[12:08] <BenC> "No packages matching 'linux-source-2.6.15' are published in Ubuntu."
[12:09] <BenC> I find that hard to believe
[12:09] <BenC> that's when searching for Source Packages
[12:10] <Mez> lol
[12:11] <Seveas> BenC, "There is no kernel"
[12:20] <CarlFK> Ok, now that I have my data in OOCalc, anyone know how to rotate it?
[12:20] <CarlFK> I want rows to be columns
[12:21] <jono> anyone know if the cairo-fied OOo will make it into dapper?
[12:22] <tseng> jono: not likely, its past feature freeze
[12:23] <tseng> jono: but its doko's baby
[12:23] <CarlFK> whops, wrong chan
[12:23] <jono> tseng, ahhh
[12:39] <jono> woo! http://www.amazon.co.uk/exec/obidos/ASIN/0132435942/qid=1141688252/sr=1-2/ref=sr_1_2_2/203-0360572-5579116
[12:40] <tseng> jono++
[12:41] <jono> :)
[12:41] <tseng> http://www.amazon.co.uk/exec/obidos/search-handle-url/index=books-uk&field-author=Bacon%2C%20Jono/203-4537622-1113542
[12:41] <lifeless> LarstiQ: /win 11
[12:41] <lifeless> bah
[12:43] <zul> heylo
[01:18] <Arrogance> jono, did you notice the amazon recommendations on your book?
[01:18] <jono> Arrogance, no?
[01:18] <Arrogance> all Windows XP stuff.  :)
[01:18] <jono> hehe
[01:18] <Triskelios> Kamion: happen to be awake?
[01:19] <Burgwork> Arrogance, which book?
[01:19] <Arrogance> http://www.amazon.co.uk/exec/obidos/ASIN/0132435942/qid=1141688252/sr=1-2/ref=sr_1_2_2/203-0360572-5579116
[01:25] <arp> Anyone lose gedit?
[02:24] <arp> Please may we have a gedit-2.13.93 for ppc?
[02:26] <infinity> arp: If the build-deps get fixed... 'Sec.
[02:26] <arp> ty infinity 
[02:26] <mjg59> Ooh, wireless /almost/ works
[02:27] <Amaranth> on what?
[02:27] <infinity> mactel?
[02:27] <mjg59> Yeah
[02:27] <ajmitch> impressive
[02:27] <arp> mactel, i.e. the new x86 macs?
[02:27] <mjg59> Yup
[02:28] <arp> nice
[02:28] <mjg59> I've got it associated with the AP
[02:28] <infinity> Close enough.  Ship it.
[02:28] <mjg59> No DHCP leases yet, though
[02:28] <infinity> Once a driver's getting that far, you can just claim the rest is user error and sit back and chuckle.
[02:28] <mjg59> Probably because the MAC address has come up as ff:ff:ff:ff:ff:ff
[02:28] <infinity> (Is this why I don't do kernel maintenance?... Maybe)
[02:29] <doko> infinity: the OOo segfaults look like nvidia related crap ...
[02:30] <jdub> ok, no one send mjg59 any more hardware
[02:30] <infinity> doko: Yeah, I'm going to test some theories and upload a new LRM (if I find an obvious fix) today.
[02:30] <jdub> it's just like feeding him crack
[02:31] <Amaranth> jdub: I'm thinking about sending him a broadcom wireless card so he can make that rock too. ;)
[02:31] <mjg59> jdub: Just think of the PR, dude
[02:31] <mjg59> Amaranth: I've got a couple...
[02:31] <mjg59> Including this Mac
[02:31] <doko> jdub: you're missing feedback on #30482
[02:31] <mjg59> The issue with the Mac is that it's a PCIE Broadcom, and we currently have no driver support for that
[02:31] <ajmitch> I'd send him my acer laptop but I think he'd just return it 
[02:32] <mjg59> So I'm just pretending that it's a PCI one, which seems to mostly work
[02:32] <doko> jdub: or just assign it to seb128 ...
[02:34] <infinity> mjg59: I see no reason why you'd need to address it as PCIE, unless you desperately need the extensions.
[02:34] <mjg59> infinity: No, but it's a different core
[02:34] <infinity> mjg59: The kernel's happy to provide a base-level PCI API for all PCI/AGP/PCI-X/PCI-E interfaces, afterall.
[02:35] <infinity> mjg59: Ahh, yes, that's more problematic. :)
[02:35] <mjg59> So I need to teach the driver that the PCIE core is much the same as a PCI one
[02:35] <mjg59> (Which seems to be mostly true)
[02:40] <doko> jdub: thanks
[02:40] <jdub> doko: sorry 8)
[02:43] <mjg59> You people would not believe just how hard I rock
[02:44] <mjg59-mac> Victory is mine
[02:45] <jdub> over wifi?
[02:45] <mjg59-mac> Yup
[02:46] <StevenK> mjg59-mac: Ouch. How many new scars did you manage to inflicit upon yourself?
[02:46] <jdub> sick fuck! (well done!)
[02:47] <mjg59-mac> Should only be a couple of lines of diff
[02:47] <mjg59-mac> Still no MAC address, but that's probably easier to fix
[02:48] <jdub> mac addresses are for hippies
[02:53] <zakame> hi devs
[02:54] <joelbryan> I have a temporary fix regarding tulip-based ethernet card that disconnects every 1 minute in dapper.
[02:59] <joelbryan> I have a davicom ethernet, my university uses those (more than 1000 pcs), I use mii-diag -r to reset the link, and I made a * * * * * cron job for it, if it isn't fix yet in release, I'll just populate mii-diag in 1000 pcs and copy the cron to those pcs.
[03:01] <dbernar1> What package are manual entries for system calls in? I do not have a man exec.
[03:01] <StevenK> manpages-dev
[03:01] <dbernar1> much appreciated.
[03:02] <joelbryan> my lovebug culprit university assigned me to install ubuntu to those pcs, (1000+) and waiting for dapper.
[03:04] <joelbryan> kernel 2.6.17 fixed the sis pci 900 ethernet card, which hangs before when detected.
[03:12] <joelbryan> DAMN!! the OpenOffice is fuckin' wicked!
[03:14] <joelbryan> oops, sorry, the new OpenOffice.org supports gtk2 file selector, and has an new splash screen. so Cool!!
[03:17] <joelbryan> it's not yet fully GNOME-ish, some are still javaish, you can easily tell the difference, but the save as and the open selector is Gtk+2 cool!
[03:17] <Amaranth> joelbryan: you already said all that in #ubuntu
[03:18] <HiddenWolf> I'd rather see OOo slimmed down and usability-fied than gnome-inised, but that's just me. :P
[03:19] <Amaranth> i'd rather see it removed from ubuntu
[03:19] <Amaranth> and replaced with KOffice and GNOME Office
[03:19] <Amaranth> but one of those isn't done yet
[03:21] <HiddenWolf> True
[03:21] <HiddenWolf> and abiword isn't up to scratch
[03:21] <joelbryan> Well I like GNOME-Office from the fact that it has Abiword, because it's fast.
[03:21] <HiddenWolf> it's fast, but it doesn't use odt, has issues with large documents, and it's interoperability with msoffice isn't as good as OOo's
[03:22] <wasabi> I really love Abiword. I'm starting to use it for some internal projects.... I'd like odt support though obviously.
[03:22] <wasabi> OOo is SO hard to program against it's not even worth it.
[03:22] <HiddenWolf> I'm in the unfortunate situation that I can't use it, since everyone in my world uses .doc :/
[03:22] <wasabi> UNO my ass.
[03:23] <HiddenWolf> Someone get me a decent replacement for impress and I'd be sold tho. :)
[03:23] <joelbryan> I just hope OpenOffice get GTK+ize.
[03:24] <joelbryan> OOo uses jar files, w/c has it's own VM, just like any java apps.
[03:25] <minghua> doesn't abiword CVS head already support ODT?
[03:25] <HiddenWolf> minghua: does, but not fully, and not by default
[03:26] <HiddenWolf> minghua: and they've stated they won't, because it doesn't fit their internal structure.
[03:26] <minghua> HiddenWolf: oh okay, thanks
[03:30] <mjg59> Hurrah!
[03:31] <mjg59> Working MAC address as well
[03:33] <jdub> mjg59: is this crazy workaround-apple stuff, or fairly normal hardware version stuff?
[03:34] <mjg59> jdub: It's just a newer Broadcom chipset than we've seen before
[04:34] <johanbr> Hi. I'm looking at the gnome-volume-manager source right now, with the aim of possibly offering a few minor patches. I noticed that the commands to run are stored as gconf keys. Is there any way to store optional commands in the gconf keys, i.e. say I want the gnome-volume-manager drop-down menu to give me a choice between running either sound-juicer or gnome-cd ?
[04:36] <tiefox> module bcm43xx is not present anymore as of 2.6.15-17 kernel
[04:37] <infinity> tiefox: A mistake, it's back in the next upload (which is just finishing up on the buildds, and should hit mirror later today)
[04:37] <tiefox> thx a lot infinity
[04:37] <tiefox> :)
[04:37] <infinity> s/mirror/mirrors/
[04:38] <arp> infinity: any news on gedit, or should I just shut up and wait?
[04:39] <infinity> arp: https://launchpad.net/+builds/+build/174002/
[04:39] <johanbr> infinity: Are you sure it's a good idea to ship bcm43xx in its current state? It was really unstable for me and my computer locked hard when I ran ifdown with the module loaded.
[04:39] <infinity> I think that translates to "shut up and wait for the binaries to be published" :)
[04:39] <arp> :)
[04:39] <infinity> johanbr: It works for enough people that we'd rather ship it, yes.
[04:40] <infinity> johanbr: Doubly-important for PowerPC users, many of whom have BCM wireless (apple notebooks), and none of whom can use ndiswrapper, which is i386-specific.
[04:40] <johanbr> infinity: Good point. I
[04:40] <johanbr> infinity: Sorry about that. What I meant to say is that you have a good point and I'm at least in the situation where I can use ndiswrapper.
[04:41] <mjg59> johanbr: It works fine on a lot of hardware
[04:41] <mjg59> (I'm using it here right now)
[04:42] <johanbr> mjg59: I don't doubt that, but on my HP nx6125 it's basically unusable.
[04:42] <mjg59> Really? Hm.
[04:42] <mjg59> Works on my nx6125 :)
[04:43] <infinity> (Note that, given its imaturity, there's a fair chance BenC will bump it to newer upstream snapshots a couple of more times before dapper releases)
[04:43] <infinity> So, here's hoping it'll work on 75~80 per cent of the bcm hardware out there by release, at least.
[04:45] <johanbr> mjg59: For me, with kernel 2.6.15-16-amd64-k8, it often dropped connections, spewed lots of errors into syslog and hard locked the kernel when I ran ifdown.
[04:45] <Mez> infinity, HARD PING
[04:46] <HiddenWolf> johanbr: did you file bugs?
[04:46] <johanbr> HiddenWolf: No. I figured with the driver being in such a rapid state of development there wouldn't be much point.
[04:47] <infinity> Mez: Half-hearted pong.
[04:47] <HiddenWolf> johanbr: there is always a point in making sure your issues are known, through the proper channels.
[04:48] <Mez> infinity, you asked me to ping you hard on monday
[04:48] <Mez> well
[04:48] <Mez> er
[04:48] <Mez> it's tuesday now ;)
[04:48] <infinity> So it is. :)
[04:48] <Mez> well - wanna work out why my control hack isnt working ?
[04:48] <infinity> Mez: /msg me with those URLs and stuff again, and I'll poke at it right after I'm done with LRM and PHP (which are top priorities right now, it seems)
[04:51] <johanbr> HiddenWolf: If you say so. I figured reporting bcm43xx bugs to ubuntu would mostly be considered noise, when the upstream driver already has changed by a large amount by the time Ben gets around to looking at the bugs, but if you think reports would still be useful I'll give the newest kernel a try with bcm43xx and file a report if it doesn't work.
[04:52] <HiddenWolf> johanbr: perhaps it's noise, but at least it'll draw more notice than the noise here in this channel when most devs are asleep.
[04:54] <mjg59> johanbr: That would be good
[04:54] <mjg59> It's not actually too hard to fix most of these bugs, as long as we get to see them
[04:56] <johanbr> Alright, thanks for the encouragement. Since apparently the latest kernel version had the bcm43xx module go awol, I'll watch the archive for a new version and give it a spin when I see it.
[04:56] <mjg59> Should be available in a couple of hours
[07:53] <pitti> Good morning
[07:55] <Mez> morning pitti
[07:55] <Mez> wtf?
[07:55] <Mez> It's morning!
[07:55] <Mez> eep
[07:56] <Mez> now: beer or no beer before I go to sleep
[07:56] <sivang> Good morning
[07:56] <pitti> hi sivang 
[07:57] <sivang> hey pitti , how's it going?
[07:57] <pitti> fine, thanks; and you?
[07:59] <sivang> pitti: pretty good, starting to see the finish of HUB :)
[08:00] <HiddenWolf> sivang: will that still be in dapper?
[08:00] <pitti> I don't see why it shouldn't be in universe
[08:00] <pitti> probably just too late for main
[08:01] <HiddenWolf> Good stuff. Time to migrate my mom then. :)
[08:01] <HiddenWolf> she wants to backup, doesn't know how, and calls me to do it once in a while
[08:02] <HiddenWolf> Then I have to figure out how windows works, etc. :)
[08:03] <sivang> HiddenWolf: The chances are dropping as we speak :) maybe if it will proove to be not full of bug as I release it and mdz will be happy to let it through realtively late in the process, it'll get included in main. But as I said, chances are low.
[08:04] <sivang> HiddenWolf: so I see I have a first tester ? :-)
[08:04] <HiddenWolf> Once I get around to ridding her of XP, and poking wine so it runs 2 little apps, hell yeah. :)
[08:06] <HiddenWolf> fixing a pc over ssh is so much easier than getting on the train for 2 hours each way. ;)
[08:08] <sivang> wow!
[08:08] <HiddenWolf> sivang: and personally, yeah. my current .tar.gz to somewhere option isn't ideal either. :)
[08:08] <sivang> has anyone seen the new Python website??
[08:08] <sivang> It went a redesign
[08:08] <sivang> HiddenWolf: hehe
[08:08] <HiddenWolf> wow
[08:09] <sivang> HiddenWolf: Hopefully, my application will be up to the expectations
[08:09] <HiddenWolf> that's a step forward to presentability.
[08:09] <HiddenWolf> sivang: I'm sure I can think of some things to nag about, no worries. :)
[08:10] <dAndy> sivang: got a link to this HUB thing so I can learn more?
[08:11] <sivang> dAndy: https://wiki.ubuntu.com/HomeUserBackup
[08:11] <dAndy> cool thanks, will look into it
[08:11] <sivang> np :) anyways, the python web site feels quicker , I wonder if they are using a trimmed version of zope or a new system.
[08:12] <HiddenWolf> I'm not convinced the new page is more usable tho.
[08:12] <HiddenWolf> News just shouldn't be at the bottom of a page.
[08:18] <freeflying> pitti: ping
[08:20] <pitti> hi freeflying 
[08:22] <freeflying> pitti: two bugs about language-support-ja
[08:22] <freeflying> pitti: https://launchpad.net/malone/bugs/31733
[08:22] <Ubugtu> malone bug 31733 in gtkhtml3.8 "Evolution mail composer crashes with SCIM" [Normal,Unconfirmed]  
[08:22] <freeflying> pitti: https://launchpad.net/malone/bugs/30905
[08:22] <Ubugtu> malone bug 30905 in language-support-ja "depends on scim-tables-ja, but not required for normal users" [Normal,Unconfirmed]  
[08:22] <pitti> well, that's a scim bug, not a l-s-ja one
[08:22] <pitti> (the first one)
[08:23] <pitti> freeflying: hm, what is scim-tables-ja good for then? It sounds useful for Japanese writers?
[08:23] <freeflying> pitti: they prefer to use scim-anthy then scim-tables-ja
[08:24] <freeflying> s/then/than
[08:24] <pitti> alright, no problem to fix that
[08:39] <pitti> freeflying: I fixed the s-t-ja bug
[08:40] <freeflying> pitti: nice 
[08:48] <freeflying> pitti: how about the scim-gtk2-immodule
[08:48] <mdke_> trappist, ping
[09:46] <Mez> eep
[09:46] <Mez> c++ *
[09:46] <Mez> anyone willing to help ?
[09:49] <pitti> Mez: maybe just ask the question?
[09:49] <Mez> for the line
[09:49] <Mez> _result = new Program(s, _useExecName)
[09:49] <Mez> how do i define it in the header file
[09:50] <Mez> I'm probs gonna cock it up anyways ;)
[09:53] <pitti> Mez: define 'it'?
[09:53] <Mez> _result
[09:54] <Mithrandir> Program *_result ?
[09:54] <pitti> ah, yes
[09:54] <Mez> aha :d
[09:54] <Mez> couldnt remember the * bit
[09:55] <pitti> new delivers a pointer
[09:55] <sivang> hmmm
[09:55] <sivang> I'm fighting with something stupid here, could use a py guru's help:
[09:56] <sivang> I have two funcitons installed on the gtk_main_iteration loop
[09:56] <sivang> one is a timeout function, to ProgressBar.pulse() ,
[09:56] <pitti> Good morning mdz_ 
[09:57] <sivang> the other monitors a pty fd , and updates the status of current file being worked on when the sub process's output has new output on the pty.
[09:57] <mdz_> morning
[09:57] <sivang> this function is installed with gobject.io_add_watch
[09:58] <sivang> now, the io watching func maintains a self.running boolean flag , so when there is no more processing to do there, it sets it to False.
[09:58] <sivang> the gui progress timeout func monitors that flag, and shoudl return false (thus stopping and automatically removed from the main loop by gtk) when self.running is false.
[09:59] <sivang> however: the timeout func only see it's value as True...:-/ am I doing anything wrong?
[09:59] <sivang> (even when the io_add_watch callback sets it to false, and stops on account of it.
[10:01] <lifeless> are you looking at the same variable?
[10:01] <sivang> it feels like the timeout func can only see the value that was current when it was installed on the gtk main loop, rather then getting it's real value everytime it's changed.
[10:02] <lifeless> which means its looking at a differnt instance
[10:02] <lifeless> I'd bet on a closure rather than an object tripping you up
[10:04] <sivang> lifeless: I made sure this is not the situation, but it is certainly worth double check :)
[10:09] <sivang> lifeless: okay, I made sure code wise this is the same variable I'm looking at. still, same issue... How can I make sure this is the same instance? if both of the callbacks produce different instances of 'self' then I'm in troulbe.
[10:10] <lifeless> past the code somewhere
[10:10] <simira> good morning, sabdfl 
[10:10] <lifeless> cannot comment in a vacuum
[10:11] <sabdfl> hey simira
[10:12] <pitti> hey sabdfl
[10:12] <sivang> lifeless: indeed, but I warn you, it's somewhat large now :)
[10:13] <seb128> pitti: thanks for the rhythmbox upload
[10:13] <sivang> hey sabdfl 
[10:15] <sivang> lifeless: muse.19inch.net/~sivan/upbackup.py
[10:16] <lifeless> whats the line I should look at
[10:17] <sivang> lifeless: 237, pushGUIBackup
[10:18] <lifeless> the variable in question ?
[10:18] <sivang> lifeless: self.running
[10:18] <sivang> lifeless: it get's updated in watch_Backup_pty_callback
[10:18] <sivang> (the function beneath)
[10:19] <lifeless> ok
[10:19] <lifeless> add
[10:20] <lifeless>           self.running = self.curBkp.next()
[10:20] <lifeless> +        if DEBUG_MODE: print "pushGUIBackup:: self.running = %s" % self.running
[10:20] <lifeless>          while gtk.events_pending():
[10:20] <lifeless> there aare two basic possibilities
[10:20] <doko> Kamion, mdz: please could you process python-pullparser and python-clientform in NEW?
[10:20] <lifeless> one is that you never run out of events
[10:21] <lifeless> another is that its never set in the first place
[10:21] <lifeless> I suggest not doing main_iteration and instead writing this as event driven
[10:21] <lifeless> but thats a suggestion ;)
[10:23] <sivang> lifeless: isn't this already event driven? that is, I already attend to the worker using io_add_watch :) do you aso suggest that when I never run out of events, then it's just stuck in there and that's why the self.running is never set?
[10:24] <sivang> lifeless: (weird thing this used to work, when the pre_build_kick was part of the .build method. I splite them since I had to propogate the child's pty up)
[10:25] <sivang> lifeless: about it never being set in the first place, I am sure it get's set. otherwise the io_add_watch would have never been removed and stopped at end of operation.
[10:27] <lifeless> you are manually iterating the event loop
[10:28] <lifeless> this means that you will have reentrancy issues that you are much less likely to have if you dont
[10:28] <lifeless> and reentrancy races are tricky enough anyway ;)
[10:29] <sivang> lifeless: I see, well, I added them such that the GUI won't freeze in the callbacks, as it did before adding those.
[10:29] <lifeless> that means your callbacks were taking too long
[10:29] <lifeless> rather than just doing whats available and returning
[10:29] <lifeless> which is what they should do in event based programs
[10:33] <sivang> mvo is also using this method to make GUI not block while doing long work in the background, so I acutally borrowed it from his code, and it seems to work so nicely so far. I wish I knew what had regressed.
[10:34] <pitti> ogra: ping
[10:40] <doko> Kamion, mdz, mdz_: please could you process python-pullparser and python-clientform in NEW?
[10:41] <sivang> lifeless: is there a real reason why the boolean falg doesn't get updated for the timeout function? 
[10:41] <sivang> lifeless: s/real/explainable/ ;)
[10:42] <lifeless> sivang: sorry. not sure on that, try one of the gnome dev channels 
[10:43] <sivang> lifeless: okay, thank anyway rob :)
[11:20] <Kinnison> is openssh-server in ship in breezy?
[11:20] <mdke> i'm fairly sure it isn't
[11:20] <Kinnison> :-(
[11:32] <mdz_> Kinnison: you know where the seed archive is, right?
[11:33] <Kinnison> mdz_: umm, if I said "no" would you fire me?
[11:34] <mdz_> Kinnison: ssh (server and client) have been on the CD since warty
[11:34] <Kinnison> mdz_: thanks
[11:34] <mdz_> Kinnison: I'd suggest a thorough investigation of https://wiki.ubuntu.com/DeveloperResources
[11:35] <Kinnison> good plan
[12:00] <Kamion> Lathiat: note how that clock-setup bug is only "Fix Committed", not "Fix Released"
[12:05] <MikeS> Hi. I've got an ubuntu kernel question: valgrind-SVN doesn't run on ubuntu kernels out of the box at the moment. It turns out that this seems to be because the ubuntu kernels have an unusual 2/2 user/kernel split, rather than the normal 3/1 (or 4/0), so the valgrind binary can't be loaded where it wants to be loaded. 
[12:06] <MikeS> Now, this can be worked around by using a different load address, but that limits processes running under valgrind to less total address space (unless we only do this for ubuntu, not generally, which seems like a nasty hack).
[12:06] <MikeS> Can anyone explain WHY the ubuntu kernels work this way?
[12:06] <Kamion> wasn't that changed recently?
[12:06] <Lathiat> Kamion: oh, ok
[12:06] <Lathiat> Kamion: cheers
[12:06] <MikeS> I don't know; I'm running dapper here
[12:07] <Lathiat> yeh it was
[12:07] <Lathiat> i saw something in the kernel changelogs about it
[12:07] <MikeS> Where would I find those?
[12:07] <Kamion> dapper-changes mail archives on lists.ubuntu.com
[12:08] <Lathiat> or /usr/share/doc/linux-image-`uname-r`/changelog.Debian.gz
[12:08] <MikeS> Lathiat: thanks
[12:09] <MikeS> "Incorrectly set memory split to 3/1 instead of the default 1/3. Revert to correct 2/2 mapping. Fixes problems with UML and Wine." - so I guess it's still 2/2, but I still don't know why :)
[12:11] <MikeS> Anyway, since that change was from BenC, I guess I'll ask him...
[12:11] <wm_eddie> doh.  The X server seems to be broken today >.<
[12:11] <Xof> oh good, I'm not the only one who noticed that kernel change
[12:11] <Xof> (well, apart from all my users too)
[12:12] <elkbuntu> is SANE going to be upgraded to the current in dapper?
[12:14] <pitti> ogra: I just uploaded a new alsa-utils which should fix sound on ppc; testing appreciated :)
[12:15] <minghua> Micksa: you may be interested in bug #30962, there is a lot of discussions about this split in that bug, although I understand none of them
[12:15] <Ubugtu> malone bug 30962 in linux-source-2.6.15 "Running wine applications instantly outputs "Killed." in the terminal" [Normal,Fix released]  http://launchpad.net/bugs/30962
[12:16] <MikeS> minghua: I assume that was aimed at me? Thanks; that looks like what the current ubuntu kernels do to valgrind...
[12:16] <pitti> Kinnison: g-p-m is not running on the current (ppc) live CD, but the package is installed; known problem?
[12:16] <Kamion> doko: done. there seems to be cruft in the source packages; please check that out
[12:17] <minghua> MikeS: ah yes.  wrong tab completion for nicks.
[12:19] <MikeS> minghua: but yes, that's exactly the same problem: apparently wine, like valgrind, wants 3/1, not 2/2.
[12:20] <wm_eddie> Man, filing a bug report with only lynx will be pretty hard...
[12:22] <doko> Kamion: just the build subdirs?
[12:22] <Kamion> doko: that and a stray changelog
[12:22] <Kamion> I noticed it in (iirc) python2.3-psycopg as well
[12:23] <Xof> regarding the 3/1 split: I have reports of the current 2/2 split breaking sbcl, cmucl and clisp on dapper.  (Whether you care about exotic programming languages I don't know)
[12:23] <mdke> wm_eddie, you can do it by email, if you have email, i think
[12:24] <crimsun> if they work in Breezy, that's a regression.
[12:24] <Xof> they work in Breezy
[12:24] <wm_eddie> :( I use gmail.
[12:24] <minghua> Xof, MikeS: sounds like we need an umbrella bug for this memory split thing in kernel :-)
[12:24] <wm_eddie> grr I can't remember how to redirect STDErr to a file so I can find out why X is bailing out after GDM.
[12:26] <MikeS> 2> foo
[12:26] <wm_eddie> thanks
[12:26] <MikeS> Xof: not surprising, it appears to be hitting most applications that have unusual memory-management- and lisps are pretty likely to do that sort of thing.
[12:27] <wm_eddie> there's an undefined symbol in libGLcore.so?
[12:28] <Lathiat> https://lists.ubuntu.com/archives/kernel-team/2006-February/000678.html
[12:40] <Kamion> pitti: not seeing gnutls11 in anastacia output yet
[12:41] <pitti> Kamion: hm, grep-dctrl does not show any deps/b-deps any more...
[12:41] <Kamion> probably on other architectures
[12:41] <pitti> I only checked openldap and rhythmbox (the two latest uploads), they built everywhere
[12:42] <Kamion> all_ubuntu_dapper_powerpc:libgnutls11                                       | gnutls11                        | gnome-control-center                         | Matthias Urlichs <smurf@debian.org>                                                   |          304474 |             712
[12:42] <pitti> but entirely possible that some package didn't built
[12:42] <Kamion> plus a bunch of stuff on hppa, ia64, sparc
[12:42] <Kamion> hppa's unsurprising since it hasn't been building recently
[12:42] <Kamion> all_ubuntu_dapper_ia64:libgnutls11                                     | gnutls11                          | libegroupwise1.2-9                             | Matthias Urlichs <smurf@debian.org>                                                   |          407344 |            1204
[12:42] <Kamion> all_ubuntu_dapper_sparc:libgnutls11                                    | gnutls11                          | evolution                                    | Matthias Urlichs <smurf@debian.org>                                                   |          291958 |             692
[12:43] <Kamion> all_ubuntu_dapper_ia64:libgnutls11-dev                                 | gnutls11                          | libedataserver1.2-dev                          | Matthias Urlichs <smurf@debian.org>                                                   |          552646 |            2104
[12:43] <Kamion> those are the others
[12:44] <Kamion> sparc seems only fractionally more out-of-date than the big three arches now; I'll put it back into anastacia
[12:44] <pitti> ah, thank you
[12:44] <Kamion> ia64 is probably still catching up from its glib2.0 breakage
[12:45] <pitti> Kamion: is that some output of a program I can access, too?
[12:46] <Kamion> pitti: not that exact output easily, but it's just germinate, you can run it yourself
[12:46] <Kamion> 'germinate --no-rdepends -s dapper -d dapper -c main,restricted,universe -a ia64' or similar
[12:47] <Kamion> should only take a few minutes
[12:49] <pitti> Kamion: ah, I see; on amd64, g-control-center depends on gnutls12, only on powerpc on 11
[12:49] <mdke> Riddell, around?
[12:50] <Kamion> hmm, the /etc/skel/Examples symlink means that we now run into Debian #152206 for every user
[12:50] <pitti> debian bug 152206
[12:51] <ogra> pitti, thanks, will test :)
[12:52] <pitti> Seveas: -Ubugtu- Error: Could not parse data returned by Debian bugtracker: need more than 1 value to unpack
[12:52] <Seveas> pitti, gracias - will look at it after the meeting
[12:53] <pitti> thank you
[12:53] <highvoltage> hi.
[12:54] <highvoltage> many of the new intel server boards map the memory from 3-4GB as 5-6GB, so with an ubuntu kernel that supports 4GB, it only sees 3.
[12:54] <highvoltage> should this be marked as a bug in ubuntu (that it should perhaps support up to 64GB RAM?)
[12:55] <Riddell> mdke: hi
[12:56] <mdke> Riddell, hiya. I've built some html with nice shiny blue kde things using xsltproc
[12:57] <wm_eddie> Doh, There isn't a bug in xorg, my hard drive just has no free space >.<
[12:57] <mdke> Riddell, see http://doc.ubuntu.com/test/kde.png
[12:57] <mdke> Riddell, I'd like to think about moving back to xsltproc now that I know how to do this, are there any other technical things that need sorting before this is possible?
[12:59] <infinity> highvoltage: You don't want the desktop kernel for those server boards.
[12:59] <highvoltage> infinity: where could i get a 'server-kernel'?
[01:00] <infinity> highvoltage: linux-image-server for 99% of server machines, and linux-image-server-bigiron for the really heavy equipment.
[01:00] <infinity> highvoltage: If neither of those seems to suit your needs, or they seem s abit wonky/goofy, please file bugs.
[01:00] <infinity> highvoltage: We need more testing of the server kernels.
[01:01] <Riddell> mdke: for website that's fine, for kubuntu-docs package we need to use meinproc for the index.cache
[01:01] <highvoltage> infinity: ok. this will be installed on 200 LTSP servers in South Africa, should be quite useful for testing then :)
[01:02] <Treenaks> infinity: I installed linux-image-server yesterday, but it didn't boot
[01:02] <mdke> Riddell, but meinproc doesn't support xincludes right? How do you do it for the server and packaging guides (which use xincludes)
[01:02] <infinity> Treenaks: Not at all?
[01:02] <ogra> infinity, talk to the industry to make bigger standard CDrom media and i'll ship the server kernel with edubuntu :)
[01:02] <Treenaks> infinity: missing initramfs
[01:03] <infinity> Treenaks: Erm..
[01:03] <infinity> Treenaks: Hardly seems to be the kernel's fault...
[01:03] <infinity> Treenaks: Are you using lilo, by any chance?
[01:03] <Treenaks> infinity: no, grub
[01:03] <mdke> Riddell, afaics normal HTML (without index.cache) opens fine in khelpcenter
[01:03] <infinity> Treenaks: Alright, bugs would be nice, then.
[01:03] <Treenaks> infinity: but if I install linux-image-server, it should install everything correctly, imho
[01:03] <Riddell> mdke: yes, for server and packaging guides I think it currently uses normal HTML
[01:03] <Treenaks> infinity: I didn't see the machine not boot :(
[01:04] <Treenaks> infinity: it's in a datacenter and I had to instruct the datacenter reboot monkey to do grub magic
[01:04] <doko> mvo: OOo upgrade ping
[01:04] <infinity> Treenaks: Yes, just like any other linux-image-* package, it should do everything for you, so I'm not sure what may have gone wrong for you.
[01:04] <Treenaks> infinity: it might be that /boot is too small..?
[01:04] <Treenaks> (but then it should complain about that)
[01:05] <mdke> Riddell, I'd really like to use normal HTML for them all, it has the advantage that we can use xincludes for them all and translation is therefore easier. If the html builds look right, I don't see a problem. But that's why I ask: I don't know khelpcenter very well
[01:05] <infinity> Treenaks: update-initramfs is lacking is some decent error conditions (that's a bug I need to fix before release), so it's possible that you could have hit an -ENOSPC condition or something...
[01:05] <infinity> Treenaks: I really couldn't tell you while grasping at straws.
[01:05] <Treenaks> infinity: I'll try again, and check before reboot this time
[01:05] <infinity> Treenaks: Was there an initrd at all?
[01:06] <Treenaks> infinity: when I dpkg --purged the package, it said there was no /boot/initrd.img-2.6.15-17-server
[01:07] <infinity> But you never actually checked after installation?
[01:07] <infinity> If you could reinstall that kernel and have a look, that'd be nice. :)
[01:07] <Treenaks> infinity: ok, doing it now [slow, slow /boot disk] 
[01:09] <Treenaks> hm, crap, I've freed up a lot of space on /boot after fixing it, so it might just work now
[01:09] <infinity> Well, if it does look like a no space issue, I know that's a bug that needs to be fixed in initramfs-tools, so no big stress...
[01:10] <Treenaks> initrd.img-2.6.15-17-server -- it's generated now
[01:11] <mdke> Riddell, I'll email
[01:14] <mvo> doko: hello
[01:15] <doko> mvo: hi
[01:17] <mdke> Riddell, emailed
[01:46] <Kinnison> Kamion: I now have bandwidth :-) (for a bit) currently branching espresso so I have a copy in case it goes out again. Will upload gparted once I've finished testing that my changes haven't broken it in non-installer mode
[02:19] <Kamion> Mithrandir: readahead is definitely still very slow when booting today's live CD in vmware
[02:21] <Mithrandir> Kamion: what's "very slow" for you?
[02:22] <Kamion> it's been sitting there for at least a minute, dunno yet
[02:22] <Kamion> it is proportionally much much slower than the rest of the system under vmware
[02:22] <Mithrandir> can't duplicate that.
[02:22] <Kamion> in vmware or not?
[02:23] <Kamion> we duplicated it in vmware at the UI sprint yesterday, so it's not just my installation
[02:23] <Mithrandir> both
[02:23] <Mithrandir> or at least with a live image as of friday or so
[02:25] <Kamion> ah, it's oopsed
[02:25] <Kamion> unfortunately the stack trace is huge and of course I don't have scrollback; the visible portion is all do_page_fault and error_code
[02:30] <Micksa> uhm
[02:30] <Micksa> where is the bug tracking system?
[02:30] <Kamion> https://launchpad.net/distros/ubuntu/+bugs
[02:34] <Kamion> Mithrandir: also, I get no worthwhile output if I select a mode other than VGA in the bootloader
[02:40] <Mithrandir> Kamion: yeah, that oops is something I see too
[02:40] <Mithrandir> it boots every third time or so
[02:41] <Kamion> mm, yeah, I also got it to boot by trying a few times
[02:41] <Kamion> is readahead a huge benefit on the live CD?
[02:41] <Mithrandir> no, it costs us about five seconds.
[02:42] <Mithrandir> but I suspect that's because it's not optimised for what we load
[02:42] <Kamion> can we just make casper turn it off for now then?
[02:42] <Kamion> I know it's a workaround but I'm worried by the bugs I'm seeing
[02:43] <Mithrandir> sure
[02:43] <Mithrandir> I haven't seen the oopses outside of vmware, so I suspect it's a bug in the scsi driver or vmware
[02:45] <Micksa> okay, so certain applications, including wine, valgrind and UML are getting SIGKILL'd starting with kernel 2.6.15-15, because of the different kernel/user address space split.
[02:46] <Micksa> is this because of bugs in the programs?  Are they able to not assume where the split is?
[03:13] <trappist> mdke_: pong
[03:23] <trappist> does it strike anyone else as ironic that the distro launched by mark shuttleworth doesn't have a valid ssl cert for its website?  I wonder if maybe mark has some pull over at thawte and could get us one.
[03:24] <Kinnison> which website?
[03:24] <trappist> https://*.ubuntu.com
[03:24] <trappist> the wiki, for example
[03:25] <Keybuk> the wiki has a valid cert
[03:25] <Kinnison> Yep, certified by Starfield
[03:26] <ogra> Kamion, could it be that the new world order in he installer (dropped base-config) changed something that prevents sshd from starting before reboot ? 
[03:26] <trappist> oh, it does now.  I quite paying attention since I said 'accept forever'.  it came to my attention again because an svn co from docteam.ubuntu.com complained about the cert.  I wonder why we don't just get a wildcard cert for ubuntu.com.
[03:26] <ogra> Kamion, d-i	preseed/late_command	string chroot /target /usr/sbin/ltsp-update-sshkeys
[03:26] <trappist> *quit
[03:26] <ogra> doesnt produce anything anymore on install 
[03:27] <ogra> any hint (from your experience) where specifically i could look
[03:28] <Kamion> ogra: /var/log/syslog would be a start
[03:28] <ogra> oki
[03:28] <Kamion> sshd certainly should *not* be started before reboot though
[03:28] <Kamion> what does ltsp-update-sshkeys do?
[03:28] <ogra> copying the ssh keys to the thin client chroot
[03:28] <ogra> but if sshd was never started it has no keys 
[03:28] <ogra> so my assumption was right ...
[03:29] <Kamion> that's not to do with starting sshd
[03:29] <Kamion> that's based on whether openssh-server has been configured
[03:29] <ogra> postinst doe it normally
[03:29] <Kamion> be careful of the distinction
[03:29] <ogra> hmm ...
[03:30] <Kamion> /var/log/syslog should have the entire output from pkgsel, including whether openssh-server has been configured
[03:30] <ogra> i wonder if its appropriate to call the ssh-keygen from ltsp-update-sshkeys ... but then i would have a new key every time ...
[03:30] <Kamion> to generate a host key? no, inappropriate
[03:30] <Kamion> you should already have a host key by that point
[03:30] <ogra> hmm
[03:30] <ogra> ok
[03:31] <ogra> i'll dig for it, i just wanted to check my assumptions before looking at it ...
[03:31] <Kamion> ok, you're using ssh-keyscan to get the host keys
[03:31] <Kamion> why?
[03:31] <Kamion> that relies on an actual running sshd
[03:31] <ogra> hmm, i didnt touch that code yet 
[03:32] <ogra> mdz, ?
[03:32] <Kamion> because it connects to the sshd to get its public host keys
[03:32] <ogra> mdz, is there a specific reason for using ssh-keyscan in ltsp-update-sshkeys ?
[03:33] <Kamion> I would think just copying them out of /etc/ssh/*.pub would be easier
[03:33] <ogra> yup
[03:34] <ogra> but before making such a change i'd like to hear mdz and his intention to use -keyscan :)
[03:34] <ogra> he must have had a reason for it ... (i suspect)
[03:35] <Kamion> probably to avoid having to hardcode host key file names
[03:35] <Kamion> which is a valid concern, but I think it's probably outweighed in this case
[03:35] <ogra> thats easily turned into a commandline option if you want to override it :)
[03:36] <Kamion> or parse sshd_config if you really want to be exact there
[03:37] <ogra> hmm, sounds reasonable ...
[03:42] <Kinnison> Kamion: I'm ready to land this gparted version
[03:42] <Seveas> ogra, "Attack of the holbachs"
[03:42] <Kinnison> Kamion: when I do, espresso will stop working with gparted until I fix it up
[03:42] <Kinnison> Kamion: will that bugger up your workflow for now?
[03:49] <MikeS> Micksa: The case I'm familiar with is valgrind: in that case, valgrind HAS to have a fixed load address compiled in (it's not like a 'normal' application for obvious reasons). The maximum amount of memory available to an application running under valgrind is limited by where this load address is, so it's desirable to have it as high as possible
[03:50] <Kamion> Kinnison: do you have a branch of espresso already that works with it?
[03:50] <Kinnison> not yet
[03:50] <Kinnison> working on it
[03:50] <Kamion> Kinnison: it will kind of bugger it up a bit, yeah :(
[03:50] <Kamion> I'd rather we waited until we can do coordinated uploads
[03:51] <Kinnison> okay
[03:51] <Kinnison> I'll hang onto it until I get a branch of espresso sorted
[03:51] <Kamion> 'cos I know the UI sprint folks want working espresso at the moment
[03:52] <Kinnison> right
[03:52] <Kinnison> best not break them then
[03:53] <Kinnison> I've branched espresso from your ubuntu repo into http://people.ubuntu.com/~dsilvers/bzr/espresso/gparted-updates
[03:53] <Kinnison> I'll tell you when I've pushed the updates there
[03:53] <Kamion> thanks
[04:08] <Micksa> mikes: do you know why dapper's kernel has had its split moved? do you know if valgrind or other progs can be made to not depend on the split being in a certain location?
[04:10] <MikeS> Micksa: I don't know when or if dapper's kernel had this moved (I thought I saw this on breezy too, but I might be wrong, I didn't investigate it at all at the time)
[04:11] <Micksa> my laptop constantly runs a uml.
[04:11] <Micksa> (naturally)
[04:11] <Micksa> when I upgraded to dapper it broke. then downgrading to kernel -14 fixed it
[04:11] <MikeS> Micksa: at least for valgrind, it HAS to depend on the split being at-or-above some location, and it makes valgrind more useful if you make that address higher.
[04:11] <Micksa> no kernels since -15 have worked with this uml binary
[04:11] <Micksa> mikes: so there is absolutely no way to make valgrind able to work regardless of where the split is?
[04:12] <Micksa> I can only guess since I have NFI how valgrind works under the hood but generally anything can be configured one way or another
[04:12] <MikeS> Micksa: note that dapper changed from a 2/2 split (abnormal) to a 1/3 split (extremely abnormal) temporarily, I think this was an accident. The latest .15 package has it returned to 2/2 (which isn't enough for valgrind, but is apparently enough for SOME but not all wine things to work)
[04:12] <Micksa> maybe this is one of those Hard Problems
[04:12] <Micksa> hmm.
[04:12] <Micksa> which latest package?...
[04:12] <MikeS> Micksa: you have to make this decision at compile time, not at runtime, so it Really Sucks if can't rely on the split being in the normal place
[04:12] <Micksa> htm.
[04:13] <Micksa> hrm, even
[04:13] <Micksa> that does blow.
[04:13] <MikeS> 2.6.15-16.22 or later
[04:14] <mjg59> It'll be back to 3:1 in the next kernel upload
[04:14] <Micksa> aw geez. I've got -17.24 and uml breaks with that :/
[04:14] <Micksa> I think
[04:14] <Micksa> mjg: oh good!
[04:14] <Micksa> mjg: also, MY LAPTOP DOESN'T S3 ANYMORE
[04:14] <Micksa> you know you love it
[04:14] <mjg59> The change was because there were some problems with suspend to disk and highmem, but we'll work on that
[04:14] <mjg59> Micksa: "Doesn't S3" isn't very descriptive
[04:14] <Micksa> I know
[04:15] <Micksa> I was just playing
[04:15] <mjg59> I'll kill you next time I see you
[04:15] <MikeS> mjg59: ok; thanks.
[04:15] <Micksa> that's quotable
[04:15] <Micksa> I gotta crash
[04:16] <Micksa> later
[04:16] <Amaranth> someone explain what a split is?
[04:16] <Kyral> Netsplit?
[04:18] <ogra> Bananasplit ? 
[04:18] <MikeS> Amaranth: in  this context: a linux process has part of its address space available to the application, and part mapped for the kernel. You can split the two parts at various places.
[04:18] <Amaranth> ah
[04:19] <mjg59> You've only got 4GB of address space available on x86, which is where the problem comes from
[04:20] <MikeS> oh yeah, I forgot to add that bit: this stuff only really matters on 32 bit architectures.
[04:24] <Xof> I don't know about that, actually.  The use of bits expands to fill the bits available
[04:27] <MikeS> Yes, but nobody has enough bits to fill the address space on a 64 bit host :-)
[04:35] <mdz_> doko: yes?
[04:37] <ogra> mdz_, due to the new order in the installer, sshd isnt started before ltsp-update-sshkeys is run in late command, is there any particular reason you coose ssh-keyscan instead of just cp to move the keys to the client chroot ? 
[04:37] <ogra> GAH !
[04:38] <doko> mdz_: could I have feedback on the mdbtools UVF exception?
[04:38] <ogra> mdz_, due to the new order in the installer, sshd isnt started before ltsp-update-sshkeys is run in late command, is there any particular reason you coose ssh-keyscan instead of just cp to move the keys to the client chroot ? 
[04:38] <ogra> (-keyscan requires a running sshd)
[04:47] <mdz_> doko: you emailed me twice about this already
[04:47] <mdz_> ogra: only that it was simpler to deal with multiple key types
[04:48] <ogra> mdz_, ok, then i can change it, thanks, just wanted to know the rationale behind it before i touch it
[04:52] <doko> mdz_: ? on Friday, and mentioned it today in the postscriptum of another mail
[04:57] <Kamion> mvo: progress bars don't seem to be shimmering at all for me in anything - it's not just espresso
[04:58] <Kamion> mvo: for example, as far as I can tell, synaptic's aren't either
[04:58] <Kinnison> I thought we turned the shimmering off
[04:59] <Kamion> I'll have to come up with some other way to indicate that espresso's still doing something during slow bits then
[04:59] <Kinnison> a throbber, mozilla-style?
[04:59] <Kinnison> an ubuntu logo, slowly spinning
[04:59] <Kinnison> :-)
[04:59] <Kinnison> or saturating/desaturating in a loop?
[05:00] <Kamion> dunno
[05:00] <seb128> Keybuk: nm-applet doesn't list any connection on my laptop :
[05:00] <seb128> :/
[05:00] <Kamion> TBH I suspect better text in the progress bar would do it
[05:00] <ogra> seb128, silbs had the same prob today iirc
[05:01] <seb128> Keybuk: it was working yesterday (versions from distro sprint), and since the update it only has a "Wired Network"
[05:01] <seb128> Keybuk: using an ipw2200 card
[05:01] <Kamion> anyone know where I can get a good magnifying-glass cursor?
[05:01] <ogra> Kinnison, after using these PM icons for a day, i must admit i liked the old ones better
[05:01] <ogra> Kamion, define good ? 
[05:01] <Kinnison> ogra: Hmm, because jdub really prefers the new jimmac ones
[05:01] <Kamion> ogra: s/good // if you like
[05:02] <Kamion> I was hoping not to have to kludge one together from an icon at run-time
[05:02] <Keybuk> seb128: what's in /etc/network/interfaces?
[05:02] <pitti> hey Keybuk 
[05:03] <ogra> Kamion, its sad that art.gnome.org dropped the icon tarballs with 1000 custom icons they once had ... give me a minut
[05:03] <ogra> e
[05:04] <seb128> Keybuk: http://pastebin.com/589065
[05:04] <seb128> Keybuk: it has the essid listed since I played with network-admin after having it broken yesterday
[05:05] <Keybuk> having wireless-mode and wireless-essid in there will prevent it being listed by Network Manager
[05:05] <ogra> Kamion, http://people.ubuntu.com/~ogra/search.png would that suffice ? 
[05:07] <seb128> Keybuk: hum, k, is working now
[05:07] <seb128> Keybuk: I probably already had a essid before but version before update was not skipping it
[05:07] <Kamion> ogra: hmm, looks a little weird somehow, but I think it'll do, thanks
[05:07] <seb128> thank you
[05:07] <Kamion> ogra: where's that from?
[05:07] <ogra> Kamion, try search2.png
[05:08] <Kamion> ogra: that's better, the longer handle helps
[05:08] <ogra> Kamion, both stolen from tango 
[05:08] <Kamion> I'll see how that looks, thans
[05:08] <ogra> locate search|grep png
[05:08] <Kamion> ah, I was trying zoom
[05:08] <ogra> ^^ helps :)
[05:09] <ogra> yeah, search returns more of the document-search icons though ...
[05:10] <Kamion> the semantics are actually zoom-in
[05:10] <Kamion> as in on a map
[05:11] <ogra> ah
[05:12] <Kamion> hmm, how about /usr/share/icons/gnome/24x24/stock/navigation/stock_zoom-in.png ?
[05:12] <jbailey> Looking at https://wiki.ubuntu.com/SeedManagement - IIRC, we have a separate CD for server now, but I don't see it mentioned.
[05:12] <jbailey> Is there a canonical place I can look for the information of what goes into the CDs?
[05:12] <Kamion> the cdimage code, probably
[05:13] <Kamion> colin.watson@canonical.com--2005/cdimage--mainline--0
[05:13] <jbailey> Kamion: Cool, thanks.
[05:13] <Kamion> I don't think anything else has been reliably kept up to date with *all* the weirdnesses
[05:13] <jbailey> Mmm.  Is that still a baz archive?
[05:13] <Kamion> yeah, I'll convert it at some point but haven't had time
[05:13] <ogra> Kamion, stock_zoom-in.png seems better for that purpose ...
[05:13] <jbailey> Is now the point where I admit that I've never learned to work baz? =)
[05:14] <jbailey> Kamion: Can I ask you to tar it up at some point?  Hande's asking me questions as to what we're shipping.
[05:14] <Kamion> baz register-archive http://people.ubuntu.com/~cjwatson/colin.watson@canonical.com--2005
[05:14] <Kamion> baz get colin.watson@canonical.com--2005/cdimage--mainline--0
[05:14] <Kamion> look in bin/list-seeds, compare with the seeds in http://people.ubuntu.com/~cjwatson/seeds/
[05:15] <jbailey> Kamion: Nice, thanks.
[05:15] <jbailey> If this works, I'll save this to a text file.  That's the best explanation I've gotten so far from people on how to check something out. =)
[05:17] <jbailey> Kamion: Ah, ~cjwatson/archives/colin.watson..., but seems to be working.
[05:17] <jbailey> Kamion: Thanks!
[05:19] <highvoltage> jbailey: you know that irritating sugababes song, push the button?
[05:20] <highvoltage> something i said? :P
[05:28] <jbailey> Meh, have to sort of the X crashes soonish.
[05:28] <jbailey> sort out, rather.
[05:30] <Kamion> jbailey: right, sorry for the bogus URL
[05:30] <jbailey> All good.  I can troubleshoot 404 errors. =)
[05:36] <mdke> Riddell: http://doc.ubuntu.com/kubuntu/desktopguide-web/C/ <-- new bling. I'd really really like to do that for the kubuntu-docs package too
[05:37] <mdke> having multiple build systems is a pain
[05:37] <jdub> Kinnison: that was totally based on seeing them outside the UI though, so take that POV with a grain of salt :-)
[05:38] <ogra> jdub, i really cant accociate the icon with a battery 
[05:38] <Kinnison> jdub: fair enough
[05:39] <ogra> at least not as it is in the panel currently
[05:39] <ogra> if we could get the svgs the current icons are based on, it would be very easy to change the colors to something more tame
[05:40] <ogra> (i.e. a 10 min job)
[05:41] <ogra> and i dont think people complain about the shape ...
[06:01] <Kyral> uuh...stupid question. Will Dapper have NTFS Write abilities?
[06:02] <highvoltage> Kyral: very unlikely
[06:02] <Kyral> highvoltage: Thats what I thought
[06:02] <Kyral> someone in #ubuntu had the idea that it would. I said no, but I just wanted to doublecheck :D
[06:03] <mkrufky> Kyral: i dont even think the stock kernel has ntfs write yet
[06:03] <mkrufky> it does, but AFAIK, it will only write to a file of equal or smaller file size, and files of smaller sized get padded with null's
[06:03] <Kyral> mkrufky: I know, I just didn't wanna wind up with proverbial egg on my face
[06:03] <mkrufky> hehe
[06:04] <highvoltage> mkrufky: i think it has partial write, but that's only changing existing files
[06:04] <mkrufky> highvoltage: yes, exactly
[06:04] <mkrufky> highvoltage: and the file sizes must remain the same, as well
[06:05] <mkrufky> highvoltage: OR get smaller (with padded nulls)  but not bigger
[06:05] <mkrufky> I have TDS functionality working correctly using php5 on my ubuntu server, communication with mssql on a windows server..... unfortunately, Breezy's php5 doesnt seem to include support for stored procedures, "mssql_init(), mssql_bind(), mssql_execute()" although all online documentation says that the functionality should be present in php4.1 and later....  anybody know what i'm missing, here?
[06:07] <Riddell> mdke_: do what?
[06:23] <desrt> mjg59; are you going to be putting all of this evil in dapper?
[06:23] <mjg59> Yes
[06:23] <Seveas> evil is mjg59's last name ;)
[06:23] <desrt> do you know if your work works on macbooks?
[06:23] <mjg59> desrt: Ought to
[06:23] <ogra> Seveas, since xgl i actually tend to belive you :)
[06:23] <mjg59> Though they have atheros wireless, not Broadcom
[06:24] <desrt> that's a weird twist
[06:24] <desrt> i guess the ibooks would probably be atheros too
[06:24] <desrt> the 'conspiracy theorist' in me wonders if this is a nod to linux users
[06:25] <highvoltage> a nod?
[06:26] <mjg59> desrt: On the one hand, wireless that works with an entirely open driver (after an hour or so of hacking), on the other hand something with a binary blob and the entire FreeBSD 802.11 stack in it?
[06:26] <Kamion> mjg59: do you think that dmidecode is reliable?
[06:26] <mjg59> Kamion: In what respect?
[06:26] <Kamion> can I rely on it in the installer as a way to tell that this is an Apple i386?
[06:26] <mjg59> Yes, that much should be reliable
[06:26] <mjg59> (As of the version I uploaded at the weekend)
[06:27] <Kamion> of course, I don't have dmidecode-udeb in the base installer system, and I'm not sure I want to
[06:27] <Kamion> although it's only 22KB I guess
[06:27] <mjg59> It's pretty tiny
[06:27] <mjg59> Now, if we could just get elilo to work...
[06:28] <Kamion> unfortunately I wouldn't get Debian to take it in the base
[06:28] <Kamion> since the base applies to e.g. floppies too
[06:28] <mjg59> Ah
[06:28] <desrt> mjg59; but at the time when apple was making hardware decisions about what to put in the new mac laptops the open source broadcom driver didn't exist
[06:28] <mjg59> desrt: Hm. Possible.
[06:29] <Kamion> mjg59: oh, can I dig it out of /sys/firmware/efi/systab instead?
[06:30] <mjg59> Kamion: Nope
[06:30] <mjg59> That just gives you the SMBIOS entry point
[06:30] <Kamion> bugger
[06:30] <mjg59> What is irritating is that the kernel /knows/ this stuff, it's just never provided to userspace
[06:31] <mjg59> I guess we could add a small amount of sysfs code to export it
[06:31] <Kamion> that would be ideal
[06:32] <Keybuk> that's what the kernel is supposed to do, indeed
[06:32] <Keybuk> sysfs attributes are trivial, as they're just kobject attributes
[06:32] <Keybuk> there's no excuse for *not* providing them
[06:32] <Kamion> what would it be an attribute on?
[06:33] <Kamion> /sys/class/misc/ somewhere?
[06:33] <mjg59> Ok. I'm not likely to have time, but basically it needs to export the dmi_ident array
[06:33] <Keybuk> /sys/devices/system I'd've said
[06:33] <Keybuk> or /sys/devices/plat
[06:33] <Keybuk> +form
[06:33] <mjg59> Mm? I'd have thought /sys/firmware
[06:33] <Keybuk> /sys/firmware is the firmware subsystem
[06:34] <mjg59> This is firmware information
[06:34] <Keybuk> that's just used for firmware requests and loading
[06:34] <mjg59> Keybuk: Uh, no
[06:34] <mjg59> /sys/firmware is the /system/ firmware
[06:34] <Keybuk> oh, sorry
[06:34] <Keybuk> yes
[06:34] <mjg59> Hence /sys/firmware/acpi and /sys/firmware/efi
[06:34] <Keybuk> which != /sys/class/firmware <g>
[06:34] <Keybuk> yay kernel guys
[06:34] <mjg59> /sys/firmware/dmi would seem to make sense
[06:35] <mjg59> Ok, yeah. Take a look at arch/i386/kernel/dmi_scan.c, then just export each member of dmi_ident[]  
[06:35] <Kamion> as a bonus, you wouldn't have to be root any more to run dmidecode if all the DMI information were exposed there
[06:36] <mjg59> Would require adding ascii names to each of them, since that information isn't kept
[06:37] <mjg59> (include/linux/dmi.h)
[06:45] <gusaweb> hi !
[06:45] <gusaweb> I have a problem with floppy disk in dapper
[06:46] <janimo> gusaweb, try #ubuntu-users
[06:46] <gusaweb> janimo I think it is a bug
[06:46] <janimo> gusaweb: the report it in LP
[06:46] <gusaweb> there is no /dev/fd0
[06:47] <gusaweb> ok
[06:47] <mkrufky> modprobe floppy ?
[06:49] <gusaweb> mkrufky it is ok now thanks
[06:49] <mkrufky> heehe im glad i could help
[06:49] <gusaweb> mkrufky do you know the reason of that problem?
[06:52] <mkrufky> gusaweb: floppy is a kernel module ... apparantly it is not being loaded automatically on your system .. you can get it to load automatically by using /etc/modprobe.conf ... but that is off-topic for this room
[06:52] <mkrufky> gusaweb: janimo was right when he pointed you to #ubuntu-users ....
[06:52] <gusaweb> mkrufky ok i am sorry. thanks a lot
[06:52] <mkrufky> gusaweb: dont be sorry ...  ;-)  
[07:32] <pitti> doko: I approved your two new python packages
[07:36] <doko> pitti: thanks
[07:39] <ogra> pitti, thanks 
[07:39] <ogra> (unbreaks the edubuntu CD)
[08:00] <mdz-sprint> Kinnison: my thinkpad's hibernate and suspend keys stopped working in Dapper; is this related to g-p-m changes?
[08:00] <Kinnison> mdz-sprint: have you checked in gnome-power-preferences to ensure they're enabled?
[08:01] <mdz-sprint> Kinnison: there is only one option, for 'sleep button'
[08:02] <Kinnison> is that set to suspend?
[08:02] <mdz-sprint> Kinnison: i have a suspend button and a hibernate button, both of which did just the right thing in breezy
[08:02] <Kinnison> the hibernate button support may be waiting on a kernel fix
[08:02] <Kinnison> the suspend button should work if it's enabled in g-p-p
[08:02] <mdz-sprint> sleep button action was set to 'do nothing'
[08:02] <Kinnison> set it to suspend and it should start working
[08:03] <Kinnison> mjg59 asked me to set it to 'do nothing' by default because many people's laptops break with suspend
[08:04] <mdz-sprint> unfortunately the net result is that laptops where suspend had been enabled in Breezy stop working
[08:04] <Kinnison> yes
[08:04] <mdz-sprint> that's worse
[08:04] <Kinnison> talk with mjg59, ultimately I'll take on whatever you two decide
[08:05] <mdz-sprint> mjg59: ?
[08:05] <Kinnison> I don't quite see how I can migrate the breezy choice though
[08:05] <Kinnison> unless I do something odd the first time g-p-m is started or something
[08:05] <mdz-sprint> Kinnison: we created a whitelist in breezy so that suspend was automatically enabled on models where it was known to work
[08:06] <Kinnison> again, it'd have to be a 'first boot' kinda action
[08:06] <Kinnison> because it's a gconf setting
[08:06] <mdz-sprint> if that policy has moved to g-p-m, the logic for choosing the default should migrate too
[08:07] <mdz-sprint> the most correct behaviour would be to use the whitelist unless the user explicitly sets the preference
[08:07] <Kinnison> right. that whitelist is where?
[08:08] <Kinnison> I've gotta go pretty soon, can you mail me to remind me to look into this
[08:08] <Kinnison> ?
[08:09] <mdz-sprint> Kinnison: in acpi-support
[08:09] <mdz-sprint> see /usr/share/acpi-support
[08:11] <mjg59> mdz-sprint: There's no good way of setting a gconf default based on acpi-support settings
[08:13] <mjg59> The default is supposed to come from the schema
[08:16] <mdz-sprint> mjg59: having suspend no longer work out of the box is a serious regression; surely there is something we can do
[08:16] <ogra> mjg59, you could set a systmwide default gconf key from postinst 
[08:20] <Burgwork> http://www.usa4id.com/ciwc/SawedOff.htm <-- mark looks entirely too smug there
[08:20] <mjg59> mdz-sprint: Hmm. I guess.
[08:20] <tseng> Burgwork: is that meant to shoot aliens?
[08:21] <tseng> Burgwork: or insane cosmonauts
[08:21] <Burgwork> tseng, no wolves in the backwoods of russia
[08:21] <tseng> 'survival kits on board all spacecraft'
[08:22] <tseng> oh
[08:22] <tseng> rtfa
[08:32] <tseng> Burgwork: http://en.wikipedia.org/wiki/Voskhod_2
[08:50] <spiekey> hello!
[08:51] <slomo> infinity, lamont: please give-back liboil on ppc
[08:52] <spiekey> has the  Bug #24468 in linux-source-2.6.15 (Ubuntu): "Breezy install kernel panics"  been fixed yet?
[08:52] <Ubugtu> malone bug 24468 in linux-source-2.6.15 "Breezy install kernel panics" [Major,Needs info]  http://launchpad.net/bugs/24468
[08:52] <spiekey> or is there a workaround?
[08:58] <ogra> mdz-sprint, did you get my PM ?
[09:12] <mdz-sprint> ogra: no
[09:15] <ogra> mdz-sprint, got it now ? 
[09:15] <Keybuk> *sigh* ... I'm so putting  "explaining to people why NM 0.6 isn't going to dapper" in my activity report, and assigning three or more hours to it
[09:15] <Mithrandir> Keybuk: write up a wiki page. :-P
[09:17] <Keybuk> NoYouCantHaveAPony? :)
[09:21] <Surak> Hello
[09:22] <Surak> Doko: there?
[09:22] <Surak> About https://launchpad.net/malone/bugs/33629 - the tip "GTK_IM_MODULE_FILE=/etc/gtk-2.0/gtk.immodules.32" does not work 
[09:22] <Ubugtu> malone bug 33629 in ia32-libs-openoffice.org "openoffice2 not showing widgets in breezy amd64" [Normal,Needs info]  
[09:24] <pitti> Keybuk: hi
[09:25] <pitti> Keybuk: do you have a minute?
[09:25] <Keybuk> pitti: for you, Martin, I have five! :D
[09:25] <pitti> Keybuk: I figured out why n-mgr doesn't work any more with l-wlan-ng
[09:25] <Keybuk> oh right?
[09:25] <Keybuk> what did I break?
[09:26] <pitti> Keybuk: 60-dispatch-more-events.patch now causes if-post-down scripts to be run
[09:26] <pitti> which is fine by itself
[09:26] <pitti> it just causes a bad series of events
[09:27] <pitti> I plug in the device, then n-m sees that it cannot control it (since it's not yet activated with wlanctl-ng)
[09:27] <pitti> so it actively disables the device, which causes the module to get unloaded
[09:27] <pitti> and thus the dbus-send for refreshing the device fails because the kernel device is gone
[09:27] <Keybuk> why does the module get unloaded?
[09:27] <Keybuk> does linux-wlan-ng have a post-down script to do that?
[09:27] <pitti> Keybuk: because l-wlan-ng has a if-post-down script that does it
[09:27] <Keybuk> ahh
[09:27] <pitti> yes :)
[09:28] <Keybuk> that's not "new world order compliant"
[09:28] <Keybuk> send the boys around in their black helicopters and make it "disappear"
[09:28] <pitti> Keybuk: ah
[09:28] <Keybuk> modules should be loaded by udev/modprobe and never unloaded, unless by hand by the user
[09:28] <pitti> Keybuk: I wondered whether it was a good idea in the first place to fiddle with a device which n-m can't manage anyway
[09:28] <Keybuk> how does n-m decide "it can't control it" ?
[09:29] <pitti> Keybuk: 
[09:29] <pitti> NetworkManager: <information>   wlan0: Device is fully-supported using driver 'prism2_usb'.
[09:29] <pitti> NetworkManager: <information>   nm_device_new(): waiting for device's worker thread to start
[09:29] <pitti> NetworkManager: <information>   nm_device_new(): device's worker thread started, continuing.
[09:29] <pitti> NetworkManager: <information>   Now managing wired device 'wlan0'.
[09:29] <pitti> NetworkManager: <information>   Deactivating device wlan0.
[09:29] <Keybuk> it could be a bug in the post-down stuff I guess
[09:29] <pitti> Keybuk: at that time it sees wlan0 as a wired interface
[09:29] <Keybuk> ahh
[09:29] <Keybuk> ok, so it thinks it can control it, just as a wired interface
[09:29] <Keybuk> so waits for link/no-link ?
[09:29] <pitti> i. e. it becomes a wireless one after wlanctl-ng
[09:30] <Keybuk> right
[09:30] <Keybuk> one way to fix this
[09:30] <pitti> Keybuk: probably something like this
[09:30] <pitti> Keybuk: if I deactivate the post-down script, it works fine, btw
[09:30] <pitti> so if that's what the new world order wants, then I'll just do that
[09:30] <Keybuk> look at debian/patches/30-blacklist-devices.patch
[09:30] <Keybuk> you could add similar code to "refuse to control" a not-yet-ready wlan-ng device
[09:30] <Keybuk> or, alternatively, you could just drop that "rmmod" script
[09:31] <Keybuk> removing drivers tends to be more problem than its worth
[09:31] <lamont> pitti: if I wanted to have pmounter do it's thing even when no one is logged in, how painful would that be for me?
[09:33] <pitti> meh, sorry
[09:33] <pitti> don't kill your foreground network-manager when talking to Keybuk 
[09:33] <Keybuk> did you get my last?
 look at debian/patches/30-blacklist-devices.patch
 you could add similar code to "refuse to control" a not-yet-ready wlan-ng device
 or, alternatively, you could just drop that "rmmod" script
 removing drivers tends to be more problem than its worth
[09:33] <pitti> Keybuk: sorry, the last thing I got was <Keybuk> look at debian/patches/30-blacklist-devices.patch
[09:34] <Keybuk> -- 
[09:34] <pitti> Keybuk: alright, I'll do that then
[09:34] <pitti> Keybuk: however, I think rmmod'ing the drivers is necessary for suspend/resume support
[09:34] <pitti> Keybuk: since the module doesn't support suspend
[09:34] <pitti> I fixed that for devices in /etc/network/interfaces
[09:35] <pitti> but it should be easy to make it work for n-m, too
[09:35] <pitti> Keybuk: ok, thanks for your wisdom :)
[09:36] <Keybuk> right, but that's better done at suspend/resume
[09:36] <Keybuk> rather than whenever anyone goes to change the interface
[09:37] <Keybuk> ifdown && ifup should always work :)
[09:40] <pitti> Keybuk: hmm, that post-down script also disables the interface (to conserve power and "to make sure it is not trying to generate interrupts"
[09:40] <pitti> Keybuk: that seems sensible to me, but would create a race condition
[09:43] <lamont> pitti: if I wanted to have pmounter do it's thing even when no one is logged in, how painful would that be for me?
[09:43] <lamont> (since you probably missed that one...)
[09:43] <pitti> lamont: Hi! yes, I missed it
[09:43] <pitti> lamont: to whom the device should belong then?
[09:43] <lamont> well, if it's ext2, root
[09:43] <pitti> ah, I see
[09:44] <lamont> etc, etc.  it's more a question of "how does pmount get bolted into the udev event fabric" 
[09:44] <pitti> lamont: the easiest might be a little udev script that checks if nobody is logged in and calls pmount if not
[09:45] <pitti> lamont: usually, it's kernel -> udev -> hal -> gnome-volume-manager -> pmount
[09:45] <Seveas> pitti, do you happen to recall the bugnumber that made ubugtu barf?
[09:45] <Seveas> my laptop crashed before I saved the log
[09:45] <lamont> mako: afternoon
[09:45] <pitti> Seveas: oh, it was a specific one?
[09:45] <Seveas> yes
[09:45] <pitti> Seveas: lemme try to find it again
[09:48] <pitti> lamont: sth. like /etc/udev/rules.d/80-root-automount.rules: SUBSYSTEM="block", BUS=="usb", RUN += "pmount-root.sh"
[09:48] <pitti> lamont: then, /lib/udev/pmount-root.sh will be called
[09:48] <Seveas> debbugs sucks quite hard for these purposes - the html is inconsistent and the mbox does not contain all information I need
[09:49] <pitti> lamont: this script could do 'pidof gnome-volume-manager', and if that's empty, call 'pmount-hal $DEVPATH
[09:49] <lamont> pitti: and /lib/udev/pmount-root.sh would be my script that DTRT
[09:49] <lamont> niceness
[09:50] <pitti> lamont: oh, wait, not DEVPATH
[09:51] <pitti> lamont: $DEVNAME is what you want
[09:53] <pitti> debian bug 355708
[09:53] <Ubugtu> Error: I tried to send you an empty message.
[09:54] <pitti> Seveas: this doesn't look right as well: <Ubugtu> Error: I tried to send you an empty message.
[09:54] <Seveas> gracias (these bogus errors it spits out now are bacause I'm working on it)
[09:54] <pitti> Seveas: ah, I see; I think I had a different bug some hours ago
[09:54] <Seveas> correct
[09:54] <Seveas> something about unpacking
[09:54] <Seveas> as I'm rewriting the debbugs part now, the exact error is irrelevant
[09:55] <pitti> Seveas: ah, I found the bug again: 152206
[09:55] <Seveas> urgh, how do you page up/page down in screen?!
[09:56] <pitti> I think you can't
[09:56] <Treenaks> Seveas: Ctrl+A, ESCAPE
[09:56] <pitti> enabling logging is your friend :)
[09:56] <Treenaks> Seveas: then use the arrow keys
[09:56] <Treenaks> termcapinfo xterm ti@:te@
[09:56] <Treenaks> in .screenrd
[09:56] <Treenaks> in .screenrc even
[09:57] <trappist> Treenaks: with that you don't have to ctrl-a esc?
[09:58] <Treenaks> trappist: only if normal scroll-up doesn't work
[09:58] <mdke> Riddell, use the same build scripts for both web and distribution. I emailed about it
[10:19] <Seveas> pitti: debian bug 152206
[10:19] <Ubugtu> debian bug 152206 in adduser ""deluser --remove-all-files" doesn't remove the complete home of the user" [Normal,Open]  http://bugs.debian.org/152206
[10:19] <pitti> Seveas: \o/
[10:19] <Seveas> and I take back what I said about debbugs - it sucks less hard than I thought ;)
[10:19] <pitti> thank you!
[10:20] <Seveas> np, thanks for reporting
[10:21] <mako> lamont: hey there
[10:24] <mdke> nice one Seveas 
[10:25] <pitti> lamont: oh, btw, your udev rule should also have ACTION=="add"
[10:39] <shaya> is gdm borked?
[10:40] <slomo> infinity, lamont: please give-back liboil on ppc
[10:40] <lamont> slomo: at this time, that's an infinity request. :-(  I hope to become schooled in LP buildd admin sometime "real soon now"?
[10:40] <lamont> s/?//
[10:40] <slomo> oh ok :/
[10:45] <lamont> mind you, that doesn't mean "quit bothering me", just "don't expect me to fix it..."
[10:46] <lamont> reminding infinity that I'd do it if I could doesn't hurt... but LP changes are needed, AFAIK
[10:53] <tepsipakki> lamont: I'm still waiting to get more info about the nfs-user-server issue with nfs4.. hopefully it will be resolved this week..
[10:53] <tepsipakki> lamont: from the nfs-team, that is
[10:54] <lamont> neato
[11:18] <fantasai> Can't seem to search in Malone, it keeps giving me an error. Is that reported already?
[11:20] <Burgwork> fantasai, ask in #launchpad