/srv/irclogs.ubuntu.com/2008/01/23/#ubuntu-devel.txt

* lamont wonders if any german speakers are awake for a totally-off-topic translation question01:10
Hobbseeheh01:11
Hobbseetry #ubuntu-de?01:11
lamontfeh01:11
lamontbe logical.01:11
Hobbseemy german won't be anywhere near good enough01:11
lamontHobbsee: they're all speaking german there!!!! :-)01:12
Hobbseethey should speak some english :P01:13
StevenKWouldn't it be funny if lamont got kicked for speaking English ...01:15
lamontStevenK: hehe01:18
lamontthey sent me to google01:18
StevenKHaha01:18
* ScottK looks around for an archive admin ...01:21
StevenKBit late for them.01:22
ScottKTrue.01:22
ScottKI thought maybe slangasek or Hobbsee might be up and about.01:22
* lamont tickes Hobbsee 01:24
* Hobbsee stomps on lamont's feet01:24
lamontScottK: I give you Hobbsee, live and in person.01:24
* Hobbsee bites01:24
ScottKHeh.01:24
* Hobbsee claws01:24
* Hobbsee uses the Long Pointy Stick of DOOM!!!!!!!!!!!!!!! ™ on ScottK01:24
* Hobbsee sets lamont and ScottK on fire01:25
ScottKHobbsee: After you're finished chewing on lamont, would you please accept the 3 uploads sitting in dapper backports?01:25
lamont"Jesus saves.  Gretsky steals, shoots,  SCORES!!!"01:25
* ScottK warms up.01:25
ScottK(It's been cold here lately)01:25
ScottKThanks.01:25
Hobbseelet me see....01:25
lamont'twas below zero (Fahrenheit even...)  this morning01:26
Hobbseesylpheed* and php*?01:26
ScottKYes.  Two for slypheed.01:26
Hobbseeyep.  done.01:26
lamontphp-clamavlib_0.12a-4~dapper1_source.changes01:26
lamontsylpheed-claws_1.0.5-5.1ubuntu0.1~dapper1_source.changes01:26
lamontsylpheed-claws-gtk2_2.6.0-1.1ubuntu1.1~dapper1_source.changes01:26
ScottKHobbsee: Thanks.01:26
Hobbseeyou're welcome01:26
lamontwe now return you to your regularly scheduled topic.01:26
lamont(and thanks)01:26
* Hobbsee sets lamont on fire again, then.01:27
lamontah.  warmth.  thanks01:27
lamontdanke, even01:27
Hobbsee:)01:27
* Hobbsee smashes edge with a large brick01:29
lamontedge or edgy?>01:31
Hobbseelamont: can you do it somehow?  launchpad is scrweing up.  again.01:32
StevenKI'm suspecting edge.launchpad.net01:32
StevenKAh. I was right01:32
Hobbseeit's edge.  and production.01:32
ScottKHow can LP be screwing up, Ubuntu doesn't have a release or freeze on right now?01:33
Hobbseeedge fails with a timeout, production fails with a non-timeout-oops.01:33
lamontoh.  edge.  yeah01:34
lamontHobbsee: what exactly are you doing that's timing out?01:34
lamont(url??)01:34
Hobbseeaccepting the dapper backports packages01:35
Hobbseehttps://edge.launchpad.net/ubuntu/dapper/+queue?queue_state=1&queue_text=01:35
lamontHobbsee: and I have your OK to accept those on your behalf?  (just to see if it'll work for me)01:35
Hobbseelamont: yeah01:36
lamontWaiting for edge.launchpad.net.................................01:36
lamontAccepting Results:01:36
lamontOK: sylpheed-claws-gtk2, OK: sylpheed-claws, OK: php-clamavlib01:36
lamontnot the quickest update I've ever seen01:36
Hobbseeright, so it works for you then?01:37
lamontyeah.01:37
lamontI should have just done one, then we could have played more.01:37
lamontsorry01:37
lamonthopefully it's not duck-related.01:37
Hobbseethanks01:37
HobbseeScottK: upload more crackports, please01:37
* Hobbsee tried it multiple times01:38
ScottKHobbsee: You need to get the tech-board to approve my core-dev application before I can do it without a beard like lamont.01:38
ScottK;-)01:38
Hobbseehehe01:39
* Hobbsee has almost no say over the techboard, sorry01:39
Hobbseewhen did they want their meeting?01:39
lamontHobbsee: I wonder if maybe it's because ScottK's sig is on the .dsc files...01:39
Hobbseelamont: as to why it got into unapproved?01:39
lamontno, as to why it OOPSed01:39
ScottKHobbsee: All backports go to unapproved.01:40
lamontbecause crackporting is an unapproved activity01:40
lamontUploaded By:  Scott Kitterman01:41
lamontWIN01:41
lamonthttps://edge.launchpad.net/ubuntu/+source/sylpheed-claws/1.0.5-5.1ubuntu0.1~dapper101:41
ScottKI got accepts for all three.01:42
lamontPENDING: Dapper  pocket Backports  in component universe  and section mail01:42
lamontit's in universe... you could have done that one01:42
lamontsame for -gtk201:43
ScottKlamont: Nope.  Backports uploads always take a core-dev01:43
* lamont grumbles at ScottK 01:43
lamontoh, really?01:43
ScottKEven for Universe.01:43
ScottKYeah.01:43
lamonthow, interesting01:43
* lamont ponders, remembers whyu01:43
zullamont has a beard?01:48
lamontzul: no, apparently I _am_ one01:48
zulah ok apparently i missed that part01:48
ScottKAnd a very helpful one at that.01:49
StevenKI don't see how beards are helpful in the usual case ...01:50
* ScottK neither (it's really pretty obsolete), but it seemed like a funny way to put it.01:50
ScottKThanks again Hobbsee and lamont.01:51
ScottKThat's 8 down.  6 to go.01:53
Hobbseeyou're welcome01:53
=== kitterma is now known as ScottK2
* lamont stares at his firewall rulesete02:25
lamonts/e$//02:25
* ScottK looks at the 1807 packages needs-building for Hardy and despairs his dapper-backports will ever build.02:28
LaserJockseriously?02:29
ScottKYeah.02:30
ScottKThat's down about 100 already from it's peak.02:30
lamontScottK: launchpad.net/+builds shows 104 for i38602:32
ScottKWell these are all arch: any, so whatever you did to hppa is going to slow things down.02:32
lamonthuh?02:33
lamontyour backports will be a while before hppa (655) gets to it, but what do you care?02:33
lamontor does it actually require it to be current everywhere before it publishes (/me thinks "no")02:33
ScottKlamont: Now, but I won't relax until it's all done.  Only clamav needed to go through binary New and that's done.  That's the only one it would have mattered for.02:37
* lamont wears his "being good" hat02:39
lamontHobbsee: did you try to accept all 3 packages each time (edge)?02:39
lamontor just one?02:40
Hobbseeall02:40
lamontif it happens again, try just one at a time... 'twould be an interesting data point02:40
lamontword is that it's not duck-related02:40
bddebianI suppose pitti is asleep this time of day?02:44
zulmost likely02:45
warp10Heya06:38
jackdawhello, i07:35
pittiGood morning08:25
pittibdmurray: hi08:25
pittibdmurray: unping; that was supposed to be 'bddebian'08:26
gaspapitti: morning.08:29
pittihey gaspa08:31
pittigaspa: getting to your mail, promised (I'm a bit backlogged on the sprint)08:31
gaspapitti: np.08:31
gaspai'm not in a hurry. :)08:32
* gaspa remembers someone's blog... :P08:32
=== carlos_ is now known as carlos
Keybukcalc: openoffice.org-core overwrites something in common08:46
mptGooooooooooooood morning Ubuntu lovers!09:02
TheMusompt: isn't that a bit old? :p09:06
mptTheMuso, old? It's only the second time I've ever said it09:06
TheMusompt: Well ok then.09:07
RAOFmpt: TheMuso works on fuzzy pattern-matching, so it *is* a bit old :)09:08
TheMusoRAOF: lol09:09
bdmurraypitti: wow, that usually happens the other way around09:10
pittibdmurray: ETABCOMPLETION :)09:15
slangasekScottK: this is sprint week, it's late for me too. :)09:23
gesergood morning09:42
pittihi geser09:46
geserHi pitti09:52
ion_Freenode.find_by_channel_and_status('#ubuntu-devel', :nonidle).each {|person| greet person }09:54
evandAnyone have an objection to me building a new i386 desktop CDs, or know what the giant warning in cdimage's crontab means?  pitti, slangasek?10:03
slangasekevand: which warning, the "do not re-enable"?10:03
slangasekthat one means that I failed to delete the warning when reenabling the cronjobs10:04
asac_the_2ndbryce: http://ubuntuforums.org/showthread.php?t=66926210:04
evandslangasek: heh, gotcha10:04
evandand yes, that one10:04
slangasekevand: it definitely doesn't apply, since the cronjobs below it are all enabled ;)  so yeah, if you need a new desktop build, go ahead10:05
evandok, I'll remove the warning as well then10:05
asac_the_2ndbryce: its bug 18203810:07
ubotuLaunchpad bug 182038 in xserver-xorg-video-nv "Black rectangle instead of image in FF3 [Hardy]" [Undecided,Confirmed] https://launchpad.net/bugs/18203810:07
calcKeybuk: yea i'll have to fix that in the next upload :\10:12
tjaaltonasac_the_2nd: I'll forward it upstream (xorg)10:12
calcKeybuk: the work around for now is to uninstall the package it is overwriting, it should have replace/conflict that old package version :\10:13
asac_the_2ndtjaalton: thanks10:13
seb128lifeless: http://bugzilla.gnome.org/show_bug.cgi?id=264264 looks like your issue but it has been duplicated and the other bug has been closed so you might want to reopen or open a new one10:26
ubotuGnome bug 264264 in Mailer "Crash: mouse grab freeze during mail DnD" [Normal,Resolved: duplicate]10:26
lifelessthanks seb12810:34
seb128you are welcome10:34
calcmjg59: ping10:35
asac_the_2ndogra1: http://bazaar.launchpad.net/~ubuntu-core-dev/network-manager/ubuntu.0.710:58
asac_the_2ndogra1: deb http://ppa.launchpad.net/asac/ubuntu hardy main11:05
emgentheya *12:00
=== Am|NickTaken is now known as Amaranth
=== \sh_away is now known as \sh
calcanyone happen to know why the regular gstreamer plugins don't do mpegdemux, just the fluendo plugin?12:49
calcis it a patent issue or something like that?12:50
calc"The main difference between this and the plugin from gst-plugins-ugly is that this one allows demuxing of MPEG2 transport streams.12:50
calc"12:50
hungerq12:54
calcseb128: i see the weather is a tooltip actually on the new clock but the temperatures seem off12:56
calcseb128: it claims it is -20C in Houston, heh12:57
calcthat would be about like Hell freezing over ;)12:58
seb128calc: it seems to be working for me12:59
seb128calc: 3.6°C for Houston here13:00
calcit actually says '-4.7F' i did the conversion in units13:00
calcis there a way to force the clock applet to update (like there is in gweather)?13:00
seb128not that I know13:01
seb12838.6°F for houston13:01
zulcalc: -20C is nothing13:01
calczul: in Houston it is13:01
calczul: Houston is probably > 15C right now13:02
zulonly :)13:02
calczul: ah well 3.6C anyway (its still early in the day there)13:02
calcseb128: switching back to C it says -20.4C heh13:02
calcseb128: oh i know what is wrong13:03
zulcalc: -9C outside where I am13:03
calcseb128: i have to pick a timezone not a city, so its giving me the temp for chicago13:03
seb128calc: picked the wrong houston?13:03
seb128calc: hum? I picked the Houston Intercontinental Airport in the locations list13:04
calcseb128: i am now more confused than before, i see what i did wrong finally13:04
calcseb128: i picked America/Chicago and didn't even notice the "Find" button13:05
calcoops13:05
* ogra1 is a bit irritated by all the popunder windows13:05
calchmm when i changed to Houston Intercontintental it changed my timezone setting13:05
calcto Monterrey, not sure if that is the same as Chicago so I switched it back13:06
calcworks now though for temp :)13:06
calcanyone else notice gweather (applet) complains about it can't find the xml locations file?13:12
calcah i see a bug is already filed13:14
pecisknautilus-connect-server will return soon? :) gvfs works nicely so far with ssh connections13:37
seb128pecisk: no13:37
seb128pecisk: gvfs does mounting now and the mounts are listed, gnomevfs worked differently13:39
peciskI see, but nautilus-connect-server dialog will be rewriten or something else will come?13:39
=== mathiaz_ is now known as mathiaz
seb128pecisk: not clear yet, you can read the mail sent by alex on the GNOME desktop list this morning on the topic13:40
bryce_seb128: btw, that bug I showed you earlier appears to also occur with xterm, so is probably not gnome terminal specific13:42
peciskseb128: I quite don't get it. How someone will create connection to server?13:43
peciskok, I will read all thread13:44
peciskthanks for info13:44
pecisk:)13:44
seb128bryce_: looks like a xorg issue ;-)13:45
bryce_:-P13:45
seb128pecisk: ctrl-L, enter your URI, the server is automatically mounted on first use and stay there13:45
seb128pecisk: gnome-vfs didn't do this mounting so the connect server was an handy way to use the same URL again13:46
seb128that was basically opening URI every time you used the shortcut13:46
peciskI see13:47
peciskbut CTRL+L is somehow very cryptic for common user13:47
seb128where now the mount stay there and is listed in all the applications during the session13:47
peciskyou have to enter all parameters there13:47
seb128not really13:47
seb128if you can connect somewhere and a password is required it'll ask for it13:48
seb128and you can browser smb shares13:48
peciskhmmm13:49
peciskyeah, I see know13:49
pecisknow13:49
peciskwhen I think, this is much easier, I have to admit13:49
peciskonly question how to create shortcuts to reuse connections - as launchers?13:52
seb128pecisk: that is not done yet13:53
peciskok13:53
peciskthanks for info, nice work guys13:53
seb128you are welcome13:53
=== enrico is now known as pippo
=== pippo is now known as enrico
=== enrico is now known as peppa
=== peppa is now known as enrico
=== \sh is now known as \sh_away
=== \sh_away is now known as \sh
=== bigon` is now known as bigon
=== ryu2 is now known as ryu
pochucjwatson_: hi. I'm still moderated in ubuntu-devel@lists.ubuntu.com. Could you please whitelist me? pochu@ubuntu.com. Thanks!15:11
Hobbseepochu: mail unmoderated, anyway15:17
=== _emgent is now known as emgent
MacSlowmvo, having keyboard-shortcuts exposed in the window-action-menu needs a patch to libwnck, which would introduce a gconf-dependency15:42
MacSlowmvo, I asked vuntz and he said it would be ok if I provide a configure/compile-time settable option for this15:42
mvoMacSlow: where does it currently get the keyboard shortcuts from?15:43
MacSlowmvo, not at all15:43
MacSlowmvo, it's not exposing them in any way15:43
mvobut when I right-click, I see "ALT-F7" here etc15:43
* MacSlow starts metacity...15:44
MacSlowhm... metacity does maybe not use libwnck for the action-menu?!15:44
MacSlowmvo, my gues was right... metacity does not use libwnck for the action-menu15:45
mvoMacSlow: right. perhaps we should also move that into gtk-window-decorator then?15:46
MacSlowmvo, I would like to fix that... I mean libwnck using gconf to obtain keyboard-shortcuts15:46
MacSlowmvo, hm... no... I don't like that15:46
mvoMacSlow: so that metacity would use it? yeah, that sounds better15:46
MacSlownot that I want to patch metacity afterwards too to use it...15:47
MacSlowjust place stuff in the correct places15:47
MacSlowmvo, argl... crap...15:49
MacSlowmvo, retrieving the shortcuts from gconf will need to look for metacity- or compiz-specific keys explicitly15:50
MacSlowthus it would be required to do window-manager-specific code in libwnck...15:51
stgrabermvo: Are you aware of a bug causing aptitude and apt to produce lots of empty lines + confirming the upgrade/install (sort of simulating the enter key) ? (only happens when apt did "Calculating upgrade" or aptitude did "Building tag database" before)15:57
stgraberI have it for quite some time now and managed to reproduce it on all my Hardy systems so far15:57
bryce_kees, bug 184492 is being blamed on the gutsy xserver security update, but afaik the changes shouldn't have effected suspend at all...  can you take a look at this one and comment?15:59
ubotuLaunchpad bug 184492 in xorg-server "xserver-xorg-core update breaks suspend in gutsy" [Undecided,New] https://launchpad.net/bugs/18449215:59
bryce_keescook: also bug 18486916:01
ubotuLaunchpad bug 184869 in xorg-server "ALT Super keys swapped after upgrade" [Undecided,New] https://launchpad.net/bugs/18486916:01
mptKeybuk, ExitStrategy involves a bit of kernel work now, so I'm not sure it belongs in the DesktopTeam/ wiki namespace16:05
slangaseklool: wrt your changelog entry for the latest nautilus upload, is there any handling for older packages that require nautilus-extensions-1.0?16:10
slangaseklool: I stumbled across this via totem, which is currently FTBFS because the build rules look for /usr/lib/nautilus/extensions-1.0; the FTBFS is an easy fix, but it set me to wondering about upgrade handling from older versions16:11
loolslangasek: seb128 wanted to upload nautilus as fast as possible and said we would add the Breaks when the packages are actually ready16:19
loolslangasek: The provides is just to be able to use Depends in the extensions instead of Breaks16:19
loolslangasek: After the extensions have been uploaded, we should add a Breaks on the old versions in nautilus16:19
loolseb128: ^16:20
slangaseklool: alrighty; as long as it's on your radar then16:21
slangasekin totem's case the nautilus extension is optional functionality, so I guess it should be exempt16:21
loolHmm that's a good point16:22
seb128slangasek: it's optional functionality for most of the other packages as well16:22
seb128slangasek: totem, file-roller, evince16:22
loolWe should use suggests or recommends then16:22
seb128Recommends when apt will install those in Ubuntu16:22
slangasekseb128: so do you want the breaks to apply to these packages?  they aren't really broken16:23
loolWe should suggests or recommends as we see fit but recommends wont work until apt supports them :)16:23
seb128slangasek: no, that's why I uploaded nautilus yesterday16:23
slangasekmaybe we need [Depends,Recommends,Suggests] and [Breaks,Bruises,Inconveniences]16:23
loolmvo: \o/16:23
loolslangasek: But what about Build-Recommends and Build-Suggests?   :-P16:24
slangasekheh16:24
Mithrandirslangasek: I want "Stabs" as well.16:24
pittidoko:    * lsb-cxx: Drop dependency on libstdc++5, not needed by LSB-3.1.16:43
pittidoko: WOHOOO!16:43
pittidoko: that sounds like we could drop gcc-3.3 to universe?16:43
dokopitti: lets see; talked with Matt Taggert and Marc Tardiff about it.16:47
dokopitti: please fix grub not to build with gcc-3.4 ;-)16:47
sorenkvm needs gcc-3.4, too.16:49
sorenPreemptive response to the inevitable "but that's not in main": Yet.16:50
slangasekpostemptive response: better fix that if you want it to be? ;)16:51
* doko will mention that in the MIR for kvm ...16:53
sorenwith a bit of luck Fabrice will fix this within a few days.16:54
soren(it's a qemu problem, really)16:54
dokoplease check as well, if it does build with gcc-snapshot16:54
sorenIt doesn't.16:54
\shseb128, would like to do me a favour and work on https://bugs.edge.launchpad.net/ubuntu/+source/libmicrohttpd/+bug/184400 It's a blocker for a sync16:59
ubotuLaunchpad bug 184400 in libmicrohttpd "[MoM SYNC] libmicrohttpd 0.2.0-1" [Undecided,Confirmed]16:59
slangasekjoejaxx: did TheMuso talk to you yet about getting the ubuntustudio-desktop seed changed for the libpt-plugins package name change?16:59
TheMusoslangasek: No I had forgotten about that one, but I did talk to him about other seed related stuff, such as merging.17:00
TheMusoThe funny thing is, if it was me doing seeds, I would have jst done it. :)17:00
TheMusoslangasek: joejaxx should be subscribed to get emails of seed changes, so he should know about that.17:01
joejaxxslangasek: nojoejaxx@venus:~/ubuntustudio.hardy$ cat desktop | grep libpt * libpt-1.10.10-plugins-v4l * libpt-1.10.10-plugins-v4l217:01
joejaxx:)17:01
joejaxxbut i already made the changes the day it was made :)17:02
joejaxxunless there was another change and i did not receive an email about it17:02
slangasekjoejaxx: that's the one, yes.  any chance it can be uploaded then, so I can make the old ones disappear? :)17:03
TheMusoslangasek: Does that require a -meta refresh?17:03
joejaxxyes :)17:04
TheMusoRight.17:04
TheMusojoejaxx: You do that, and I;ll do the -meta.17:04
joejaxxTheMuso: ??17:04
pittidoko: grub> ah, that17:05
slangasekTheMuso: he said he already changed the seed, so isn't the -meta all that's left? :)17:05
TheMusoslangasek: Yes, but I thought there was also other stuff to be done as well.17:06
TheMusofor the studio seeds that is17:06
slangasekah, ok; nothing else from my end, dunno what else you guys have pending17:07
TheMusoIn any case, I'll do the meta. joejaxx if there is anything that I need to wait for you to do, plesae let me know now.17:08
TheMusoplease17:08
joejaxxnope there is not :)17:08
TheMusojoejaxx: Ok thanks.17:09
joejaxxTheMuso: you are most welcome17:09
joejaxx:)17:09
joejaxx /win 10717:09
joejaxxklfjgkfjgk17:09
ion_107? Whoa.17:10
keescookbryce_: odd bugs.  there was nothing in the CVE fixes that were related to video drivers...17:22
bryce_keescook: well neither issue seems driver related17:23
bryce_keescook: otoh, I can't fathom why those would have broken from the update17:23
keescookyeah, and the 2nd one scares me a lot -- they have selinux enabled?  if so, their machine isn't running upstart.  who know what else is weird.17:24
\shhmm...will  synced "new packages" (not in ubuntu yet) from debian directly be delivered to the archives or are they just pushed like "new packages to ubuntu" into the new queue?17:30
slangasekthey have to go through binary new17:31
=== Am|NickTaken is now known as Amaranth
seb128\sh: just did the syncs, the new sources will be in NEW but we usually accept debian packages to universe quickly17:42
* _MMA_ thanx seb128 for the Ardour sync.17:42
\shseb128, thx :)17:42
TheMusoslangasek: About to upload the new meta now.17:51
TheMusoSo you can do what you were going to do.17:51
seb128mr_pouit: around?17:51
slangasekTheMuso: cheers :)17:52
mr_pouitseb128: yes17:53
seb128mr_pouit: I did demote to xfce to universe, could you close all the tasks?17:54
mr_pouitseb128: ok ^^17:54
seb128thanks17:55
seb128mr_pouit: you can likely do it by mail like you opened the tasks ;-)17:55
mr_pouitseb128: I hope so :p17:55
=== mathiaz_ is now known as mathiaz
matteohi18:08
matteoNow running lintian...18:08
matteoE: amrenc_0.5.1~gutsy1_source.changes: bad-distribution-in-changes-file gutsy18:08
matteothat's definitely a lintian bug18:08
loolseb128: http://people.ubuntu.com/~lool/packages/cheese/2.21.5-0ubuntu1/hardy/cheese_2.21.5-0ubuntu1.dsc18:09
loolseb128: thanks18:10
seb128you are welcome18:11
sistpoty|workmatteo: yes. looks like lintian was synced from debian instead of merged18:17
ScottKmatteo: That's not a proper Ubuntu revision number18:29
ScottKmatteo: If it's ubuntu something it'll know about gutsy18:29
matteoit's ubuntu hardy18:29
ScottKIt looks like a backport to Gutsy from Hardy.18:29
ScottKIt it's a sync from Debian, .changes should say unstable, and it'd be fine.18:30
jdongshould there really be a linux64 binary on i386 Ubuntu? :)18:30
sistpoty|workScottK: however lintian was synced from unstable... though I guess lintian doesn't know anything about gutsy, hardy etc. any longer18:31
ScottKsistpoty|work: It does.18:31
ScottKsistpoty|work: It just looks for an ubuntu revision before it remembers18:31
ScottKsistpoty|work: Debian incorporated all the Ubuntu changes in their version.18:32
ScottKsistpoty|work: If they missed a spot, then it's a bug.18:32
sistpoty|workScottK: oh, cool.. didn't know that :)18:32
ScottKNow for backports, it'd be useful to have lintian look for either ubuntu or an ubuntu release name in the revision.18:33
\shseb128, one sync you forgot (it's also new to ubuntu but in debian): bug #184389 :)18:33
ubotuLaunchpad bug 184389 in ubuntu "Sync liboglappth from debian unstable" [Undecided,Confirmed] https://launchpad.net/bugs/18438918:33
seb128\sh: synced18:38
\shseb128, thx a lot :)18:38
seb128you are welcome18:39
Keybukmpt: it's probably easier to separate off the kernel work into a spec we can give to that team, and depending on it for various pieces18:40
mptKeybuk, ok18:41
mpt(though it's only two sentences or so)18:41
jdstrandmsg NICKSERV help18:43
jdstrandoops...18:43
* zul hands jdstrand a "/"18:49
jdstrandthanks zul!18:50
CarlFKsomething changed in the last 12 hours to cause Jan 23 19:14:40 main-menu[3081]: WARNING **: Configuring 'grub-installer' failed with error code 1   http://dev.personnelware.com/carl/temp/Jan23/b/log/syslog19:18
CarlFKcjwatson: ping ^^^19:19
=== ln-_ is now known as ln-
shayaI'm wondering who is in charge of apt?19:27
shayaor who knows its internals well?19:27
jdongshaya: ask your question and see if we know the answer?19:28
shayawant to know if there's some place I can look to see how apt does dependency management19:28
shayai.e. determining which set of packages is "good"19:29
shayafor research project, don't want to reinvent the wheel19:29
shayawant to credit debian/apt for the work19:29
ToyKeepershaya: You probably should just get the source and start reading...  with any luck, it may include enough documentation to answer your questions, but if not, the code shouldn't be too bad.19:42
desrtdearest #ubuntu-devel, plz upload new flashplugin-nonfree package to gutsy!!20:18
desrtmy fingers are getting sore from explaining how to fix this to everyone20:19
lamontslangasek: is it to late to ask for a sync of git-core from sid?20:24
rzrdesrt: ask louder ..20:28
Kmoslamont: i think the feature freeze is only near 4th february20:30
ScottK2Kmos: 14 Feb.20:30
Kmosor that :)20:30
Kmosso it now has a fixed date, nice.20:30
ScottK2Kmos: It's been 14 Feb for a very long time.20:30
KmosScottK2: ok20:31
Kmosthx20:31
matteohi all21:28
matteowhere should I make a proposal about many packages?21:28
matteoi'll explain21:28
matteonewer dpkg supports other compression modes other than gzip21:28
matteoit supports bzip2 and lzma21:29
matteousing lzma with packages with docs will help a lot to riduce package size21:29
matteosince text compresses very well21:29
ion_I’m quite sure dpkg and apt in Ubuntu support lzma.21:29
jdongion_: isn't that a hardy thing though?21:30
ion_Most likely.21:30
matteoyes, it's new in hardy21:31
=== \sh is now known as \sh_away
matteobtw compressing it's very slow21:42
matteobut it's done once, in the buildd21:43
matteohttp://pastebin.ca/87017921:43
matteo~4MB saved on 2621:44
matteoexactly 15%21:44
infinitymatteo: No need to "propose" it, we added the support for a reason.21:44
matteo"we"?21:45
infinitymatteo: But we need to finish up infrastructure support before we start compressing anything with lzma.21:45
matteoare you in the core team?21:45
jdongmatteo: he's buildd admin :)21:45
infinitymatteo: Err, I run the buildds, among other things.21:45
matteocool21:46
matteoman bc21:47
matteooops21:47
matteoit's good to hear that the ubuntu team often change things21:48
matteonot being very conservative a-la-debian21:49
matteoalso, the slowdown with bzip2 isn't too much21:49
infinityDebian's not all that conservative, per se, they just have more people to convince every time something changes.21:49
matteoso it could became the default compress mode21:49
Mithrandirmatteo: uh, no.  It's much, much slower.21:49
matteoMithrandir: twice21:50
jdongmatteo: twice isn't much?21:50
jdongmatteo: I guess when compared to lzma ultra, sure...21:50
Mithrandirit's not like people aren't complaining about unpack speed already..21:50
matteojdong: imagine how many things you can put in the ubuntu CD21:50
jdongMithrandir: lzma's unpack is in line with gzip IIRC21:50
matteounpack is ever fast21:50
jdongdefinitely advertised by its author to be faster than bzip221:50
Mithrandirjdong: it needs more memory or something, I thought?21:50
matteojust compressing is slow21:50
jdongMithrandir: no, less memory than bzip221:51
jdongMithrandir: just compressing is a BEAST.21:51
matteobut compression is made *once*21:51
jdongright21:51
jdongand we have control over the compression hardware21:51
matteothe buildd make it, the it's just a mirror push21:51
infinityDecompression of lzma is still memory-intensive, comparatively.21:51
mantiena-baltixhi all21:51
mantiena-baltixlool: hi21:51
infinityUsers with 4GB machines might not care, people with more conservative setups could be upset if we changed the default compression (so we won't)21:52
* Mithrandir yawns and wanders off to bed. See you all tomorrow.21:52
infinityThe whole plan was to first change all the bzip2 packages to lzma (since that's an obvious win), then change some others where the space saving is too good to pass up.21:52
* sladen really should just do a modified version of gzip with larger dictionary support. It would have 95% of the cases we actually want21:52
jdonginfinity: you're right, lzma levels 5 and higher are significantl more memory intensive then gzip21:53
matteoqt4-docs needs 34MB to uncompress with LZMA21:53
matteothat's on my system21:53
sladenlzma decompress (_not_ compress) is basically  dictionary size + extra21:53
jdongmatteo: if we don't use level 9 to compress, it'd be a lot less21:53
matteoso it's definitely not a memory issue21:53
jdongmatteo: hold on, 34MB decompression memory is not an issue?21:54
matteoimpressive, lzma unpacks faster than bz21:54
jdongI'd say anything over 5MB is an issue21:54
matteojdong: why?21:54
jdongmatteo: on a minimal-requirements system, it would easily push into swap21:54
matteoubuntu reccomends minimum 265MB ram21:54
mantiena-baltixsorry for disturbing, why you are talking about LZMA compression ?21:54
jdongmantiena-baltix: yes21:54
sladenbzip2 is ~2-4 * block size (900kB), so 4MB.  dictionary based (Deflate, LZMA) are going to win on decompression, with compression being an exp factor of dictionary size21:55
=== rzr is now known as rZr
jdongmatteo: ok, so booted to GNOME you're already using 130 or so in DE, open Firefox is 50MB, add restrited modules in there you barely have 30MB left over21:55
jdongmatteo: and we *require* 64MB21:55
jdongmatteo: we suggest 192MB21:55
jdongwe only recommend 256 for desktop effects21:56
sladenjdong: I would cap the LZMA dictionary at 4MB, 8MB tops21:56
jdongsladen: agreed, unless the package significantly benefits and is only for high-end systems anyway21:56
mantiena-baltixjdong: you answer is pretty short for question "why ?" ;)21:56
jdongsladen: what's clear is that LZMA level 1 both compresses faster and better than bzip default....21:57
jdongwell always faster, most times better21:57
sladenjdong: which is still 100x-200x the length for deflate21:57
matteohttp://pastebin.ca/87019921:58
jdongsladen: oh I'm not gonna try to compete with deflate for resource usage :D21:58
sladenjdong: that's what I'd expect.  lzma is a dictionary compressor; compression is a case of string matching21:58
sladenjdong: bzip2 is a qsort of 900,000 strings of length 900,000 bytes21:59
matteoquicksort?21:59
jdongsladen: well can you show a deflate implementation with similar ratio to lzma without the resource usage?22:00
=== Kmos_ is now known as Kmos
sladenjdong: sure.  for only for a dictionary of  < 2^1522:01
matteoyou talks so much about decompression speed, ignoring completely download speed22:01
sladenjdong: deflate has hard-coded 20 year old limits.  LZMA doesn't22:01
sladenmatteo: download *RAM* usage22:01
jdongsladen: I see. Isn't deflate with a bigger dictionary more or less lzma?22:01
matteowhat's the problem with a 30MB usage?22:02
sladenjdong: yes.  But the coding stage on the backend that saves another 1% is semi-heavy, and could be left as the plain Huffman used in Deflate22:02
matteoapt uncompresses an archive at once22:02
sladenmatteo: ...requiring 30MB to unarchive it?22:03
matteoand so?22:03
* sladen hands matteo a 128MB machine with a LiveCD22:03
mantiena-baltixI have a simple question for any ubuntu developers - why wine packages in ubuntu gutsy and hardy are more than 3x bigger comparing with same wine packages in official winehq.com repository and also with older wine package from ubuntu feisty, edgy and dapper ?22:04
_MMA_sladen: Be realistic. :P22:04
jdongmatteo: look you might not care about RAM usage just like how I am hooked up to 125MB/s internet and don't care about download speed :)22:04
jdongmatteo: but 30MB is quite extreme for decompression RAM usage22:04
sladen_MMA_: what's unrealistic about that?22:05
sladenthe problem here is that 30MB _is required_.  Otherwise decompression is _impossible_22:05
_MMA_jdong: So is having 80MB OO.o updates on dial-up.22:05
jdong_MMA_: I agree. both scenarios are unwanted22:05
sladenit is not using more RAM to make it go faster.  It is /necessary/ to have that much scrollback to copy and paste from22:05
jdong_MMA_: but 80MB download at a slow speed is POSSIBLE , while having insufficient free RAM to decompress is not very possible22:06
jdong_MMA_: it's not like if you feed it half the RAM it wants, ti'll decompress half as fast.22:06
matteoit's always possble22:06
matteoput something in the swap22:06
matteoextract22:06
matteothen restore it22:07
matteothe kernel does daily milion times22:07
matteowhy can't we do once at setup time?22:07
jdongmatteo: what about an openoffice update? You'd have an update manager, a GNOME DE, plus whatever else the user has running22:07
_MMA_Im just not always keep on holding back progress while worrying over 8 year old hardware.22:07
jdong_MMA_: well IMO dial-up is 20-year-old hardware. :P22:08
sladen_MMA_: "just works" means being compatible22:08
jdongdone.22:08
sladendamn those 20Kbit/s connections people   ...via their mobile phone22:08
_MMA_sladen: Just my opinion. You have to cut things off sometimes. Not Vista-like but at some point things should move along.22:09
_MMA_If all you have is 128MB RAM use a Alt disk and Fluxbox.22:09
sladen_MMA_: indeed, so I think that finding a trade-off using ~3MB-4MB of dictionary would be good22:09
jdong_MMA_: it doesn't matter *WHAT* disk you use, an update later can easily require that you have 30MB RAM.22:09
jdong_MMA_: and on a system like that, there's not much idle stuff you can just swap out22:10
mantiena-baltix_MMA_: lots of people have slow connection in also in these days - for example in lithuania about 50% of internet users have slower than 128 kbit/s internet speed22:10
matteosladen: i can't immagine the kernel unable to allocate 30MB ram22:11
sladenI've looked at many areas of compression within Ubuntu.   LiveCD, download-size, upgrade-minisize {size, download, time}, mirror size, ---the problem is that we can't optimised for all of them at once22:11
_MMA_jdong: Hence my opinion that our minimum required RAM is more like 256. Thats really not so much to ask.22:11
_MMA_s/is/should be22:11
matteoand RAM is very cheap22:11
sladen_MMA_: for X.  But not for a virtual server running Apache, that you might allocate 64MB to22:12
jdongmatteo: I have several special-purpose systems with 64MB RAM running Ubuntu22:12
jdongmatteo: and there's no reason why they should be unable to unpack linux-headers-generic now22:12
_MMA_sladen: Still I fell thats a corner-case.22:12
mantiena-baltix_MMA_: High Ubuntu RAM requiremens are because of putting ~35 MB of unneeded restricted modules in RAM22:12
sladenRAM does not grown on trees.  It grows in fast-eastern-Asian fabs22:13
sladen_MMA_: four corners, and you're trapping in a cage.22:13
_MMA_:)22:13
jdongmantiena-baltix: how do you propose storing them in license-compliant form? :)22:13
sladenjdong: I thought they were pruned if the hardware wasn't present after boot22:14
jdongsladen: nope22:14
jdongsladen: all three nvidia.o's, at 10MB/each or so, are in a tmpfs22:14
jdongsladen: it also adds around 2 secs to bootup22:14
mantiena-baltixjdong: there are at least 2 very simple solutions:22:16
mantiena-baltix1. compile and store in RAM only these modules, which are really needed on that hardware22:16
mantiena-baltix2. after installation restricted modules could be stored into hard disk instead of tmpfs in RAM22:16
mantiena-baltixalmost nobody will need 3 nvidia modules at once22:17
mantiena-baltixeach takes about 9 MB of RAM22:17
jdongmantiena-baltix: nobody, indeed, will use all 3 nvidia modules at once. the unused ones should certainly be removed from tmpfs22:17
jdongmantiena-baltix: our packaging doesn't even allow 3 userlands to be installed side by side22:17
sladenmantiena-baltix: (it _can't_ be stored onto the hardware disk => distribution in a linked form)22:18
mantiena-baltixso, I can't install different nvidia-glx packages if I have 2 different NVIDIA video cards on my system, which require different nvidia-glx package ?22:18
jdongmantiena-baltix: that's correct, they conflict each other22:19
matteosorry22:19
matteogot disconnected22:19
matteoi was sayng22:19
matteoi work in the openwrt project22:19
matteoand we use lzma in routers22:19
matteoboth kernel and rootfs22:19
matteoour machines has usually 8,12 or 32MB ram22:19
jdongmatteo: well at boot time, you guys have full control over what's in the RAM (namely, nothing)22:20
jdongmatteo: if we compressed the kernel with LZMA, I'd have zero complaints22:20
matteoand the rootfs?22:20
jdongmatteo: what if you guys packaged your updates in LZMA that needed a 16MB dictionary to unpack? Would the nat conntrack tables be happy?22:20
sladenjdong: solution to the Nvidia issue might be to have a initramfs hook that causes a rebuild and adds a marker for which ones could be used22:20
mantiena-baltixjdong: If nvidia-glx packages conflicts each other and, AFAIK, nvidia-glx packages also conflicts with ATI/AMD xorg-driver-fglrx package, I think I have simple solution for this situation22:23
mantiena-baltixwe don't need to link and put into tmpfs any of such modules until one of nvidia-glx or xorg-driver-fglrx packages is installed22:23
jdongmantiena-baltix: true, they can't be used simultaneously22:23
mantiena-baltixit's very simple to detect during boot time which of nvidia-glx or xorg-driver-fglrx packages is isntalled and link only needed nvidia.ko/fglrx.ko module or no modules at all if there are no nvidia-glx/xorg-driver-fglrx packages installed22:25
mantiena-baltixhow you like my idea ?22:26
bigondoes somebody know where http://paste.debian.net/47563 comes from?22:27
bigonI have reinstalled my hardy machine yesterday and I get that now22:28
bigoneverytime I install something that call scrollkeeper22:28
mantiena-baltixalso we need to check if NVIDIA/ATI hardware is on that system - one simple command: lspci |grep VGA does this job22:29
mantiena-baltixSo, could anyone tell me why wine packages in ubuntu gutsy and hardy are more than 3x bigger comparing with same wine packages in official winehq.com repository and also with older wine package from ubuntu feisty, edgy and dapper ?22:36
mantiena-baltixI think it's a bug22:36
mantiena-baltixlook at http://archive.ubuntu.com/ubuntu/pool/universe/w/wine/?C=S;O=D22:39
BlackDiamondscould some one take a look at this bug -> https://bugs.launchpad.net/ubuntu/+source/linux-wlan-ng/+bug/18517922:46
ubotuLaunchpad bug 185179 in linux-wlan-ng "card is not recognized in live session" [Undecided,New]22:46
BlackDiamondsand say if the proposed solution is possible ?22:46
mantiena-baltixBlackDiamonds: I can look - I have linux-wlan-ng supported card now :)22:47
BlackDiamondsis it usb ?22:48
mantiena-baltixyes22:48
BlackDiamondsthe current work around on the problem is to install the linux-wlan-ng package on a livecd session and then unplug, then replug the device22:48
BlackDiamondsthis is really really annoying because the chipset is quite popular and a lot of people use it22:49
mantiena-baltixyea, I meet with this problem too22:49
BlackDiamondsI have been using work arounds for ages on various linux distros and it would be nice to finially see a solution to make it "just work"22:49
BlackDiamondsthe best solution for the long term would have some one recode the drivers to make it support all of the kernel wireless extentions and then have it in the kernel but that will take a while, I hope an effort will start when the chipset is no longer being used in new cards22:50
BlackDiamondsmantiena-baltix, thank you very much for looking at the bug22:52
mantiena-baltixBlackDiamonds: Don't thank me - I'm not official ubuntu developer - I think you should make a bugreport to include linux-wlan-ng package into official ubuntu CD22:53
BlackDiamondsanother bug report ?22:53
BlackDiamondswouldn't it be best to keep it on this bug ?22:54
mantiena-baltixBlackDiamonds: I don't know - ask official ubuntu developers or search in wiki.ubuntu.com about this procedure22:56
BlackDiamondsalright thanks22:56
BlackDiamondsalso any place to get an offical ubuntu developer ?22:56
mantiena-baltixI think here22:57
=== bigon is now known as bigon`
=== Traxer is now known as Traxer|on
=== Traxer|on is now known as Traxer
Kmosif some admin here.. the apport-retracer on datacenter is down..23:38
mantiena-baltixcool23:42
* Mez points Kmos to #canonical-sysadmin23:43
KmosMez: thanks23:45
=== johanbr_ is now known as johanbr

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