb128no, need coffee first this morning08:32
seb128ah, and builds have been retried during the night08:33
simiramvo: hmm... good suggestion08:34
slytherinHobbsee: by the way, it is not i386 problem. All the CDs are very small08:34
Hobbseeslytherin: it wouldnt' surprise me.  cjwatson should know why08:35
gaspapitti: i saw usplash... hm. there really was some dumb mistakes...08:49
StevenKpitti: Do you think my libosso hack is too gruesome?08:51
pittigaspa: working fine now :) ; however, it's still not quite what I need, so I added another command (INPUTCHAR); will upload today08:52
pittiStevenK: haven't looked at it yet (will do in a few), but it sounds like a great hack :)08:52
pitti(in the *sniff* "good crack" sense)08:53
gaspapitti: ok... so you're not angry with me, isn't it ? :D :D08:53
pittigaspa: no, why should I? :)08:53
sorenIt's *Tuesday*?!?08:53
gaspapitti: there's something other that should be done in usplash?08:53
pittisoren: does that come as a big surprise for you?08:53
=== DrPepperKid is now known as MacSlow
sorenpitti: Yes!08:54
pittigaspa: there are tons of bug reports...08:54
gaspa( someone's going to fosdem, this month? )08:54
sorenpitti: This is awesome! I just got an extra day until FF!08:54
pittigood morning macd08:54
pittigood morning MacSlow08:54
pittigood morning Pi"I'm a lazy tab key user"tti08:54
gaspapitti: yes, but much of them seems arch-dependent...08:54
MacSlowhi pitti08:54
MacSlowodd... I thought I shut down xchat yesterday08:55
gaspaso i wasn't able to reproduce them.08:55
sorenMacSlow: Shut down xchat? But that would log you off irc, wouldn't it?08:55
pittiznc FTW!08:55
MacSlowsoren, yes... in the evening before going to bed this makes sense :)08:56
pittisoren: apt-cache show znc (I have used this for some weeks now, it works great)08:56
sorenMacSlow: But, but... You'd be logged off irc!!!11!!!one!08:57
pittiseb128: I'm about to look at MacSlow's sponsor request, or did you already?08:57
seb128pitti: feel free, there is new upstream versions08:57
seb128pitti: so if you want to do the update in the same time you are welcome08:58
Mithrandirsoren: you'd take IRC through a drip feed directly into your brain if you could. :-P08:58
seb128pitti: the libwnck one should be easy08:58
pittisoren: still better than wasting power over the night :) (but with a proxy you can have both)08:58
pittiseb128: erm, I was just going to upload them08:58
pittiseb128: *sigh*, ok, will do08:58
seb128pitti: ok, do it, I'll pick those when I do the update08:58
seb128pitti: no, don't bother, I'm fast at doing those updates ;-)08:58
sorenpitti: You turn off your computer too?!08:59
sorenGood thing I'm sitting down already..08:59
pittisoren: sure, whenever I don't need it for at least an hour08:59
pittiMacSlow: just a nitpick for the future: can you please use patch tags (https://wiki.ubuntu.com/UbuntuDevelopment/PatchTaggingGuidelines) for future patches, to record upstream bugs and descriptions, etc.?09:01
pittiMacSlow: and please put the source.changes there, too (rebuilding locally now, so don't worry for now)09:03
MacSlowpitti, oh... sorry didn't knew about the comments09:04
MacSlowpitti, I only knew about stating the LP-bug entry in the changelog09:04
seb128hate cupsys and apparmor09:05
pittiMacSlow: no need to be sorry, it's a relatively new thing09:05
pittiMacSlow: that's why I'm pointing it out, we want to try to establish it a bit09:05
pittiseb128: --verbose?09:05
seb128pitti: still the broken update for a week, I didn't take time to investigate but it's getting annoying09:05
seb128Loading AppArmor module: Failed.09:05
seb128invoke-rc.d: initscript apparmor, action "force-reload" failed.09:05
seb128so cupsys can't be configured09:06
MacSlowpitti, so the suggested comments in the patch itself is newer than the (LP: #12346)-thing?09:06
pittiugh, I thought apparmor was working again on current kernels?09:06
pittiMacSlow: yes, much newer09:06
pittiMacSlow: in the desktop team we want to add some metadata to patches to record description, whether it's ubuntu specific, and various bug tracker links09:07
pittiMacSlow: g-c-c> any particular reason why you hcanged the function names?09:08
pittiMacSlow: wouldn't it be enough to just change the called programs?09:09
MacSlowpitti, I like consistency and predictability09:11
MacSlowpitti, and since it's not a public API-call I felt save changing it09:12
MacSlowpitti, thanks for the uploads!09:13
gesergood morning09:28
loolpersia: I tried to clarify the "situation" with UME-handled packages in #188130; I'm around to chat about them if you need further clarifications09:34
pittiMacSlow: you're welcome *hug*09:34
pittihi geser09:34
pittiStevenK: hah, nice patch!09:35
pittiStevenK: I take it you verified that using the same pointer doesn't lead to double-free() or other crashes when the session terminates?09:36
pittiStevenK: I wait with the promotion until it gets built everywhere09:37
geserHi pitti09:46
seb128StevenK: could you look at the gimp update sponsor request? getting the new version before the freeze would be nice09:53
StevenKpitti: I've found where it disconnects. It doesn't free them, it calls dbus_bus_release_name() -- as long as that copes, I think it should be fine.09:53
StevenKseb128: Yeah, in a little while.09:53
* StevenK runs off to buy part of dinner.09:53
simiracan someone fix the locobot?10:06
cjwatsonslytherin: ok, I'm sure it's my fault, I'll look at it shortly10:06
Hobbseesoren: no, it's really monday.10:14
HobbseeMithrandir: do they do those drips now?  that'd be nice!10:15
simiraHobbsee: he's in transit I believe10:16
Hobbseeyay for backscroll, and irc proxies10:17
simirahe's on in a few mins10:17
Hobbseejust demand he comes on now :P10:17
simiraI don't need him, I'm in class :)10:18
Hobbseesimira: i thought you were supposed to listen in class?10:20
Hobbseeand not be on irc?  :)10:20
Hobbseethat being said, i've been on irc during uni too.10:20
simiraHobbsee: yes, when my french lecturer learns Norwegian properly, maybe...10:20
Hobbseeoh, this is *still* the french lecturer?10:21
Hobbseesheesh, how long is he giving a lecture for?10:21
simirayes, we mostly got only him :p10:21
simirafour hours today10:21
simirafive yesterday, and three tomorrow, I believe10:21
Hobbseewhat's he attempting to teach?10:22
simiraa norwegian standardization system for archive management for official purposes10:22
simirapretty much like ISO 15489, just made and adjusted to Norwegian standards10:24
simiralucky for me, it's very easy to understand from an IT/developer perspective10:25
* mpt orders a copy of ISO 15489 to keep on his bedside table10:25
simirampt: you'd need a bad French lecturer to have any use of it. The standardization itself is somewhat interesting10:26
simira(if you are interested in archive managment, that is)10:26
pittiyay, seems I sufficiently convinced usplash to do what I want it to do \o/10:28
pittisane fsck integration into usplash!10:28
TheMusopitti: DO those recent usplash changes affect themes at all?10:28
pittiwith pressing Esc to skip routine checks10:28
TheMusopitti: ooooo shiny!10:28
mptpitti, cool10:29
pittiTheMuso: no, they don't; I just fixed the input handling10:29
TheMusopitti: Right.10:29
pittiI currently hijack the progress bar to show the fsck progress10:29
pittiwhich means that it'll jump around a little10:29
pitti(i. e. used for init script progress, then fsck progress, then again init script progress10:29
pittiI'd like to use a different colour for fsck, but usplash currently doesn't allow that10:30
TheMusoBetter than nothing.10:30
pittiit's probably good enough for a first upload10:30
pittimpt: WDYT? should I use the progress bar or output some text10:30
cjwatsonpitti: ooh, congratulations10:30
pitti(like percentage)10:30
cjwatsonpitti: which fsck backends does it support?10:30
pitticjwatson: only ext3 actually provides progress reading10:30
mptpitti, how soon in the progress bar do/can you find out that you need to do fsck?10:30
pittiI'll check it in a bit how it looks with reiser and xfs10:31
pittishould be fairly early10:31
pittiI didn't run a complete boot yet10:31
cjwatsonbut only when fsck actually starts10:31
mptpitti, I mean, do you know that you'll have to do a fsck before the progress bar begins?10:31
pittiI just start usplash and checkfs. manually10:31
pittiright, what cjwatson said10:32
pittimpt: let me test this with a real boot and come back to describe how it looks like in the entire boot sequence10:32
slytherincjwatson: thanks for looking into it. :-)10:35
gaspapitti: next step is a kernel ooops... i want to see it in a usplash screen... :D :D10:39
MithrandirHobbsee: I'm sure you could get one if you paid enough..10:39
pittimpt: so, it actually looks a bit ugly10:40
mptpitti, my usual suggestion is to retain one progress bar for overall progress of the task (in this case, starting up), and using text for subtasks10:40
pittiinit script progress starts from 0 to 30%, then checkfs kicks in and does 0 to 100, then init script continues from 30 to 10010:40
mptSo it's ok if the progress bar gets stuck for a few minutes, as long as there's a line of text changing regularly underneath10:40
pittimpt: ok, I I should rather output the percentage?10:40
mptwhere "regularly" > 1/second10:40
mptpitti, is there anything more fine-grained you can report than the percentage?10:41
Hobbseesimira: fun10:41
mptFor a large disk, a single percentage could take many seconds10:41
pittimpt: fsck has 5 stages, and reports a percentage for each of the stages10:41
pittibut they are fairly meaningless10:41
mptIs there a MB measurement, for example?10:41
mptor disk blocks, or something?10:42
pittiwell, blocks maybe10:42
pittiit outputs some numbers 'cur' and 'max'10:42
HobbseeMithrandir: heh10:42
mptpitti, do they represent a fraction?10:42
mpti.e. progress = cur/max?10:42
pittiI can output the numbers (stage X/5, cur, max) directly instead of percentages10:42
pittimpt: yes10:42
pittimpt: that's the percentage withing a stage10:42
mptand max > 100?10:42
pittifor a small disk, max < 10010:43
pittifor a large one I suppose it's much bigger10:43
pittiI have a 6 GB test partition, where it's 9410:43
mptHow long does it take overall?10:43
mptactually, sorry, wrong question10:43
mptHow long does the slowest stage take?10:44
mptfor that 6 GB10:44
pittioh, stage 1: 46 (70% of time), stage 2: 473 (20% of time), stages 3 to 5 are very quick10:44
mpt473 seconds?10:44
pittino, that's the 'max' number reported10:44
pittian absolute number of files or blocks to check, or so10:44
mpta different number for each stage10:44
pittiI have a conversion function which accumulates stage, cur, and max to a single percentage10:45
pitti(adapted from fsck.ext3)10:45
mptThe reason I'm asking this, is that it's good for the text to update at least every couple of seconds10:45
mptso that it never looks frozen10:45
mptAs a (very) last resort we could include the time elapsed, I'm just trying to work out whether that's necessary10:45
pittimpt: what about including all?10:46
Mithrandirmpt: you don't have any meaningful output that changes every couple of seconds by default, at least.10:46
pittiChecking disc..... 23% (stage 1, 234/437)10:46
Mithrandirwell, maybe on a 6G disk, but not on a 1TB disk or thereabouts.10:46
mptMithrandir, "by default" as in the existing text display?10:46
Mithrandirmpt: yes.  You have a text throbber and a progress bar.10:46
mptyes, I'm familiar with it10:47
mptbut not familiar with how bad it gets on huge disks10:47
pitti(I currently mimic that text mode progress bar behaviour in usplash)10:47
Mithrandirit just throbs slower.10:47
pittiI think the numbers (cur/max) will get higher, so I could output those10:47
pittifsck seems to update progress several times a second10:48
mptWhat makes it spin the throbber?10:48
pittimpt: the text mode has a certain threshold for the percentage10:48
mptI mean, what has it finished when it rotates one segment10:48
pittie. g. if one "=" represents 2.3%, it updates the progress bar every 2.3%10:48
pittioh, it has that rotating thingy, too10:49
Mithrandirpitti: "throbber". :-)10:49
pittiMithrandir: I think I just learned a new word :)10:50
mptSo, if we can assume that (a) fsck will usually take more than a couple of minutes, and (b) for most disks max > 100, then I think it's better to show cur & max10:50
pittimpt: maybe the percentage in addition, to give a feeling for an ETA?10:50
mptthrobber, I think comes from Netscape 1.1, where it actually throbbed10:50
pittisince stage 1 takes 70% and stages 2 to 5 together 30%, it would look less intimidating10:50
mptpitti, you're right10:51
pittimpt: WDYT about "23% (stage 1/5, block 234/437)"10:51
mptYour suggestion above is perfect10:51
mptexcept three dots, not five :-)10:51
Mithrandirpitti: where do you get the cur and max numbers from?10:51
pittisure :)10:51
pittiMithrandir: fsck -C10:52
pittiMithrandir: rather, -C310:52
pitti(output progress to fd 3)10:52
pitti-C == -C0 is magic, it uses the text mode progress bar10:52
pittibut for an fd you get three numbers (stage cur max) per line10:52
pittione line for each cur10:52
Mithrandirmax for a 2.6T volume is ~21k, so a percentage would be good10:53
persialool: Thanks for the clarification on MIC.  Sounds like a bit of a mess.  I'd like to wait to hear back from smagaoun about it, but will push pre-FF if nobody else hits it.11:02
Hobbseepersia: see my comment in -motu11:03
persiaLonger term, I'd think setting maintainer to UME, and granting UME access to the VCS might be a sensible solution, to avoid tracking multiple debian/ directories.  Alternately, one of the two repos can be the master, and the other can merge/sync as Ubuntu does with Debian.11:04
cjwatsonslytherin: OK, I see the problem - the code that generates the "master task" for cdimage doesn't cope with following more than one level of seed dependencies11:04
persiaHobbsee: I refuse to accept anything as "crack" that can be fixed.  Mortar is cheap :p11:04
cjwatsonslytherin: the upshot being that required, minimal, standard, and desktop-common go missing11:04
slytherincjwatson: That is almost everything one would need to use Ubuntu. ;-)11:07
slytherindesktop I mean.11:07
loolpersia: Thanks11:07
persialool: On an extended note, aside from other things, doesn't it make sense for someone in UME to be sponsoring UME stuff, rather that UUS?  My comfort level for uploads when wearing a UUS hat goes down significantly when the resulting maintainer is not MOTU.11:09
loolpersia: There are little uploaders for UME at the moment; most people aren't MOTU and we only have a couple of core dev while the packages are being promoted to main11:11
persialool: Makes sense.  I remember when myth was that way.  UUS works until it gets resolved (although it may soon require UMS)11:12
loolpersia: While folks started the MOTU process and will probably be core dev in the next weeks or months when they will have demoed some experience, I fear it would be too hard to sponsor everything internally ATM11:12
Hobbseepersia: sure, but sanity is not.11:15
TheMusoHrm. Is it known that one doesn't get a /dev/cdrom symlink in /dev?11:51
TheMusoRunning latest updates, from my local mirror.11:52
sucotronichello everybody11:54
sucotronicone question11:54
sucotronicHow often the maintainers checks the original projects for new releases?11:54
cjwatsonsucotronic: "from time to time"11:55
cjwatsonsome maintainers are subscribed to announcement lists and see them immediately; some just check every so often; some only update when prodded11:55
sucotronicthere isn't any mechanism to notify maintainers?11:56
sucotronicor is more easy to become a maintainer?11:58
cjwatsonsucotronic: subscribing to upstream announcement lists is the preferred mechanism for diligent maintainers, though they can also create a debian/watch file and use uscan --report --verbose regularly11:59
sucotronicthen, I've to contact with the maintainer to ask him?12:00
cjwatsonsucotronic: filing a bug would be the usual method12:08
sucotroniccjwatson: sorry, I'm new. How I can do it?12:10
cjwatsonsucotronic: https://bugs.launchpad.net/ubuntu/+filebug; also read https://wiki.ubuntu.com/ContributeToUbuntu12:12
cjwatsonsucotronic: please ask further support questions in #ubuntu12:12
sucotroniccjwatson: thank you a lot12:14
Mezhmm, does anyone know anyone here who works for yahoo ?12:14
sucotroniccjwatson: I think you are wrong. I don't want to report a bug. I'm a maintanier of a project, and I only want to know how to notificate the ubuntu maintainer the new releases12:17
persiasucotronic: A bug of the form "Please upgrade to new upstream release X.Y" is the preferred format if you want to push changes, rather than waiting for a pull.12:18
=== ember_ is now known as ember
sucotronicpersia: ok, that's what I want to know. Thanks12:20
DarkSun88Hi all12:28
cjwatsonslytherin: Ubuntu daily CD builds fixed now12:33
slytherincjwatson: Thanks. Does that mean that new images will be generated again?12:33
cjwatsonslytherin: I just generated Ubuntu ones12:36
cjwatsonand am building Kubuntu now12:36
slytherincjwatson: kudos to you. :-)12:37
slytherinAny of the buildd admins present here? I have a debconf preseed request.12:37
cjwatsonI thought that particular preseed had been done12:37
slytherincjwatson: which one? I am referring to batik build, it uses older j2sdk because Sun JDK is too new for it.12:38
cjwatsonslytherin: the RT ticket was marked resolved on 12 Dec; I'll see if I can dig up exactly what was preseeded12:40
slytherincjwatson: AFAIK, only Sun java packages have been preseeded.12:41
gesercjwatson: sun-java-* works on the buildds after the preseeding was done (at least in hardy)12:41
slytherinand what we need is preseeding for j2sdk/j2re package (Blackdown JDK/JRE). This is needed for some packages.12:43
cjwatsonshared/accepted-sun-dlj-v1-1 is preseeded, but nothing else seems to be12:47
cjwatsonslytherin: mail rt@admin.canonical.com with your request12:47
pitticjwatson: do you know if there will be any problems if I add a "Breaks: usplash (<< 0.5.12)" to initscripts? (which is a required package)12:48
pitticjwatson: I need the new usplash features for the fsck integration12:48
cjwatsonpitti: don't think so12:48
pittiok, thanks12:49
slytherincjwatson: Difficult task if you ask me. Access to my primary mail account (gmail) is blocked for me. I will see what I can do. :-)12:49
cjwatsonslytherin: (or ask somebody else to do so) and say exactly which debconf question you need preseeded12:50
slytherincjwatson: How do I identify debconf question?12:50
cjwatsonslytherin: you'll need to read the maintainer scripts and figure out what they're doing12:54
slytherincjwatson: Looks like there are two different debconf questions, one for sdk and one for runtime. 'j2re1.4/license: true' and 'j2sdk1.4/license: true'13:03
StevenKpitti: libosso built everywhere - do you need to wait for the publisher to finish, or can you promote it while it's running?13:07
pittiStevenK: if it's uploaded, I can promote it13:11
Hobbseepitti: bad idea.13:11
Hobbseepitti: you need to wait for it to finish building, else it gets a "failed to upload" status.  iz YALPB.13:12
MithrandirHobbsee: but it's shiny!13:12
* Hobbsee takes the shiny away from Mithrandir13:12
pittiHobbsee: "if it's uploaded..."13:12
Hobbseepitti: i thought you meant source uploaded13:12
pittino, I meant the binary builds were uploaded13:12
Hobbseeoh right.   go ahead, then :)13:13
* Hobbsee hugs Mithrandir13:13
pittianyway, I'm finally off for lunch for a bit; I'll review/promote it once I'm back13:13
* Hobbsee blinks13:14
Hobbseeso, uh...13:15
Hobbseeeither iv'e forgotten how to use powerpoints, or thsi cable has just died on me.13:15
Hobbseeor only works sometimes13:20
MithrandirHobbsee: it's probably to wear down your nerves.13:22
HobbseeMithrandir: it was working fine prior to this.  but now that i'm going on holidays, where it'll be my only phone, it does this.  grrr.13:23
=== cr3_ is now known as cr3
saispoBenC: ping ?14:19
BenCsaispo: ?14:20
saispoBenC: hi :) i have a little question about git kernel ubuntu, i see on list *changes that kernel 2.6.20 and 2.6.22 are fixed but nothing in gitweb, it's normal ?14:21
BenCsaispo: what do you mean "fixed"?14:22
saispoabout CVE-2008-0600.14:22
ubotu** RESERVED ** This candidate has been reserved by an organization or individual that will use it when announcing a new security problem.  When the candidate has been publicized, the details for this candidate will be provided. (http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2008-0600)14:22
* pitti promotes libosso and osso-gwconnect to main and waves to StevenK14:25
RiddellDarkSun88: are you going to rebulid all the libglew rdepends?14:34
DarkSun88Yes. Are there problems?14:34
RiddellDarkSun88: no, just confirming it's in your sights, I've let it through New14:34
RiddellDarkSun88: also I've uploaded a new koffice2 so you can ignore krita-kde414:35
DarkSun88Well, thanks. :)14:35
saispoBenC: changes haven't been commited in git ?14:38
BenCsaispo: I'm sure they have, but maybe they weren't pushed...I can check14:40
saispook, thanks :) because i must rebuild them14:40
saispoand i use git for doing this14:40
saispoBenC: thanks, good for gutsy :)14:46
saispowaiting feisty :)14:47
dramenhow can i install a manpage when i build a (debian)package by using dpkg-buildpackage14:59
ion_See dh_installman(1)15:00
persiadramen: You likely want to ask that question in #ubuntu-motu, and read the dh_installman manual page15:00
dramenin debian/rules there is a call dh_installman15:03
dramennormally, it should install the manual page automatically on the right location15:04
=== hunger_t_ is now known as hunger
cjwatsonas dh_installman's own manual page says, you have to tell it which manual pages to install, either with debian/manpages (or debian/PACKAGENAME.manpages) or with command-line arguments15:06
dramenok, i just put the deprecated dh_installmanpages into my debian/rules-file and now it works15:16
dramenalthough, the output is as followed: "dh_installmanpages: This program is deprecated, switch to dh_installman."15:17
pitticjwatson: ooh! ubuntu-meta with reorganized seeds?15:19
cjwatsondramen: dh_installmanpages has a different interface, which includes automatically searching for manual pages. In practice this turned out to be more trouble than it was worth.15:25
cjwatsonpitti: yes15:25
=== mathiaz_ is now known as mathiaz
seb128ogra: will you look at ted's gnome-screensaver updates, the bug is assigned to you, or should I do it?15:39
seb128soren: could you look at the gtk-vnc update sponsor request? it should be easy and required by the new GNOME15:41
sorenseb128: It's on my list.15:42
loolHmm today's Tech Board meeting doesn't appear on the fridge which is the official place to list tech board meetings (says the wiki); who should I ping to get this fixed?  Tech board folks to ask fridge people or fridge people directly?15:43
xivulonhi heno15:44
persialool: Mail to fridge-devel@15:44
henohey xivulon!15:45
persiaMore generally, TB should send such a mail when a meeting date is decided.15:45
henoxivulon: that last version looks nice :)15:45
Keybuklool: fridge *still* can't do recurring meetings15:46
Keybukso they have to add them one by one by hand15:46
saisposeb128: \o/15:46
seb128saispo: what?15:46
* lool mailed fridge-devel15:46
Keybukand occasionally we hit the horizon of Corey's boredom of typing the same meeting in15:46
saisposeb128: nothing, just say "hello" :)15:47
seb128hey saispo ;-)15:47
xivulonheno: glad to hear that15:48
xivulonheno re skinning, I should be made the labels and buttons solid color and/or transparent15:49
xivulonat the moment the image is 314x16415:49
xivulonassumes white background15:49
henoxivulon: can the buttons be made transparent?15:49
xivulonheno: never tried that15:50
ograseb128, hmm, i'm pretty packed atm if you have a spare cycle for that it would be great15:50
xivulonshould be possible though15:50
seb128ogra: ok, will look at it, thanks15:50
henoxivulon: ok, grey will be fine as well15:50
ograthanks a lot15:51
henohaving a pixmap on the entire background would be great15:51
xivulonthat should be possible too, but haven't played with that either, requires overriding default nsis behaviour (164x314) expected15:52
xivulonthat said when the image goes below text readibility is affected15:52
xivulonI think that a 3Dish logo with shadow on white background might be good enough15:53
henoI agree15:53
xivulonif they can keep it within the above size much easier for me15:54
henoIt's just useful to know what the technical constraints are when talking to the art team15:54
xivulonthat can be branded (at compile time though)15:54
henolet's start with that then15:54
xivulonI mean I can always override the nsis behaviour here and there, but I have the schedule quite busy before friday already15:55
xivulontranslation I think is a more urgent issue15:55
xivulonI can reuse wubi scripts (po <-> nsh) or win32loader approach (c dll + gettext)15:56
xivulonthe transparency by the way depends on mfc api, if it is supported by standard flags or via sendmessage then it can be done15:59
henoxivulon: NSIS gets the locale from Windows?15:59
xivulonif not it requires a separate dll which is not worth it IMO15:59
xivulonuses the default windows language16:00
xivulonbut at the moment I only have english in there anyway16:00
henoI think the text is ok for translation now (though I'd like to see even less of it :) )16:01
xivulonif there are mfc gui experts here, tips are welcome :P16:01
xivuloneasiest way for me is if you change the text in the wiki16:01
xivulonI'll grab it from there16:01
henoOK, I'll try condensing it some more if possible16:02
xivulonI'll upload the code tonight, building it should be quite straightforward, you need wine+nsis2.3416:03
xivulonheno: quick googling shows that transparent buttons are not obvious, transparent labels should be ok16:09
henothat would be great16:10
xivuloneh did not look at what is required to enlarge the image though16:10
xivulonheno: http://img444.imageshack.us/img444/9167/umenu1wy6.jpg16:21
xivulonthis is to show transparency capabilities16:21
xivulonI think it looks better on the side though16:22
henoxivulon: right, but we wouldn't use that image in that case. most likely the image would have the logo on the left and a faint texture on the right16:25
henoxivulon: what's the image dimension on that one?16:25
xivulonI appreciate that, note that the same image would probably also go in the reboot page16:25
xivulonheno that is a bit tricky, since the size changes slightly with font sizes (the image can be stretched or we can end off into a background color)16:34
\shsoren, ping dpkg-dev...as I don't like to fight with dpkg, would you like to bring the "Description" tag back to _source.changes files, as described here: http://git.debian.org/?p=dpkg/dpkg.git;a=blobdiff;f=scripts/dpkg-genchanges.pl;h=0fcd529b847c5f51f7f4c5e3ea8c51e186d3730d;hp=7dada21a531c568298764c76dbe9a44b56471101;hb=15fa75bd7143d6ec54f0a77c482ff0cfb72bf440;hpb=1f6feb233d4e7088cb920f386292f8d31d694d3a ?16:36
soren\sh: File a bug.16:36
\shsoren, right now, this field is policy and MIA in last dpkg from debian/ubuntu :)16:36
\shsoren, ah well, I'll provide a debdiff16:37
sorenCool, thanks.16:37
\shsoren, np16:39
=== \sh is now known as \sh_away
xivulonheno the nsis recommended size for a vertical image is 164x314, for a full image 496x31416:48
xivulonheno: checkboxes are like buttons, so cannot be transparent16:56
xivulonthat basically rules out a full size image16:57
henoxivulon: ok (you mean for the radio buttons on the reboot page?)16:58
xivulonthey will have a solid background16:58
henobtw, is there any limitation on colour depth?16:59
xivulonIt has to be a bitmap not sure what depth is supported, but I that depends on windows defaults16:59
xivulonheno: http://img166.imageshack.us/img166/1228/umenu2zo3.jpg17:02
xivulonIMO Simplest option is to have a vertical image (164x314) with 3Dish logo + shadow, all with a white background.17:02
henoxivulon: agree. I'll look at making one (and ask for some art help if I fail)17:04
henothanks for investigating17:04
neighborleeso is hardy close'ish to average user ready , from what I read about current status it would seem so  :)17:42
jpatrickneighborlee: feature freeze soon17:43
neighborleeohhhhh btw17:43
neighborleewhat was the decision on mono17:43
neighborleejpatrick, sounds nice :)17:43
persianeighborlee: The (nearly) final set of upstreams should be determined this week, but there are still lots of bugs that need closing.  Testing appreciated, but know that you're testing.17:43
neighborleepersia, yes17:43
neighborleepersia, current issues seemed doable ;)17:44
neighborleepersia, what is the status of mono in hardy or is it already gone ;)17:45
neighborlee I saw wiki and  was wondering17:45
persianeighborlee: You'll get a better answer from the daily CDs or the metapackages than from me (I don't know, except that there are at least mono packages in universe)17:46
neighborleeyes that wont change from what I read17:46
neighborleejust  wont be installed out of box or on install CD17:46
neighborleepersia, daily cd ??17:46
LaserJockI can't imagine mono being gone, we have f-spot and tomboy using it17:47
neighborleepersia, surely someone knows17:47
neighborleeLaserJock, you dont know about the  wiki do you ;)17:47
persianeighborlee: Check the hardy ubuntu-desktop package17:47
LaserJockneighborlee: I know there is a lot of stuff on the wiki, much of it has no baring on what's actually done17:47
neighborleepersia, ok fine I was hoping someone would just know what status was ;))17:47
neighborleeLaserJock, ok, odd.17:47
LaserJock*bearing, methinks17:47
LaserJockneighborlee: why is that odd? The wiki is open to editing by anyone17:48
neighborleewell the idea is the remove mono entirely, but leave on repo at least17:48
neighborleeLaserJock, I assumed it was a serious issue r aised by a developer, but maybe not.17:48
neighborleeLaserJock, it looked very serious to me.17:48
mjg59Mono's not being removed17:48
neighborleemjg59, ok ..as I say the wiki looked done very professsionally so I had zero reason to think anything less17:48
LaserJockneighborlee: well, I'm sure the person who wrote it  was serious, but if it's the wiki page I'm thinking of it was not by a developer17:49
mjg59Someone's written a spec about it. The spec hasn't been approved.17:49
neighborleeic alright17:49
mjg59You can see that by following the link from the spec to the status page on Launchpad17:49
neighborleewell it did makes alot of sense though if you think about it17:50
neighborleeMUCH less MB footprint  ,  memory too ?..and there are valid  replacements to all apps according to the wiki author anyway17:50
neighborleeanyway I was curious, as on the outset it sounded very logical.17:50
neighborleeand now with winforms being deprecated for avalon it does beg the question possibly ;)17:51
mjg59There's no decent replacement for Tomboy17:51
neighborleeyes actually there is, at least according to the wiki author17:52
neighborleeoffhand I dont recall.17:52
mjg59He's wrong17:52
neighborleebut I think there was some feature the repalcement didn't have.17:52
neighborleebut I now he mentioned something17:52
neighborleeand what about tracker..it replaced beagle :)..makes you wonder at least it does me.17:53
neighborleehm lets see17:53
neighborleefound it17:54
neighborleeok this is on forum however17:54
* persia notes bug #190862 may be interesting in any discussion of tomboy17:55
ubotuLaunchpad bug 190862 in tomboy "License headers missing" [High,Confirmed] https://launchpad.net/bugs/19086217:55
neighborleeoh gawd clealry this is a HUGE issue for those involved in the discussions..a sorta 'heated' debate it looks like.17:55
neighborleepersia, hmm17:55
neighborleelicense headers oh thats inntersting ;)17:55
LaserJockneighborlee: yes, because almost are all arguments are a technical "smoke and mirrors" around "we don't want it"17:55
mjg59persia: Unless there's any reason to think otherwise, the presence of a global copyright file is generally sufficient17:57
* mjg59 heads out17:57
xivulonheno: once you get the artwork for umenu could you please also produce a set of images of size 150x57 to be used in wubi header?18:12
xivulonideally there should be one for each flavor covered by wubi18:12
xivulonmjg59, there used to be some code in powermanagement-interface/pmi.acpi to disable suspend (ram/disk) for loopinstallations18:16
xivulonthat for some reason did not go through to pm-utils18:16
xivulonwould it be possible to add it back? grep host in pmi.acpi18:16
xivulonah missed the heads out part...18:18
=== \sh_away is now known as \sh
pranithis there any update for the vmsplice exploit for ubuntu yet?18:41
sistpoty|workpranith: yes18:41
pranithi did apt-get upgrade today, and no new kernel...18:42
sistpoty|workpranith: at least I've seen a new kernel from security today18:42
sistpoty|workpranith: arrived not too long ago18:42
jdongit was put in git like 23hrs ago18:42
pranithok, thanks18:42
jdongupdate released on all supported versions of Ubuntu :)18:43
geserslangasek: the last comment in bug #187890 says that the package got synced but the archive has still the old version. A bug in a script not catching an error?18:52
ubotuLaunchpad bug 187890 in onscripter "Please sync onscripter 0.0.20080121-1 (universe) from Debian unstable (main)" [Wishlist,Fix released] https://launchpad.net/bugs/18789018:52
\shsoren, would you like to take a look at bug #191324 and upload please, so linda and everyone else is satisfied again using dpkg-genchanges.pl ? :)18:52
ubotuLaunchpad bug 191324 in dpkg "dpkg-genchanges.pl missing the "Description" field in *_source.changes files" [Undecided,New] https://launchpad.net/bugs/19132418:52
\shsoren, I just tested it , and everything is fine again and following debian policy18:53
sistpoty|work\sh: revu is already fixed *g*, and sparky's linda as well *g*18:53
slangasekgeser: mmm, could be.  the sync-source script apparently exits with success when you fail to ask it to override the Ubuntu changes, and so pitti's most recent wrapper script sends some false-acks :/18:53
\shsistpoty|work, well, if dpkg is fixed, you can drop the workaround ,-)18:54
slangasekRiddell: I guess you have the sync lock right now, could you add 187890 to your queue?18:54
persiageser: There's an issue with package versions that small.  I had to manually force a couple packages earlier in the cycle18:54
\shsistpoty|work, but I'm uploading my package again, ok?18:55
persiainfon 0~r144-1 and infon-devel 0=r198-2.118:55
sistpoty|work\sh: I'd rather not drop the workaround... revu shouldn't throw backtraces because of a missing entry in the changes file (of course I should really fix much more for that *g*)18:55
Riddellslangasek: ok18:55
sistpoty|work\sh: sure (btw.: I put back your last upload)18:55
sistpoty|work\sh: so that's up there already18:55
\shsistpoty|work, so it's already on the list, cool :)18:55
Riddellslangasek: that can't be synced, it has a smaller version number than the current version we have19:01
slangasekRiddell: oh, that's also a good reason for not syncing then19:02
geserslangasek: so we need to add an epoch to onscripter?19:03
slangasekgeser: to be syncable again, yes19:03
jaaltoWhere can I find step by step instructions how to make a release and upload files to it in launcpad. I have forgotten the details19:03
slangasekgeser: or else drop the leading zeroes, and worry about epoching later?19:03
ScottKDebian would need to add the epoch, right?19:04
slangasekepoch, or drop the leading zeroes19:04
geserScottK: if we don't want to carry it forever, yes19:04
geserslangasek: wouldn't dropping the zeroes also require to rename the orig.tar.gz?19:05
* ScottK was thinking if we wanted to be able to sync it. Once you add an epoch for Ubuntu, you aren't syncing any more.19:05
slangasekgeser: well, yes - why would that hurt anything?19:05
geserslangasek: not really19:06
slangasekScottK: we already can't sync it because the current Ubuntu versioning sorts greater than the Debian versioning19:06
ScottKRight, but us adding an epoch doesn't help that.19:06
geserScottK: how does dropping the zeros help us?19:07
* ScottK didn't say it does.19:07
slangasekScottK: sorry, I read "we need to add an epoch to onscripter" as "we need to get the Debian maintainer to add an epoch to onscripter" :)19:08
ScottKAh.  Maybe that's what he meant and I'm to much of a literalist.  Dunno.19:09
* geser will upload onscripter without the leading zeroes19:12
=== Ubulette_ is now known as Ubulette
slangasekgeser: upload it to Debian?19:12
geserslangasek: to Ubuntu :)19:14
geseras a fake-sync19:14
slangasekoh, gotcha19:14
geserwhat's the correct version in this case? -1build1 as there is no real change or is it -1ubuntu1 due to the changed versioning?19:15
xivulonslangasek, is it possible to agree on metalink urls, even without dealing with generating the metalinks? I need to hardcode those in wubi.19:15
xivulonotherwise I will need to make amendments post feature freeze19:16
persiageser: Maybe -0ubuntu1 because of the changed versioning?19:16
geserthat's also an option19:17
xivulononly need to know where the metalinks will be, and convention for filename therein (relevant before final release), we can use static metalinks for the time being pointing at daily-builds19:22
xivulonslangsek: something like: ubuntu.com/metalinks/ubuntu-8.04-desktop.metalink -> ubuntu-8.04-desktop.iso (filename)19:25
slangasekxivulon: let me see about working through that over the next few hours.  Any new URLs under ubuntu.com need to be discussed with the webmaster first19:30
xivulonslangasek, keep in mind that 1: the metalink url should not change over time within a release cycle, 2: the filename within it may or may not change19:34
xivulonto make clear the second point, assume the metalink is called ubuntu-8.04-desktop.metalink, that will now point to urls that will retrieve the file hardy-desktop.iso19:35
xivulonon top it contains a filename property (which most downloaders use to rename the downloaded file). That can be hardy-desktop.iso or ubuntu-8.04-desktop.iso19:36
ScottKRiddell: Could I please have a give back on testresources.  I believe it'll build this time.19:36
xivulonI'd opt for the second option19:36
RiddellScottK: not from me19:37
RiddellScottK: try slangasek19:37
ScottKOK.  Riddell: who from then?19:37
ScottKslangasek: Would you please give back testresources.  I'm reasonably certain it will build.19:38
geserScottK: ask pitti, Mithrandir or lamont for a give-back19:38
geserScottK: you need an build admin not an archive admin19:39
lamontgeser: s/lamont/infinity/19:39
* ScottK hopes one of them see that then.19:39
slangasekScottK: not me either, I'm not a buildd admin19:39
geserlamont: aren't you a buildd admin anymore or do you simple don't handel give-back requests?19:40
lamontgeser: well.....19:40
geserScottK: Hobbsee can also do give-backs, but that doesn't help you right now19:40
lamontI'm an lp-admin now, so I can certainly do it.19:41
* ScottK would need her long stick to reach out and wake her up.19:41
lamontI prefer to not abuse the duck that way19:41
geserlamont: I looked at https://edge.launchpad.net/~launchpad-buildd-admins/+members and you're still listed so I guessed you could do it19:42
lamontyeah.  I should really fire myself from that group, to be proper19:43
ScottKlamont: Please give back testresources first.19:44
lamontScottK: dropping myself from the group won't make it so I can't give it back..19:44
ScottKTrue, but if you did it first, it would actually be done.19:45
lamontboth done19:47
=== awalton__dos is now known as awalton__
ScottKlamont: Thanks.19:50
slangasekxivulon: ok, I've given the ubuntu.com webmaster food for thought; there are various concerns about how much of a difference this will make to www.ubuntu.com load at release time and so forth, but we'll work through those and I'll let you know what conclusions we reach.  If this is just a matter of adding the metalink URLs to wubi, though, I wouldn't worry about not having the decision before FF20:07
bddebianOh, who is doing give-backs?  Can someone do lordsawar?20:08
pittibddebian: can do20:10
pittibddebian: people in the ~launchpad-buildd-admin team can20:10
pittibddebian: nothing to give back, it's built everywhere but hppa (and it's needsbuild there)20:11
pittihi LaserJock20:11
LaserJockpitti: you gotta a sec for a NEW consultation?20:12
newz2000xivulon: Hi, I'm the webmaster. I have some questions but I need to do a little investigation in order to speak intelligibly on the matter. Can I e-mail your or /msg you later on this week?20:12
pittiactually not, trying to concentrate to get this usplash thing fixed and then go to bed20:12
LaserJockpitti: fine20:12
pittiLaserJock: but just ask your question,20:12
bddebianpitti: Ah, someone must have done it, thanks20:12
LaserJockI'm totally switching the packaging for squeak (multiverse smalltalk stuff) so that we're syncing to the official unofficial packages20:13
pitti'official unofficial'? :)20:13
LaserJockthe source package names are completely different20:13
pittiI see20:13
pittiso that requires some removals/NEWs20:13
LaserJockpitti: Debian won't take it officially, but squeak maintains official packages20:13
LaserJockbut there is some binary package overlap20:13
pittiLaserJock: that's not a problem; just make sure that the versions are higher20:14
LaserJockyes, the versions will all be higher20:14
pittiand that the packages with the same names have the same purpose :)20:14
pittiand keep transitional empty packages until after hardy for upgrades20:14
LaserJockok, so I shouldn't ask for removals?20:14
pittithe old source packages should be removed20:15
pittiand old libraries, etc.20:15
LaserJockk, and have the new packages do the transitionals20:15
pittiold application packages should become transitional packages20:15
LaserJockok, how hard would it be to get somebody to remove squeak-image, squeak-sources, and squeak-vm for me? :-)20:16
pittiLaserJock: let's first get the new packages in, then remove the old onese20:16
pittiLaserJock: it's not hard, just file a bug against them asking for removal and sub ~ubuntu-archive20:16
* pitti sighs and goes for another round of reboot/usplash testing; brb20:18
pitti4 usplash uploads in two days...20:18
LaserJockfun :/20:19
ScottKpitti: If you're going to be around for a bit, you might idle in #ubuntu-meeting and 'fanboy' my core-dev application in person when the time comes (meeting in progress now).20:20
MezQuestion: with all the idiots trying to get people to sudo rm -rf / - why not create a patch that gives you an "are you sure - are you really sure - are you REALLY REALLY sure" kind of warning?20:21
persiaMez: Fails the silent-on-success guideline20:22
LjLMez: as in, putting an alias by default in ~/.bashrc for --preserve-root?20:22
ScottKbddebian: Testresources FTBFS again.  It's your fix, so have fun.20:22
MezLjL, something like that20:24
=== jelmer_ is now known as jelmer
MezLjL, actually, thats not a half bad idea20:25
MezAnyone willing to sponsor a patch for that20:26
LjLMez: not sure. guidelines aside, people would probably starting to give "alternative commands" to bypass that check20:26
MezLjL, it's built into rm, so wouldnt fail the guidelines20:26
infinityMez: Preventing users from shooting themselves in the foot with cute aliases and such never works.20:26
Mezinfinity, wow - havent seen you in ages20:27
Mezhow're things20:27
infinityDecent, I just don't speak up in -devel much anymore20:27
MezI've actually looked for you specifically on a couple of occasions infinity ;)20:28
Mezcant remember why now20:28
Mezprob something to do with Debian and NM20:28
slangasekor turtles20:28
infinityMez: (Note that we could just change it to "rm -rf /*" etc... If people will run random commands like that, they'll always get what's coming to them)20:28
infinityslangasek: *glare*20:28
Mezinfinity, ;)20:29
Mezslangasek, do I want to know20:29
slangasekyes, but I won't tell you20:30
=== mdz_ is now known as mdz
=== Keybuk_ is now known as Keybuk
xivulonslangasek: thanks, yes for the time being it's mostly a matter of URL only I can point to my website for the time being and repoint later.20:42
xivulonNote that if the "filename" attribute within the metalink is not constant throughout a release cycle, I will also need to recode a bit.20:42
slangasekxivulon: right.  I'm not sure what the best solution is for that side of things; everywhere else, we make a point to distinguish between pre-release codenames, and release versions20:43
xivulonslangsek, if you keep the metalink url fixed, either way is wrong20:44
xivulonslangasek: ^20:44
xivuloni.e. you either have a ubuntu-8.04-desktop.metalink pointing at hardy-8.04-destop.iso or you have ubuntu-8.04-desktop.metalink which downloads hardy-desktop-iso and calls it ubuntu-8.04-desktop.iso20:45
slangasekxivulon: yes, it's also wrong because we don't want to have /metalink/ files labelled "hardy" after release, or labelled "8.04" before release20:47
xivulonan alternive is to use server side redirection20:48
xivulonso I can point to ubuntu-8.04-desktop.metalink.latest20:48
xivulonwhich will redirect me to either ubuntu-8.04-desktop.metalink or ubuntu-8.04-alpha3-desktop.metalink20:49
xivulonthe metalink file itself by the way should be lighter than an average webpage if you are concerned about bandwidth20:50
slangasekxivulon: what about if the installer were to look for a file under the release number, and fall back to the codename if it's not available?20:50
xivulonI can do that but will hardcode some ubuntu specific practices20:51
slangasekthat would also give us the option of hosting them in two different places (i.e., releases.u.c, cdimage.u.c, which is how the images are split as well)20:51
xivulonI was thinking somthing like:20:51
xivulonubuntu.com/metalinks/8.04 with subdirs: final, beta, alph3, current20:52
xivuloncurrent contains "fixed" urls which are redirected20:53
slangasekthe subdirs seem unnecessary, since we don't generally keep previous milestone images around after the next one has been released20:53
xivulonagree, on top of that server side redirection do not require files visible to the user20:54
xivulonso it would be ubuntu.com/metalinks/8.04 containing ubuntu-8.04-desktop.metalink OR ubuntu-8.04beta-desktop.metalink + ubuntu-8.04-desktop.current.metalink redirection20:55
xivulonthe last one being "invisible"20:55
xivulonthe filenames therein should match whatever a user would grab manually from the mirrors20:56
xivulonwould the above seem reasonable?20:57
slangasekit still has the flaw that it implies using the "8.04" designation before 8.04 is released20:57
slangasekI expect newz2000 will have some opinions on the layout20:58
Mithrandirwe could use metalinks/NOT-YET-RELEASED/8.04/ or something like that.20:58
xivuloncould be called hardy-desktop, the issue is that all distros share the same codename so can't have that...20:58
infinityWe should definitely not use numbers before release.20:59
xivulonubuntu-beta-desktop.metalink ubuntu-finall-desktop.metalink ...?20:59
xivulonall in 8.0420:59
infinityI still remember the s/6.04/6.06/ pain.20:59
xivulonI mean either one or the other20:59
slangasekxivulon: I would probably do ubuntu-hardy-desktop-i386.metalink or such; prefixing the distro name to the image name20:59
slangasekor else mirroring the per-flavor heirarchy currently present on releases.u.c/cdimage.u.c21:00
xivulonslangasek: I really only care about having redirections (.htaccess) which are fixed across a cycle and gets updated21:00
xivulonThe rest is for "real users" more than programmatic access21:01
xivulonprovided the redirections get me to the right metalink, I do not really mind what is called21:01
infinityWhat's the actual goal here?21:01
infinityA static URL that's always correct?21:01
slangasekinfinity: giving wubi a static url that can be embedded, yes21:02
infinityIt's my opinion that having the current alpha turn into the release is very much incorrect.21:02
infinityThis is one reason why we do the rename on release.21:02
infinityubuntu-hardy-desktop.iso != ubuntu-8.04-desktop.iso, and never should we promote anything that claims otherwise.21:03
xivulondon't think there is much confusion there21:03
slangasekthere isn't *today*, because we make the distinction21:04
infinityThere would be with a static URL that tracks through to release.21:04
Mithrandirany reason wubi can't try two URLs and fall back to the not-yet-released one if the released one doesn't exist?21:04
Mithrandirso it'd be embedding two URLs, not just one21:04
infinityHaving different locations and filenames is a big, blinking warning that you're using unreleased code.21:04
infinityAnything that obscures that without warning the user is a no-no to me. :/21:05
infinitys/unreleased/unsupported/, if you prefer.21:05
infinity(It's obviously all "released")21:05
xivulonThat is of course not an issue for the stand-alone version, but it is problematic for the version bundled with the CD21:05
infinityWhy is it an issue?21:06
slangasekfor the version bundled with the CD, it could read metadata from the CD image itself telling it which url to check...?21:06
xivulonI will have to change the embedded urls just before release21:06
infinityif (download release.iso) { continue; } else { download beta.iso; warn user that it's beta; }21:07
xivulonslangasek: I do parse .disk/info anyway...21:07
slangasekxivulon: in that case, you already have the metadata that tells you whether it's released21:07
xivulonsorry infinity misunderstood you above, yes I can do 2 urls21:07
xivulonslangasek: true.21:08
xivulonif a user gets a copy of wubi.exe and pass it to a friend, that becomes stand-alone, and will not "know" what to download21:09
slangasekare there identifying version numbers in wubi itself?21:10
Mithrandirxivulon: make it look at http://releases.ubuntu.com/.manifest or something then?21:10
xivulonNow I use 8.04+bzr revision21:11
xivulonBut can of course change that21:11
xivulonMithrandir: you mean the wubi name?21:11
xivulonah no21:12
Mithrandirxivulon: if you have a standalone .exe, what information does wubi want to collect before it can install?21:12
xivulonit needs to know the metalink url21:12
Mithrandirit needs to be able to construct the right metalink URL, right?21:13
xivulonif the standalone is produced after the ISO ships, that is trivial, but if the standalone is on the CD that is less so21:13
Mithrandirit's not needed for wubi from 8.04 to be able to install 8.10?21:13
slangasekhuh? why does it make a difference when the .exe is built?21:14
xivulonslangasek: if I release an exe on my website after ISO release I can embedded the right urls21:14
slangasekwe will know, a priori, what the urls are for "pre-release" and "release" metalink files, whether these are the same or different21:14
slangasekI think we're all agreed that the urls should be *defined* well in advance21:14
slangasekit's just that the "release" url shouldn't be *populated* prior to the release21:15
Mithrandirthey'll just 404 until we actually have released.21:15
xivulonslangasek: correct21:15
slangasekxivulon: ok, so where's the problem then?21:15
xivulonnot really a problem, so far there seem to be 2 solutions: I try 2 different urls or we use redirection21:15
xivulonjust need to knwo which route you guys prefer, for me redirection is less work :P21:16
cody-somervilleWould someone be so kind to summarize what the problem is? :)21:17
slangasekwell, yes, but I think the above 800 lines of scrollback make it clear that there's resistance to using a single, user-visible url for pre- and post-release files :)21:17
xivulonah but that will not be user-visible21:17
cody-somervilleWhy would you want that anyhow?21:17
slangasekcody-somerville: the subject at hand is wubi using metalink files; I don't know that there is a problem currently21:18
slangasekxivulon: uh, it'll be on a website, users can access websites, therefore it's user-visible :-)21:18
xivulonin the sense the redirection is serer side (.htaccess) so if you look in the metalinks directory you do not see redirecting urls21:19
xivulonof course you can also opt for visible redirections21:19
slangasekmm, ok21:20
slangasekeh, "server-side redirect" is a very vague term without any implications of .htaccess or any other mechanism; that was my confusion, citing wikipedia doesn't really help :)21:23
xivulonWhat I mean is that I can type "http://ubuntu.com/metalinks/hidden_latest.metalink" and get "http://ubuntu.com/metalinks/ubuntu-8.04-desktop.metalink"21:25
xivulonbut if you look in "http://ubuntu.com/metalinks" you only see "ubuntu-8.04-desktop.metalink"21:25
slangasekyes.  "server-side redirect" implies the former, it didn't imply the latter.  But I'm on the same page now.21:26
xivulonin any case the downloaded metalink (and hence iso filename) would reflect the actual status21:28
xivulonso do we use (hidden) redirection or do I test for 2 URLs?21:29
slangaseknot our decision to make yet, still needs the webmaster's buy-in :)21:29
slangasek(and if the webmaster nack's hosting it on www.ubuntu.com, then it'll have to be two urls since pre-release images are hosted on a completely different server from releases)21:30
=== \sh is now known as \sh_away
xivulonI would assume that in any case, the filename within the metalink will match the underlying iso file URL (i.e. now it would be hardy-desktop-i386.iso)21:31
xivulonslangsek thanks a lot for bearing with me21:32
slangasekno problem at all; sorry for having deferred this so long21:33
xivulonnot strictly a problem for me, I simply wanted to avoid post feature-freeze changes21:33
sorenpitti: debian bug 46534021:51
ubotuDebian bug 465340 in dpkg "dpkg: Broken call to open in Dpkg/Control.pm" [Important,Open] http://bugs.debian.org/46534021:51
=== crevette is now known as crevette_
=== crevette_ is now known as crevette
=== crevette is now known as crevette_
sorenpitti: Short version: Please use dpkg-gencontrol differently :)21:57
* soren gives doko "the eye"22:10
* ion_ has two.22:11
doko_siren: ?22:13
doko_soren: ?22:13
=== doko_ is now known as doko
sorendoko: Hm.... I'm having odd problems with grub and right now, my prime suspect is your dpkg patch :)22:14
sorendoko: Yup, it seems to be accurate.22:17
sorendoko: It checks:22:17
sorenif test "x$CFLAGS" = x; then default_CFLAGS=yes22:17
sorenAnd if $default_CFLAGS isn't "yes", it doesn't check for e.g. -fno-stack-protector and then it blows up.22:18
dokosoren: ohh great, it does the CFLAGS settings stuff in debian/rules, but then it doesn't pass it to the upstream make22:21
sorendoko: What would you suggest? Adding -fno-stack-protector to CFLAGS? unset CFLAGS before calling configure?22:21
dokoor configure22:21
sorendoko: No, it *does* get passed to configure. That's the problem. configure behaves differently depending on whether or not CFLAGS is unset(/empty) or not.22:22
dokosoren: this is insane. the conservative thing would be to unexport CFLAGS, and mention it in debian/rules22:25
sorendoko: That's your recommendation?22:27
sorendoko: I just tested it, and it does indeed fix it. Not surprising, I'm just saying.22:27
dokosoren: unexport CFLAGS, and maybe remove the setting of CLAGS in the rules file, it's just misleading22:28
sorendoko: Hm... I'm not sure I remember this correctly..22:30
sorenIf a makefile specifies CFLAGS, but there's also a CFLAGS environment variable, what happens?22:30
sorenThe environment variable takes precedence, right?22:31
dokosoren: in debian/rules, CFLAGS is not passed to configure, so CFLAGS from debian/files is ignored, configure takes it from the env. in the generated Makefile, it's available and taken, unless you call make with the -e option22:32
sorendoko: Ah.. Ok.22:33
emgentdebian #46298422:39
ubotuDebian bug 462984 in python-moinmoin "python-moinmoin: MOIN_ID cookie bug" [Serious,Open] http://bugs.debian.org/46298422:39
=== ion___ is now known as ion_

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