[00:02] <Caesar> slangasek: good question
[00:02] <Caesar> I suspect badly
[00:15] <Keybuk> Caesar: ifupdown calls dhclient with -pf /var/run/... -lf /var/run/...
[00:16] <Keybuk>    [[ifconfig %iface% hw %hwaddress%]]
[00:16] <Keybuk>     dhclient3 [[-e IF_METRIC=%metric%]] -pf /var/run/dhclient.%iface%.pid -lf /var/lib/dhcp3/dhclient.%iface%.leases %iface% \
[00:16] <Keybuk> oh, no, the latter is /var/lib
[00:17] <Keybuk> ifupdown (0.6.7ubuntu2) dapper; urgency=low
[00:17] <Keybuk>   * Change the dhclient3 leases path to /var/lib/dhcp3 so the leases survive
[00:17] <Keybuk>     a reboot.  (Ubuntu: #18148)
[00:17] <Keybuk>  -- Scott James Remnant <scott@ubuntu.com>  Wed, 23 Nov 2005 16:11:30 +0000
[00:17] <Keybuk> #
[00:17] <Keybuk> amusing
[00:19] <Keybuk> looks like there's a bug there
[00:19] <Keybuk> NM uses /var/run anyway
[00:19] <Keybuk> so the two are out of sync
[00:20] <Caesar> Keybuk: that issue is an aside
[00:20] <Caesar> I'm more interested in the network being unconfigured when the machine switches to runlevel 2
[00:20] <Keybuk> does dhclient really fail if it can't write the lease file?
[00:20] <Caesar> The /var/lib, /var/run thing just got noticed in the process
[00:21] <Caesar> Keybuk: I haven't checked the failure behaviour yet
[00:21] <Keybuk> seems odd that it would do so, since it would try and write the lease file *after* setting the interface up
[00:21] <Caesar> Here, we have /var on the root filesystem so it's not an issue for us
[00:21] <Keybuk> we've had occasional bugs about the network being unconfigured throughout ubuntu's life
[00:21] <Keybuk> never tracked them down, because the people who can replicate it disappear
[00:22] <Caesar> But is my understanding correct? The system can exit single-user mode with the network still in a semi-configured state?
[00:22] <Keybuk> right
[00:22] <Caesar> That's... apalling
[00:23] <Caesar> I'm going to assume yanking the udev rules will return normalcy
[00:23] <Keybuk> there's a bug that we even have networking while in single-user mode
[00:23] <Keybuk> is that what you mean?
[00:23] <Caesar> Debian's single user mode is rather ridiculous
[00:23] <Keybuk> right, if you comment out the ifup rule then it'll bring networking up as before
[00:23] <Caesar> I think we'll do that
[00:23] <Keybuk> which will obviously block on dhcp in the boot process
[00:24] <Keybuk> but you probably want that for network-depending hosts
[00:24] <Caesar> Which would be most of them
[00:24] <Keybuk> at Google, sure
[00:24] <Caesar> I can't see how we're that exceptional in this regard
[00:24] <Keybuk> our default desktop install is quite the opposite, networking is generally brought up after login
[00:25] <Keybuk> and blocking the boot for 60s just because theres no network cable plugged in to your laptop is pessimal
[00:25] <Caesar> Yeah, that's the downside
[00:25] <Caesar> But we're not using NetworkManager
[00:25] <Caesar> Especially not on servers
[00:25] <Caesar> (or workstations for that matter)
[00:26] <Keybuk> why does it fail your boot though?  what are you using networking for during boot?
[00:26] <Keybuk> oh, syslog
[00:26] <Caesar> Yes
[00:26] <Keybuk> so syslog's nslookup fails?
[00:26] <Caesar> Yes
[00:26] <Caesar> This use case is fairly typical I'd have thought
[00:26] <Caesar> There's going to be all sorts of network-dependent things going on in runlevel 2
[00:26] <Keybuk> dunno, you're the first person in ~3 yeras to notice <g>
[00:27] <Keybuk> technically nothing in runlevel 2 should be network dependant
[00:27] <Keybuk> since you might not have a network
[00:27] <Caesar> Oh come on
[00:27] <Caesar> So where's your webserver supposed to go?
[00:27] <Keybuk> web servers aren't generally configured by hostname?
[00:27] <Keybuk> apache's conf uses ips
[00:28] <Caesar> They could have config that requires a DNS lookup
[00:28] <Caesar> You can't tell me that nothing that starts at boot shouldn't have a DNS dependency
[00:28] <Keybuk> well, nothing in the default install
[00:28] <Keybuk> otherwise the boot will fail every time, since you won't have DNS until the user logs in :)
[00:28] <Caesar> Think servers for a moment
[00:28] <Keybuk> bear in mind that there's other things that lockstep the networking
[00:29] <Keybuk> we call ifup -a in rcS
[00:29] <Keybuk> that locks out if another ifup is still running
[00:29] <Caesar> Hmm
[00:29] <Caesar> So we have an ifup that runs at S10
[00:29] <Keybuk> so if your network card is plugged in, detected and firmware uploaded before that point - you're fine
[00:30] <Caesar> It's unclear if that continues running, or fires off dhclient and returns fairly quickly
[00:30] <Keybuk> it continues running until after dhclient finishes
[00:30] <Caesar> ok
[00:30] <Keybuk> (since it does the if-up.d stuff afterwards)
[00:30] <Caesar> Of course
[00:30] <Keybuk> I last looked at this stuff for dapper, so it's possible that someone's broken it
[00:30] <Keybuk> but a quick glance says it's ok
[00:31] <Caesar> We've seen the second ifup -a in S40networking not configure eth0 because /var/run/network/ifstate claimed it already was configured
[00:31] <Keybuk> sure
[00:31] <Keybuk> you get that with a dhclient timeout
[00:31] <Keybuk> ifup can't tell the dhcp failed
[00:31] <Keybuk> or the wpa failed
[00:31] <Keybuk> etc.
[00:31] <Keybuk> it's a general design bug
[00:31] <Keybuk> ifstate only really records "someone tried to bring it up"
[00:31] <Caesar> Yeah
[00:32] <Keybuk> I have seen reports where dhcp fails during boot for no explicable reason
[00:32] <Caesar> If only we had dhclient output from the udev invocation
[00:32] <Keybuk> the logs show it apparently trying to discover an ip
[00:32] <Keybuk> and if you do it later, it works
[00:32] <Keybuk> I usually suggest init=/bin/bash, start syslog, then exec /sbin/init
[00:33] <Keybuk> (with syslog dumping to a file in /dev)
[00:33] <Caesar> Hmm, that's a new one
[00:33] <Caesar> We can try that
[00:33] <Caesar> Poor man's boot logging eh?
[00:33] <Keybuk> indeed
[00:33]  * Keybuk is a very poor man :p
[00:33] <Caesar> If only upstart did it ;-)
[00:34] <Keybuk> one day
[00:34] <Keybuk> it really is on my todo list
[00:34] <Keybuk> and I'm making forward progress through that again now :p
[00:34] <Caesar> Cool
[00:34] <Keybuk> so, from the sound of it:
[00:34] <Keybuk> ifup eth0 has run because
[00:34] <Keybuk>  - ifstate says it's up
[00:34] <Keybuk>  - ifup -a in S40networking doesn't block
[00:35] <Keybuk> if ifup hadn't run, ifstate wouldn't say it had, so S40networking would bring it up
[00:35] <Keybuk> if ifup was still running, it would hold a lock, so ifup -a would be locked out
[00:35] <Caesar> So it'd be nice to know wtf dhclient has done by the time S40networking has completed, one way or the other
[00:37] <Keybuk> also see whether commenting out the udev rule helps or not
[00:37] <Keybuk> if it does, it implies there's something amiss with when dhcp is running
[00:37] <Keybuk> but I can't think what that might be
[00:38]  * Keybuk heads for bed (which is where I was on my way to :p)
[00:38] <Caesar> Okay, I'll try and let you know
[00:38] <Caesar> Goodnight!
[01:25] <ScottK> slangasek: if something's wanting to get promoted to Main only because it's an alternate depends (and not even the first one listed) and we don't otherwise want it in Main, is adding it to extra-excludes in the appropriate see the correct solution?
[01:58] <bbs> hey, i'm writing an initscript for a program that my company is writing ... we are going to package it very nicely for ubuntu because i said you guys are where a lot of heat is at right now -- personally i use source based distros (how silly of me i know :) ) -- but i was wondering where a good guide for writing initialization scripts (upstart -- not exactly sure how your system works) -- so that i can properly instantiate things that are required for a p
[01:58] <bbs> any links?
[01:58] <bbs> or documentation?
[02:00] <sommer> bbs: you might try: http://upstart.ubuntu.com/getting-started.html
[02:02] <bbs> does that go over discussion of creating pid files?
[02:02] <bbs> sommer: i'm looking to really support ubuntu the best most likely
[02:10] <slangasek> ScottK: bomber kapman killbots promoted, thanks
[02:11] <slangasek> stgraber: the update-manager one sounds transient; the file-roller one is resolved for the next run; and the xserver one is in tjaalton's hands
[02:11] <ScottK> slangasek: Great.  Thanks.  A little furhter along you'll run into a request for kdeedu-kvtml-data to get promoted to fix kdeedu.
[02:11] <slangasek> ScottK: yep, was just about to do that one now :)
[02:12] <ScottK> slangasek: Great.
[02:12]  * ScottK will wait for you to get to his seeds question in due course then.
[02:14] <slangasek> Keybuk: er, Caesar is most certainly not "the first person in three years" to notice network-dependent stuff going on in runlevel 2
[02:15] <slangasek> I've just been holding my tongue and praying we can migrate to upstart soon
[02:15] <slangasek> s/upstart/& jobs/
[02:16] <slangasek> Keybuk: also, your definition of runlevel 2 is de facto wrong for Debian, so yeah.
[02:17] <Chipzz> slangasek: I've alwyas wondered why the usage of runlevels differed in debian anyway
[02:17] <slangasek> because they differ everywhere, because they're not a standard?
[02:18] <Chipzz> let me put it otherwise: why they differed fmor the de facto standard, being redhat
[02:18] <Chipzz> (de facto standard: 3 -> text; 5 -> graphical login)
[02:19] <StevenK> And why is RedHat is the de facto standard?
[02:19] <slangasek> ScottK: extra-excludes> hmm, I would've expected germinate to be smart enough to sort that out on its own?
[02:19] <slangasek> or maybe not
[02:19] <ScottK> slangasek: The package in question is kdesvn.
[02:19] <Chipzz> because the way redhat did it, like it or not, is how most other distro's did it. and the fact that redhat has long been one of the most popular distro's
[02:20] <ScottK> It's listed on component mismatches and as nearly as I can determine only becuase one of it's binary packages is an alternate depends for kdesdk.
[02:20]  * slangasek throws bits of rpm at Chipzz ;P
[02:21] <Chipzz> to my knowledge, but I may be mistaken on that, debian and ubuntu are the only significant distro's that have 2 as the equivalent runlevel of what redhat et use for 3
[02:21] <Chipzz> (well, given the different way a grpahical login is started that's not 100% true either, but...)
[02:21] <Chipzz> s
[02:22]  * Chipzz chews on them and burps ;P
[02:22] <Chipzz> (slangasek: as an aside: I have done lots of rpm packaging in a different life ;))
[02:22]  * ScottK will go out on a limb and guess the other Debian derivatives do it similarly.
[02:22] <slangasek> I have, too; you'll note what I do now instead
[02:22] <StevenK> Hah
[02:23] <Chipzz> ScottK: yeah, but can you call those debian derivatives significant? :P
[02:23] <pwnguin> what about the embedded debian derivatives?
[02:24] <Chipzz> pwnguin: isn't emdebian (or whatever it's called) a subproject of debian?
[02:24] <pwnguin> Chipzz: there's others
[02:24] <Chipzz> afaik it's maintained by a debian maintainer
[02:24] <pwnguin> angstrom
[02:25] <pwnguin> openembedded
[02:28] <slangasek> ScottK: yeah, I guess excluding that package is the only way to clear it out of the mismatches, doh
[02:29] <slangasek> Chipzz: it's maintained by a Debian developer; that doesn't mean it's a subproject of Debian, it's not really hosted using Debian resources...
[04:58] <ScottK> slangasek: If you're still around ... Need binary promotion for kpartloader to fix kdesdk installability.
[04:58] <slangasek> done
[04:59] <ScottK> slangasek: Thanks.
[05:39] <ScottK> slangasek: I think python-plasma is the last binary promotion Kubuntu CD/DvD are needing.
[05:45] <ScottK> Nope.  I'm wrong.
[05:46] <ScottK> slangasek: Need libkdcraw7 libkdcraw7-dev libkexiv2-7 libkexiv2-7-dev too.
[05:47] <NCommander> oooh
[05:47] <NCommander> I think ...
[05:47] <NCommander> I think PyKDE4 built ...
[05:49] <NCommander> ahaha
[05:49] <NCommander> AWESOME
[05:49] <NCommander> It builds!
[05:49] <ScottK> Congratulations.
[05:49] <NCommander> I just need to make cmake work properly now
[05:49] <NCommander> and we're fully in business
[05:49] <pitti> Good morning
[05:49] <NCommander> hey Pici
[05:49] <NCommander> *pitti
[05:50] <pitti> hey NCommander
[05:50] <NCommander> I just about got kde4bindings building on arm
[05:50] <NCommander> wooo
[05:51] <StevenK> Morning pitti
[05:51] <pitti> slangasek, LaserJock: we currently don't have a vehicle for moving the GNOME help files from packages to langpacks, UDS discussion was insufficient; I plan to continue it with the soyuz team via email
[05:51] <StevenK> pitti: It's only 6:50am! \o/
[05:51] <ScottK> Good monring pitti.
[05:52] <ScottK> morning even.
[05:52] <pitti> StevenK: hey, at least I slept through until the alarm went off at 6 (my wife has to get up at that time)
[05:52] <StevenK> pitti: Woot! Progress!
[05:53] <ScottK> pitti: Would you be up for some binary promotions?  slangasek seems to have dropped off.
[05:54] <ScottK> python-plasma libkdcraw7 libkdcraw7-dev libkexiv2-7 libkexiv2-7-dev are the remaining ones needed to get KDE installable for the Alpha milestone.
[05:55]  * ScottK also casually mentions that there are two MIR waiting for ubuntu-mir needed to get lintian installable again ...
[05:57]  * StevenK NBS's out libboost1.36, half tempted to NEW it, just so he can remove it again
[06:00]  * ScottK notes the time in his TZ and decides to go to bed.
[06:00] <ScottK> Good night all.
[06:02] <pitti> ScottK: good night!
[06:05] <mahesh> i compiled a fresh kernel and then now want to add features like.... automounting the any scsi devices..... i need help.....!
[06:09] <LaserJock> pitti: so the plan would be to have binarymangler pull out all the C/ .xml files and put them on LP somewhere and then put them in the langpacks?
[06:10] <pitti> LaserJock: well, everything *except* C, but yes
[06:11] <LaserJock> oh, sorry
[06:16] <StevenK> -- Found FFMPEG INCLUDES: LIBFFMPEG_INCLUDE_DIR-NOTFOUND
[06:16] <StevenK> kdenlive, you fail
[06:17] <LaserJock> pitti: but the plan isn't to be translating though? my experience of xml2po, etc. are that I wouldn't rely on them for complete automation
[06:17] <ScottK> StevenK: Which version are you looking at?
[06:18] <StevenK> ScottK: 0.5.svn20071228-0.0ubuntu3
[06:19] <ScottK> There's a 0.7 that someone is working on packaging.
[06:19] <StevenK> Then I'll stop caring, since I'm only caring for NBS
[06:20] <ebroder> Any bored backporters I could get to ACK on LP #301283, LP #305001, or LP #308800?
[06:20] <StevenK> ScottK: Apparently, you were sleeping
[06:21] <ScottK> Not quite yet.  I'm slow.
[06:21] <mahesh> i compiled a fresh kernel and then now want to add features like.... automounting the any scsi devices..... i need help.....!
[06:21] <ScottK> ebroder: Did anyone answer the rdepends testing question on sqlalchemy?
[06:22] <ebroder> ScottK: I...didn't know there was an rdepends testing question
[06:23] <ebroder> ScottK: Give me a second to look at what the rdepends are
[06:25] <ScottK> ebroder: In the bug now.
[06:31] <ebroder> ScottK: Would you consider it sufficient to just test against the versions in Hardy?
[06:31] <calc> pitti: got them uploaded :)
[06:32] <ScottK> ebroder: yes.  If those work, it's great.  If they don't, they need to be backported too.
[06:32] <pitti> calc: \o/
[06:32] <mahesh> i compiled a fresh kernel and then now want to add features like.... automounting the any scsi devices..... i need help.....!
[06:34] <calc> pitti: also stripped out ubuntu specific bits and created debdiffs and filed the patches in debian bts :) heh
[06:34] <ScottK> mahesh: This isn't a help channel.  Asking the same question every 10 minutes doesn't change that.  Ask in #ubuntu.
[06:34] <pitti> calc: appreciated, so let's hope we can sync them again soon
[06:35] <calc> of course as debian is in freeze it may take them a while to upload with the patches
[06:39] <mahesh> ScottK>> i thought some one would help.... #ubuntu is for end-user support
[06:43] <calc> hmm my xom build failed on the buildd, worked fine on my local box
[06:43]  * calc is going to upgrade his chroot and verify it still works locally
[06:50] <calc> oh this failure is ugly it worked in my non-clean chroot
[06:50] <calc> when i cleaned it and rebuilt it fails :\
[06:50] <calc> so apparently it needs something else
[06:50] <calc>     [javac] The system is out of resources.
[06:50] <calc>     [javac] Consult the following stack trace for details.
[06:50] <calc>     [javac] java.lang.StackOverflowError
[06:50] <calc> lovely
[06:51] <StevenK> calc: More RAM? :-P
[06:51] <calc> heh
[06:51] <calc> not likely the problem
[06:52] <calc> i just wish i knew what i had installed previously that allowed it to build ok
[06:52] <calc> hehe
[06:52] <calc> i found it but i don't know why it helps
[06:53] <calc> default-jdk-builddep works just default-jdk by itself does not
[06:55]  * calc uploaded fixed version
[06:56] <calc> pitti: can you process the rest of the packages once they hit the archive in an hour or so? :)
[06:57] <pitti> calc: right, -builddep is for building, default-jdk-headless is for binary dependencies
[06:57] <pitti> calc: I'll try (however, note that I'll block at least one pacakge, since we are in alpha-2 freeze)
[06:59] <calc> pitti: well default-jdk vs default-jdk-builddep (builddep was supposed to only be needed when building gcj packages)
[06:59] <calc> pitti: ok
[06:59] <pitti> oh?
[06:59] <pitti> calc: anyway, are you using -headless? for most libraries that should suffice
[07:00] <calc> yes for the runtime dep
[07:00] <calc> " - build dependencies: If a source package builds a "-gcj" package, don't
[07:00] <calc>    stop building it yet. This will slow down the gij runtime. Instead,
[07:00] <calc>    use "default-jdk-builddep", which depends on the default jdk and
[07:00] <calc>    java-gcj-compat-dev. If no "-gcj" package is built, use default-jdk
[07:00] <calc>    as a build dependency."
[07:00] <pitti> hm, ok
[07:00] <calc> that was from doko's email, sorry for the crummy line-breaks
[07:02]  * calc headed to bed now, bbl
[07:03] <pitti> calc: sleep well!
[07:09] <dholbach> good morning
[07:13] <LaserJock> morning dholbach
[07:13] <dholbach> hi LaserJock
[07:24] <highvoltage> morning LaserJock and dholbach
[07:25] <dholbach> hi highvoltage
[07:30]  * LaserJock waves to highvoltage 
[07:38] <highvoltage> LaserJock: did you get a chance to go by to the UDS?
[07:45] <LaserJock> highvoltage: no, just did some remote stuff
[07:46] <pitti> stgraber: yay MOTU! congratulations!
[08:18] <pitti> hi tkamppeter
[08:25] <LeviTheSmith2> THANK YOU!
[08:26] <LeviTheSmith2> I'll just copy and paste what I wrote in the other channel
[08:26] <LeviTheSmith2> My old laptop I got didn't work well with Windows(function keys/quick keys didnt work), Fedora(same as windows, plus wireless and audio didnt work), Solaris(same as fedora). Ubuntu: Audio worked after the first boot - no need to install drivers. Function keys work - i can turn the volume done now!. Wireless worked out of the box too! :D
[08:26] <LeviTheSmith2> s/done/down
[08:27] <LeviTheSmith2> If anyone's wondering, the model is Asus A2H 2.66ghz
[08:27] <tkamppeter> pitti, hi
[08:57] <lool> pitti: I didn't get very clear feedback on libv4l versions fixing what or breaking what (didn't get any regression report so far I think), and I was tempted to stop with the 0.5.6 SRU, but the last comment on LP #286070 encourages me to look at SRU-ing 0.5.7; the diff is minimal and fixes a missing argument in a syscall
[08:58] <lool> pitti: What do you think of http://people.ubuntu.com/~lool/libv4l-0.5.7.diff ?
[08:59] <lool> It a) adds debug code b) handles more error cases nicely c) fixes the number of args to SYS_mmap2
[09:00] <NCommander> hey lool
[09:01] <lool> hey NCommander, how is it going?
[09:01] <NCommander> lool, finally fixed kde4bindings on ARM
[09:02] <NCommander> It was a combination of five bugs that had to be fixed to get it going
[09:02] <lool> NCommander: I read yesterday that you found a broken configure check
[09:02] <NCommander> yeah
[09:02] <NCommander> About the only thing that WASN'T broken was sip -_-;
[09:02] <NCommander> go figure
[09:03] <lool> NCommander: When do you plan to upload this?  I'm happy to sponsor ASAP to get more armel stuff building
[09:03] <NCommander> We're in freeze
[09:03] <NCommander> So Friday assuming alpha 2 releases on time
[09:03] <lool> You've got sponsoring in place for this?
[09:03] <NCommander> (all the patches are on launchpad and/or in the appropriate bazaar branches)
[09:03] <NCommander> ScottK, apachelogger or Riddell will do it
[09:04]  * NCommander is using the opportunity to rebuild all the KDE packages again from scratch just to make sure there are no overlooked OOPS
[09:04] <lool> Cool
[09:04] <NCommander> I'm also sending the necessary patchsets to the apporiate upstreams
[09:05] <lool> Great; would you mind pointing me at the corresponding upstream bugs when these are filed?  I'd like to subscribe to them
[09:05] <NCommander> I can't find a bugtracker for PyQT4
[09:05] <NCommander> so that's just going to be an email to their mailing list
[09:05] <NCommander> lool, https://bugs.kde.org/show_bug.cgi?id=177965 - theres one of the KDE ones, I have to file two more to get all the ARM patches in
[09:06] <directhex> NCommander, any issues in the cli bindings on arm?
[09:06] <NCommander> None to my knowledge
[09:06] <NCommander> But I only built pykde4 (which takes two hours by itself)
[09:06] <NCommander> So I'm doing a full rebuild to make sure there is nothing broken
[09:06] <directhex> yeah, kdebindings is bad enough on a 2+ghz core2
[09:07] <NCommander> Go 233MHz ARM :-)!
[09:07]  * NCommander actually has to wait for kde4libs to finish building since I needed to change parts of its build system
[09:09] <directhex> i wonder whether the kubuntu people applied my cli plasma examples fixes
[09:09] <tjaalton> pitti: hey, should jockey not offer to install nvidia/fglrx before they are fixed to support the new ABI, or how should it be handled
[09:10] <pitti> lool: doesn't look too bad, should get proper testing, though
[09:11] <pitti> tjaalton: I can temporarily disable it if you want
[09:11] <lool> pitti: Naturally
[09:11] <NCommander> directhex, nope, we disabled the examples
[09:11] <lool> pitti: I did test it in Debian and in Ubuntu before pushing to my ppa
[09:11] <lool> pitti: I'll open a new SRU bug for this, thanks
[09:12] <tjaalton> pitti: I believe it should be disabled, although apt would refuse to install them too
[09:13] <directhex> NCommander, that's option 2, of course. it seems some silly bugger in upstream changed a property into a method, so foo.Size became foo.Size() (which is what the 2 patches changed)
[09:13] <NCommander> directhex, blame the guy who did kde4bindings
[09:14]  * NCommander would include the patch if it was risk free for breaking ARM builds
[09:15] <directhex> NCommander, it's adding six brackets between two .cs files. i imagine the risk of regression is small - unless there are other issues building the examples which didn't show up on amd64
[09:15]  * NCommander just got kde4bindings working on ARM
[09:15]  * NCommander has no desire to touch it until after its uploaded and ACCEPTED
[09:18] <lool> What's the syntax in launchpad to link to a comment in another bug again?  I thought it was bug #1234#c12 but it's not
[09:18] <lool> ups
[09:23] <directhex> NCommander, i posted both patches with a 72h lifetime, so can you save them someplace?
[09:23] <NCommander> directhex, attach them to a bug against kde4bindings, and set the milestone to alpha 3
[09:28] <directhex> NCommander, do i have the ability to set milestones?
[09:28] <NCommander> oh right
[09:28] <NCommander> your not in bugcontrol
[09:28] <NCommander> Just link me to the bug, and I'll set the milestones and such
[09:28] <directhex> bug 308891
[09:30] <pitti> tjaalton: uploaded
[09:30] <tjaalton> pitti: thanks
[09:57] <tseliot> mvo, seb128: gnome-control-center and gnome-desktop build again (and actually work too) in Jaunty now.
[09:57] <mvo> Keybuk: did you had a chance to look at the sponsor request for #304774 (sysklogd)? I will take it unless you have some objections
[10:00] <tseliot> mvo: can you merge from lp:~albertomilone/gnome-desktop/tseliot-fixes and from lp:~albertomilone/gnome-control-center/tseliot-fixes ?
[10:01] <mvo> tseliot: thanks a lot! will do the merge now
[10:01] <tseliot> mvo: good
[10:02] <mvo> cjwatson_: could you please add me (temprarely) to the debian-installer team so that I can push a (trivial) patch against the desktop file of usb-creator (bug #286924)
[10:20] <mvo> tseliot: merged and pushed, thanks again!
[10:21] <tseliot> mvo: great, thanks :-)
[10:28] <cjwatson> mvo: ubuntu-installer you mean? done
[10:30] <mvo> cjwatson: thanks
[10:40] <emgent> norsetto: \o/
[10:41] <norsetto> emgent: ah, somebody alive! Hi emgent
[11:45] <asac> hmm ... does the accessibility toolkit we ship in gnome by default have a name? how is it called?
[11:48] <pitti> ATK?
[12:04] <fabbione> soren: if the server team is interested, i just released redhat-cluster 3.0.0-alpha1 upstream.
[12:08] <soren> fabbione: Cool.
[12:16] <asac> pitti: ok ... isnt that an abstract layer on top of specific implementations?
[12:17] <pitti> asac: maybe it is a spec, too, but it's literally called atk1.0 (source package)
[12:17] <asac> ok thanks
[12:18] <asac> pitti: did we turn this on by default in intrepid or was that on foreever?
[12:19] <pitti> asac: it's a basic property of GTK, but some features aren't enabled by default
[12:19] <pitti> asac: see "Hilfstechnologien" in System -> Einstellungen
[12:20] <asac> pitti: yeah i know. just wonder if we enabled more features by default since intrepid
[12:20] <asac> there is a firefox bug which happens much more reliably in intrepid than in hardy
[12:20] <asac> which could be related to atk
[12:46] <stgraber> pitti: thanks
[12:48] <ogra> stgraber, so now on to the main upload permission :)
[12:48] <ogra> sadly TB meeint is only in a week
[12:48] <ogra> *meeting
[12:50] <stgraber> ogra: yep :) btw, I'll likely have a new ldm and ltsp to upload today, I'm doing some testing and trying to fix that ldm-ubuntu-theme thing (other than just reverting to the old gtk)
[12:51] <ogra> stgraber, did you just copy the human gtkrc ?
[12:52] <ogra> actually that shouldnt even be needed ... the old python code had an automatic fallback to use the system gtkrc, i wonder why that doesnt happen anymoe
[12:52] <stgraber> ogra: at the moment I just replaced ubuntu-theme by gtk2-engines-ubuntulooks
[12:53] <ogra> ubuntulooks is dead
[12:53] <ogra> if you upload that ldm will pull it back on the CD
[12:53] <ogra> so that wont work
[12:54] <stgraber> hmm, now that I think of it, is ubuntu-theme also supposed to work on Intrepid ?
[12:54] <ogra> yes, its the default
[12:54] <ogra> and it uses murrine
[12:55] <ogra> just copy /usr/share/themes/Human/gtk-2.0/gtkrc into the ldm-theme-ubuntu package
[12:55] <ogra> that should make it work
[12:55] <stgraber> ok, so the problem isn't that I'm testing with a backport on Intrepid ... (I'm still waiting for an instalable jaunty)
[12:57] <ogra> do you have murrine installed in the chroot ?
[12:57] <ogra> the intrepid package didnt pull it in ... (ldm should still fall back to the default gtkrc though)
[13:09] <stgraber> ogra: ah, ok so I'll try installing murrine by hand (I really need a Jaunty box ...)
[13:09] <ogra> just do a vbox intrepid install and upgrade ?
[13:11] <stgraber> well, I'm just waiting for ubuntu-desktop to actually install, then I'll reinstall my laptop (it was last installed with Intrepid alpha-1)
[13:14] <ogra> oh
[13:14] <ogra> you dont upgrade by final release ?
[14:16] <cinnamon> i hope this is the right place to ask. i think it is not good that bugs in ubuntu launchpad time out, when not fixed. who made this decision and why?
[14:20] <cjwatson> they don't actually (at the moment), it just says they do
[14:21] <cjwatson> https://help.launchpad.net/Bugs/Expiry
[14:22] <cjwatson> the implementation was more a Launchpad decision than one of ours, although our bug team did comment on it
[14:29] <cinnamon> thanks cjwatson
[14:46] <rudchenkos> hi there. could anyone point me to an info how to use debug symbols from a *-dbg package?
[14:47] <Keybuk> gdb will pick them up automatically
[14:49] <rudchenkos> Keybuk, oh, it knows the "/usr/lib/debug/" path?
[14:59] <Notch-1> hi, i'm looking for the persistent/casper mode developers, can somebody help me?
[15:10] <luckyone> Dear ubuntu-devel, The new 64-bit package for pulseaudio that hit repos last night is such an amazing improvement. I want to sincerely thank everyone who played a part in putting it together. Best regards, luckyone!
[15:12] <ogra> luckyone, send flowers to TheMuso (once he shows up)
[15:18] <DktrKranz> ogra: Debian recently used cookies to hug people :)
[15:19] <ogra> so we'll see a lot fat DDs soon ?
[15:19] <DktrKranz> probably
[15:20] <DktrKranz> especially if they plan to release before 2009!
[15:26] <Mez> ogra: there'll be one mopre fat DD whenever FD get around to approving me ;)
[15:27] <ogra> haha
[16:31]  * steron is away (will be back !)
[16:32] <stgraber> ogra: I copied /usr/share/themes/Human/gtk-2.0/gtkrc to greeter-gtkrc and now it works :)
[16:32] <ogra> great :)
[16:35] <stgraber> ogra: hmm, I see half of the ldm themes use gartoon and the other half use Human, which one should we use ? (icon theme)
[16:35] <ogra> Human
[16:36] <ogra> only keep gartoon for Edubuntu
[16:36] <ogra> (we'll likely need to fiddle with deps here)
[16:47] <mok0> stgraber: congrats!
[16:48] <stgraber> mok0: thanks
[16:48] <stgraber> ogra: ok
[16:50] <sianis> hi
[16:50] <kelemengabor> hi sianis :)
[16:50] <sianis> hi kelemengabor
[16:52]  * Keybuk just can't get git fast-export|bzr fast-import to work on the udev tree :-/
[16:55] <mvo> hello kelemengabor, sianis
[16:55] <kelemengabor> sianis: here is my idea: we need to archive the ddtp po-files in bzr at the time of changing the translated distribution in rosetta
[16:55] <calc> pitti: do you have time to review the uploads i did last night wrt the mir?
[16:56] <kelemengabor> translators will translate that, and we export that
[16:56] <kelemengabor> then merge it to the archived translations of current-1
[16:56] <kelemengabor> and not to the pot of current-1
[16:57] <mvo> kelemengabor: we could create projects in rosetta for each release, not sure if rosetta would be able to handle that
[16:57] <flowolf> hi all
[16:57] <kelemengabor> because we would lose changed strings
[16:57] <flowolf> I just noticed that many -dev packages got removed from older ubuntu releases
[16:57] <mvo> kelemengabor: this would give us the "suggest strings from other releases" feature as well
[16:57] <flowolf> what's the reason for this?
[16:57] <kelemengabor> mvo: yeah, but this is unnecessary
[16:57] <mvo> kelemengabor: ok, even better
[16:58] <kelemengabor> mvo: suggest strings is great, but not for 1000 strings :)
[16:58] <flowolf> and is it possible to have them back?
[16:58] <kelemengabor> the main thing is that translators would ideally have nothing to do
[16:58] <calc> flowolf: mentioning at least one of the packages by name might help someone determine the cause
[16:58] <kelemengabor> to see their translations in older releases
[16:58] <flowolf> calc: http://packages.ubuntu.com/feisty/i386/build-essential/download
[16:59] <kelemengabor> we should merge back automagically
[16:59] <flowolf> calc: try to download this
[16:59] <calc> flowolf: ah yea feisty is dead
[16:59] <calc> flowolf: you didn't see the notice go out a while back?
[16:59] <kelemengabor> and take care to not to lose anything that already existed
[16:59] <calc> "Ubuntu 7.04 reaches end-of-life on October 19, 2008"
[16:59] <flowolf> calc: so you removed the -dev packages ?
[17:00] <calc> flowolf: i didn't but the archive admin probably removed the entirety of feisty
[17:00] <flowolf> nope, feasty is out there
[17:00] <kelemengabor> because packages can move between repositories, and strings can change, we need to keep existing po-files in bzr,
[17:00] <flowolf> you can still download and install "regular" packages
[17:01] <flowolf> it's only the -dev packages, like build-essential or libc6-dev that disappeared
[17:01] <calc> flowolf: ah ok, well i have no idea but since it is completely unsupported now i wouldn't be surprised if they removed bits of it
[17:01] <kelemengabor> and at updates, create a compendium file from newly exported files
[17:01] <flowolf> calc: still the packages are still there, http://it.archive.ubuntu.com/ubuntu/pool/main/b/build-essential/
[17:01] <kelemengabor> and merge that compendium to the po files
[17:01] <flowolf> it is only their names to be wrong in the package file
[17:02] <calc> http://packages.ubuntu.com/feisty/i386/build-essential/download <- that page comes up to the mirror list
[17:02] <flowolf> so you guys broke development packages on purpose for some misterious reason...
[17:02] <kelemengabor> this will protect us against strings ﻿moved from one po file to the other
[17:03] <Keybuk> cjwatson: didn't you have some patches to make bzr fast-import work?
[17:04] <cjwatson> Keybuk: I don't think udev worked for me either so they probably aren't relevant
[17:04] <kelemengabor> I see this is rather complicated and resource-wasting, but should be a little more elegant than just using the current po files for all the older releases
[17:04] <cjwatson> flowolf: no, you need to use old-releases.ubuntu.com if you're using feisty
[17:04] <calc> flowolf: hmm actually no its gone completely, 11.3ubuntu1 was in gutsy which is why that package is still on the site (i think)
[17:04] <calc> flowolf: plain 11.3 was what was in edgy/feisty
[17:05] <cjwatson> flowolf: sorry but we simply don't have space to keep feisty on the standard mirror hierarchy
[17:05] <Keybuk> tbh, git fast-export crashes with the udev repository
[17:05] <Keybuk> so I'm probably doomed anyway
[17:05] <flowolf> cjwatson: that's good enough
[17:06] <mvo> kelemengabor: the apt-ddtp-tools can generate Translation-$lang -> po -> Translation-$lang so the current (latest) translation status in the archive can be used to create po/pot files
[17:07] <flowolf> it is working
[17:07] <flowolf> cjwatson: thank you
[17:07]  * calc forgot that it went somewhere else when it was removed :)
[17:07] <kelemengabor> mvo: cool, but using the current po files for older releases will lose some data, unfortunately
[17:08] <mvo> kelemengabor: true, it would have to be a merge of old and new I guess
[17:09] <kelemengabor> yeah, this is why I mentioned storing the po files in bzr
[17:09] <mvo> kelemengabor: I need to go for dinner, I will be back in ~30min, ok? then we can talk more about it. let me know if you need anything in apt-ddtp-tools itself
[17:09] <kelemengabor> ok, good appetite
[18:07] <slangasek> directhex: working with meebey, I have mono-addins fixed up now for the 2.0 transition; this looks like it's going to be the best way to get us down to size on CDs for alpha-2 after all, so I'm going to follow through on it today
[18:22] <Baby> do you know which man page section would correspond to a tcl tk command?
[18:22] <slangasek> Riddell, ScottK: kubuntu ISOs are still very much oversized, is there any hope of bringing that down in time for alpha-2?
[18:30] <ScottK> Dunno.  Let me see what I can come up with.
[18:52] <ScottK> slangasek: There are a couple of changes I'm confortable with making.  Doing then now.
[18:52] <slangasek> ScottK: uploads, or seed changes?
[18:52] <slangasek> (i.e., how soon should I try rebuilding)
[18:52] <ScottK> seed changes.
[18:53] <ScottK> slangasek: You need a new kubuntu-meta upload too, right?
[18:53] <ScottK> So the stuff gets dropped from kubuntu-desktop
[18:53] <slangasek> yeah
[18:54]  * ScottK notes that Riddell has been on vacation for both Alpha's so far this cycle.  Very convenient.
[18:56] <ogra> he would have a hard time testing images where he is now
[18:58] <ScottK> True.
[19:00] <hyperair> jcastro: hey you around?
[19:01]  * jdong wonders if a digital camera would work well as a scanner.
[19:01] <hyperair> jdong: try super macro mode. some cameras work well enough
[19:03] <jdong> WOW
[19:03] <jdong> doesn't look bad at all
[19:03] <jdong> (powershot sd1100, i.e. ixus 82)
[19:04]  * hyperair chuckles
[19:09] <ScottK> slangasek: I'm trying to get my kubuntu-meta update out the door before I have to go get kids at school.  Not sure if I'm going to make it.
[19:09] <ScottK> slangasek: Is there a convenient place I can see how far off we are on CD size?
[19:13] <Gadi> can someone here tell me whether ssh grabs bits from /dev/urandom only upon the initiation of the session or whether it is constantly pulling from /dev/urandom to encrypt each packet?
[19:18] <ScottK> slangasek: Uploaded what I have.
[19:34] <slangasek> ScottK: convenient place> well, the directory index gives you the current size, and the target size is $((700 * 1024 * 1024))
[19:34] <slangasek> uploads> thanks
[19:35] <highvoltage> hey slangasek
[19:37] <slangasek> highvoltage: hello
[19:37] <ogra> whoops, what happened to http://qa.ubuntuwire.com/ftbfs/
[19:37] <ogra> no more ftbfs at all \o/
[19:38] <highvoltage> whohoo!
[19:38] <ogra> heh
[20:07] <ScottK> slangasek: How about go ahead and NBS out kdeplasma-addons-libs4.  The only remaining rdepends are transitional packages on ia64 and hppa.  Getting them really gone will take a while.
[20:16] <ScottK> LaserJock: The package kpercentage has been removed by upstream, so it needs to be removed for Edubuntu seeds/meta (it's on the NBS list currently).
[20:17] <LaserJock> ScottK: right, I've got it removed locally, just having pushed+uploaded, I'll do that nowish
[20:17] <ScottK> LaserJock: Great.  Thanks.
[20:19] <ogra> infinity, oh, did you just do another mass give back ?
[21:04] <bigon> libv4l has an higher version in intrepid than in jaunty should I report that?
[21:06] <doko> NCommander: any results with your kde4bindings experiment?
[21:07] <slangasek> ScottK: kdeplasma-addons NBSed
[21:07] <Keybuk> would I be right in thinking that there's no particular standard to the content of Vcs-Bzr ?
[21:07] <NCommander> doko, I built it on ARM :-)
[21:07] <ScottK> slangasek: Thanks.
[21:08] <Keybuk> Vcs-Bzr: lp:~ubuntu-core-dev/udev/ubuntu
[21:09] <Keybuk> would that be the "right thing" to put there?
[21:09] <pochu> Keybuk: I think it needs to be an URI
[21:09] <Keybuk> that is a URI
[21:10] <ScottK> slangasek: We have a proposed solution for Kubuntu oversize.  It'll be an hour or two before I can do it.
[21:10] <slangasek> Keybuk: to date I've spelled it out rather than using lp:, but there's no standard other than it being a URI
[21:10] <slangasek> ScottK: ok, cheers
[21:10] <Keybuk> slangasek: the advantage to the bazaar http URL is that you can stick it in a web browser as well
[21:10] <ScottK> slangasek: In the meantime, if you want to move kmail and kdepim from cd to dvd that'll also let mysql and akonadi fall off the CD.
[21:10] <slangasek> Keybuk: true
[21:11] <Keybuk> the disadvantage is that you have to rewrite it by hand to be able to *commit*
[21:11] <Keybuk> the advantage to the lp: URI is that bzr will use bzr+ssh if you can commit, http if not
[21:11] <ScottK> slangasek: If an hour or two doesn't hurt, just leave it to me.
[21:11] <pochu> Keybuk: I guess if debcheckout can use it properly, it's fine then. But I'm not sure :)
[21:12] <Keybuk> ah, maybe debcheckout handles the http/bzr-ssh transition
[21:12] <slangasek> Keybuk: well, 'debcheckout -a' knows how to rewrite the http urls; and once you've checked out from lp: you still have to rewrite by hand to turn that into a commitable url if, say, you gained commit access after you did the checkout
[21:12] <Keybuk> I'll use http then, since that's what most packages seem to do
[21:13] <slangasek> ScottK: I'll probably leave it to you, since I'm still fighting with ubuntu CD sizes
[21:13] <ScottK> OK
[21:55] <calc> ugh launchpad maintenance at 4pm, lol
[21:55] <calc> asac: do you have MIR approval access?
[21:56] <cjwatson> https://launchpad.net/~ubuntu-mir/+members
[22:05] <calc> cjwatson: thanks :)
[22:09] <slangasek> doko: do you know why pygtk was "ported" to numpy in Debian?  What was wrong with the way it was built previously?
[22:09] <slangasek> doko: (this adds a number of packages to the CDs that weren't needed, previously)
[22:11] <slangasek> doko: hmm, I guess we were using python-numeric previously, which has much simpler depends
[22:12] <pochu> slangasek: according to http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=478452, because python-num{eric,array} are unmaintained upstream
[22:13] <slangasek> but the one that is maintained upstream has bunches more deps, hmm :/
[22:14] <doko> slangasek: yes, we do want to get rid of the old numeric packages. not sure, if we can modularize numpy
[22:15] <slangasek> doko: ok.  FWIW, the switch from numeric to numpy is a net loss of ~6MB on the CDs
[22:16] <slangasek> (libblas3gf+liblapack3gf+libgfortran3, plus a size delta between numpy and numeric themselves)
[22:16] <directhex> bloaty python :/
[22:17] <doko> pochu: could you ask upstream if they can support a minimal build/split?
[22:18] <doko> slangasek: yes, we had these dependencies in python-numeric-ext, which was not on the CD
[22:18] <bryce> slangasek: https://wiki.ubuntu.com/Hotkeys added a 'should be' diagram based on a photo of mdz's whiteboard art.  Let me know if I got anything incorrect there
[22:19] <slangasek> bryce: doh, I had already been working on one of those; but I also discovered that hotkey-setup does some stuff that's outside of what can be done with hal fdi, so it probably remains for the near term
[22:20] <bryce> slangasek: ok I'll stick that back in
[22:21] <bryce> how about atkbd?  timo thought that it was going away
[22:21] <pochu> doko: I don't know them (I never touch it) but Ondrej does
[22:22] <slangasek> bryce: I don't think that's going to cease to exist inside the kernel, is it?  It's just going to be accessed via the event interface
[22:22] <bryce> yeah I dunno the details
[22:22] <bryce> https://wiki.ubuntu.com/Hotkeys updated
[22:23] <directhex> hm. RIP adept.
[22:24] <slangasek> ?
[22:24] <directhex> http://web.mornfall.net/blog/farewell__44___adept.html
[22:30] <NCommander> ugh
[22:30] <NCommander> Of course LP shuts down the moment I need it
[22:35] <calc> heh the information about the new LP release is linked back to... LP so you can't even read about it while it is being done
[22:35] <bryce> calc,  >o<
[22:36] <ScottK> slangasek: I have a fairly draconian whack at the Kubuntu CD prepped for whenever LP wakes up from being broken.
[22:36] <Nafallo> not broken :-)
[22:36] <Nafallo> upgraded
[22:38] <ScottK> Nafallo: If it's not working it's broken.  These long outages for upgrades are, IMO, a sign of poor design.
[22:39] <ScottK> Scheduled broken instead of randomly broken, but still broken.
[22:42] <NCommander> StevenK, did you ever find out if the bug in germinate was fixed w.r.t. to fbreader?
[22:42] <StevenK> NCommander: Nope
[22:42] <StevenK> Was plotting to ask Colin
[22:42] <NCommander> StevenK, ah good
[22:53] <ScottK> slangasek: Any idea how https://launchpad.net/ubuntu/jaunty/i386/kdeplasma-addons/4:4.1.85-0ubuntu2 got deleted?
[22:54] <slangasek> ScottK: <cough> probably the line where I said "<slangasek> ScottK: kdeplasma-addons NBSed", where neither of us seem to have noticed I got the package name wrong
[22:54] <ScottK> Ahh.  That'd do it.
[22:55] <ScottK> slangasek: Do I upload it again or can you put it back?
[22:55] <slangasek> I think we established at UDS that it's possible to restore NBSed binaries
[22:56]  * slangasek takes out the right package instead, while he waits
[22:56] <ScottK> slangasek: If you'd do that, I'd appreciate it.  Excellent.
[22:57] <ScottK> slangasek: OK.  When I updated kubuntu-meta it took that out of the desktop (since it didn't exist).  I'll wait until you put it back and run it again.
[22:57]  * slangasek nods
[22:57] <ScottK> It'll be a bit because I've got more kids to get to ballet lessons.
[22:59] <bryce> calc: now that lp is back up, where do we find the release notes?
[23:04] <slangasek> ScottK: fwiw, the people I need in order to get the binaries restored aren't responding - a reupload may be simpler
[23:09] <calc> bryce: well not sure if it is the release notes but the news.launchpad.net referred to the release via a link to LP
[23:09] <calc> https://launchpad.net/launchpad-project/+milestone/2.1.12
[23:10] <calc> i guess that sorta tells you what got fixed its just a list of bugs :)
[23:13] <bryce> ahh, good idea
[23:19] <bryce> calc, mm, lots of bug fixes, although nothing that I recall affecting me
[23:21] <bryce> calc, the signed ppa's sounds interesting; shame that the Read Only Launchpad feature doesn't appear to have gotten done
[23:22] <FAJALOU> just a question:  is a rolling release a possiibility for ubuntu?
[23:23] <directhex> FAJALOU, a 6-month release cycle is too long for you?
[23:23] <FAJALOU> directhex:  trying to update to a new system every 6 months is yes.
[23:23] <FAJALOU> i hate doing it b/c i always end up having to take days to reconfigure this and that.
[23:23] <bryce> FAJALOU: how would you define 'rolling release'?
[23:23] <FAJALOU> directhex:  it's more of the opposite; it's too short.
[23:24] <directhex> then use the LTS releases.
[23:24] <FAJALOU> bryce: as in i never have to go through a 'major' upgrade again.
[23:24] <ogra> FAJALOU, if you file bugs abut this and that these can be fixed for next release ;)
[23:24] <FAJALOU> directhex:  i would love that,,, except then there are things in 8.10 that don't show up in 8.04, even if they are both supported.
[23:24] <bryce> FAJALOU: so why not just stick with LTS?
[23:24] <directhex> short answer: no
[23:24] <directhex> long answer: noooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
[23:25] <FAJALOU> bryce:  because i just upgraded to 8.10... stupid decision on my part ;)
[23:25]  * bryce shrugs  worked for me
[23:26] <FAJALOU> bryce:  ya if only it 'just works' for all of us.
[23:27] <calc> the only time i have to bother with reconfiguring things is if i blow away my system entirely including my home dir, which doesn't happen often
[23:27] <calc> if you just do a real upgrade then it shouldn't require too much reconfiguration
[23:27] <bryce> FAJALOU: so it's not that you want a rolling release, you just want perfection
[23:27] <ogra> rolling perfection :)
[23:27] <calc> it shouldn't really require more reconfiguration even if it is broken, it either works or it doesn't generally, heh
[23:28] <FAJALOU> bryce:  well i see rolling release as closer to perfection,,, because you don't have to worry about total reconfigurement ;)
[23:28] <calc> FAJALOU: so just don't reinstall, just upgrade?
[23:28] <ogra> but you will have 100 times more bugs
[23:28] <bryce> FAJALOU: sure you do
[23:28] <bryce> FAJALOU: consider if you got a rolling upgrade to a new xserver, or a new kernel...
[23:28] <FAJALOU> bryce; ok ya i guess that makes sense.  you too ogra
[23:29] <ogra> in a 6 month release cycle less than half of it is used for development, the rest is stabilization and bugfixes
[23:29] <FAJALOU> point.
[23:32]  * calc bbl, going to give a speech at my LOCO
[23:32] <kees> doko: around? looks like the mysql jaunty pain is between -O1(works) and -O2(explodes).  (using gcc-snapshot)
[23:33] <directhex> FAJALOU, rolling releases are fine if you want zero QA, e.g. gentoo. stable releases imply someone's actually making sure it works
[23:33] <FAJALOU> QA?
[23:33] <ogra> quality assurance
[23:34] <doko> kees: you might want to file an upstream bug report with your current findings, and update this one later
[23:34] <slangasek> ScottK: correction, cprov rescued them for me
[23:34] <kees> doko: okay.  can I still do .o bisection even though some are -O1 and some are -O2 ?
[23:35] <doko> kees: sure, please do. we need at least find the miscompiled file
[23:35] <kees> doko: okay, cool.  /me continues...
[23:36] <kees> doko: (thought I suspect it's much more than just 1 file)
[23:37] <doko> kees: hmm, ok, please file the upstream report, maybe you get feedback there as well
[23:46] <LaserJock> slangasek: mind if I upload edubuntu-meta?
[23:46] <slangasek> LaserJock: go ahead
[23:47] <LaserJock> slangasek: thanks