[12:43] <nemish> Can someone tell me yes or no but I thought in xorg that radeon driver was deprecated and supposed to use ati driver?
[12:44] <Burgwork> nemish, not really, but this is not a support channel
[12:44] <nemish> Burgwork, thanks.. sorry
[12:45] <Burgwork> nemish, no worries
[12:50] <TheMuso> c
[01:02] <BenC> isn't Breaks supposed to obsolete the provides/conflicts/replaces trio?
[01:03] <Keybuk> BenC: not obsolete, no
[01:03] <Keybuk> Breaks does replace most uses of Conflicts: <<
[01:04] <Keybuk> it's for "these don't work together", leaving conflicts for "these contain the same files"
[01:04] <BenC> so if a package is renamed, then I should still use the old provides/conflicts/replaces setup to replace the old package?
[01:06] <BenC> this is for linux-kernel-headers being renamed to libc-dev-linux-headers, which only libc6-dev depends on, and it's not a versioned depends
[01:07] <BenC> means libc-dev-linux-headers wont really be installed until libc6-dev is updated, but at least things wont break
[01:13] <ghee22> hi
[01:13] <ghee22> I am trying to make my program l10n compatible or I18n compatible
[01:13] <Burgwork> ghee22, you need gettext support
[01:13] <ghee22> anyone have directions on how to get this started?  I am reading through the debian intro on this topic and it's very thourough
[01:14] <ghee22> hey burg, I am not familiar with gettext
[01:15] <Burgwork> sorry, nor am I
[01:15] <ghee22> ok, if anyone is, is the gnu gettext? http://www.gnu.org/software/gettext/
[01:18] <allee> bug or feature that pmount version in dapper-updates is higher than in edgy?
[02:18] <BenC> Kamion: nic-firmware will be in the 2.6.17-6 kernel uploading tonight
[02:20] <gnomefreak> BenC: Kamio^n is onvacation this week 
[02:21] <Burgwork> gnomefreak, people on vacation at canonical have an alarming habit of still being on irc
[02:21] <gnomefreak> that doesnt sound like a vacation :(
[02:22] <imbrandon> that and they read irc logs too ;)
[02:23] <Burgwork> imbrandon, are you telling my people who develop ubuntu can read? I am shocked ;)
[02:23] <imbrandon> hahah 
[02:24] <imbrandon> they read / dream in c++^W<fav prog lang here>
[02:25] <Burgwork> imbrandon, this is ubuntu. They dream in python
[02:26] <imbrandon> ahh right 
[02:32] <ohoel> was xorg 7.1 in knot1?
[02:42] <crimsun> no
[02:43] <crimsun> (well, it depends how much of 7.1 you're referring to. It had bits and pieces already.)
[03:01] <ohoel> thanks crimsun
[06:50] <wasabi> Heh. X in edgy is actually fine! It's fglrx that's messed. =(
[09:16] <lmanul> Hi desrt :)
[09:17] <lmanul> desrt: Somehow "Manu the brown" doesn't sound as shiny as "Gandalf the white", I don't know why ;-)
[09:17] <desrt> white is shiner than brown.
[09:18] <lmanul> Oh, that must be it.
[09:18] <lmanul> :)
[09:18] <jdub> ha ha ha
[09:18] <jdub> manu the brown
[09:18] <lmanul> Hi jdub :)
[09:18] <pitti> Good morning
[09:19] <jdub> hey hey pitti 
[09:22] (sivang/#ubuntu-devel) morning
[09:22] <pitti> jdub: hey Jeff, how are your pants today?
[09:23] <thom> pitti: that's a pretty brave question, on balance
[09:23] <thom> morning
[09:23] <jdub> pitti: funny you should ask
[09:23] <pitti> thom: I feel lucky today :)
[09:23] <pitti> thom: good morning, too!
[09:23] <jdub> pitti: i had a string of pants-wearing days
[09:23] <jdub> pitti: and today i found that all my pants needed to go in the wash
[09:23] <pitti> jdub: not too surprising in your winter
[09:23] <jdub> so i am hanging loose today
[09:24] <jdub> pitti: yeah, it is pretty cold here atm
[09:24] <jdub> at least at night
[09:24] <jdub> that's the worst bit; it's reasonable during the day, but cold enough to feel very different at night
[09:24] <pitti> desrt: you cannot write a program that proofs that these bugs don't exist ;)
[09:24] <pitti> desrt: (very similar to the halting problem)
[09:24] <desrt> pitti; you're wrong.
[09:25] <pitti> desrt: however, there are some programs that do static code analysis
[09:25] <desrt> you confuse recognisers with deciders
[09:25] <pitti> usually they can find something
[09:25] <pitti> desrt: but I never played with them, sorry :/
[09:25] <desrt> have any names to drop?
[09:26] <jdub> bad idea: watching the whole billie piper series of doctor who, wondering why you thought her music was so bad, and finding some to listen to.
[09:26] <thom> jdub: define "pretty cold"
[09:26] <thom> jdub: hahaha
[09:27] <desrt> pitti; in any case, you could have a program that outputs either "safe", "unsafe" or "i don't know"
[09:27] <pitti> desrt: http://www.dwheeler.com/flawfinder/ maybe
[09:27] <jdub> thom: it's not all that cold really, but the difference between day and night makes it seem worse
[09:27] <desrt> pitti; and you can improve the chances of getting "safe" or "unsafe" with code annotation
[09:27] <pitti> desrt: right, finding potential overflows is not that difficult
[09:28] <pitti> desrt: however, so far I usually used the tools 'grep' and 'eagle eye' :)
[09:28] <desrt> i was actually sort of thinking more along the lines of the output being one of:
[09:28] <pitti> desrt: I'm pretty interested in automatic tools, just never found the time
[09:28] <desrt> - "i can prove that it is safe."
[09:28] <desrt> - "i can not prove that it is safe."
[09:28] <pitti> desrt: maybe 'I can prove that it is safe against buffer overflows'
[09:28] <desrt> since the people i am working for are definitely interested in hard proof
[09:28] <desrt> they're all about formal methods
[09:28] <pitti> but that will only be the case in rare circumstances
[09:29] <pitti> desrt: heh, rewrite the program in prolog :)
[09:29] <desrt> ok.  maybe i define safe as "will not smash its stack for any user input"
[09:29] <pitti> then it is straightforward to proof
[09:29] <pitti> s/proof/prove/
[09:29] <desrt> i want to do it for C
[09:29] <pitti> (just kidding)
[09:29] <desrt> since it's not even vaguely interesting for other "safer" languages
[09:30] <desrt> and since everyone always uses C anyway
[09:30] <desrt> pfah.  that's garbage.
[09:30] <pitti> desrt: well, since security vulnerabilities happen on many different levels, you cannot really proof 'safety'
[09:30] <Spads> billie piper sings?
[09:30] <desrt> pitti; oh.  i am only interested in the very small sort of definition of safety that i just stated
[09:30] <pitti> desrt: but memory checking is of course a confined domain which should have good tool support
[09:30] <thom> Spads: no. but she produced records
[09:31] <desrt> memory checking implies runtime......
[09:31] <desrt> anyway.  lots to think about
[09:31] <desrt> and 2 years in which to think about it
[09:32] <pitti> desrt: ok, flawfinder and http://www.hp.com/go/cadvise/ are the two in my bookmark list
[09:32] <desrt> i guess i'll probably talk to you again some time in those 2 years :)
[09:32] <desrt> goodnight.
[09:33] <pitti> desrt: sleep well!
[09:48] <lucas> pitti, desrt: there's the stanford checker too
[09:49] <lucas> it has gone commercial now, I think
[09:50] <crimsun> as coverity iirc
[09:52] <lucas> ah yes
[10:09] <bluefoxicy> hmm
[10:10] <bluefoxicy> I didn't destroy my system
[10:10] <bluefoxicy> but I'm not quite sure I actually did anything useful either.
[10:10] <bluefoxicy> https://launchpad.net/distros/ubuntu/+spec/networking-security-sysctls  You be the judge; I'm going to sleep, it's 4am.
[10:15] <fabbione> net.ipv4.conf.all.accept_redirects = 0
[10:15] <fabbione> AHHH
[10:15] <ogra> yay
[10:15] <Kamion> which means I can *finally* start ignoring IRC :P
[10:15] <StevenK> Heh
[10:15] <Hobbsee> Kamion: hah.  never
[10:15] <ogra> how many vacation days did you loose now ? 
[10:15] <Burgundavia> Kamion: go away, you are on holiday
[10:16] <pitti> Kamion: \o/
[10:17] <Kamion> ogra: on aggregate I think about 1-1.5
[10:17] <infinity> Kamion: Awesome job, dude.  Go away.
[10:18] <infinity> Kamion: I owe you some slack cover when you get back.  Just let me know what slavery you need performed.
[10:19] <bluefoxicy> fabbione:  redirects == one computer tells you that packets destined for <host> or <network> should instead be routed through <address>
[10:19] <fabbione> bluefoxicy: i know exactly what a redirect is
[10:19] <fabbione> bluefoxicy: turning it to off is BAD
[10:19] <bluefoxicy> fabbione:  in other words, un-authenticated remote administration of routing tables.  The best "secure" method we have is "look at gateways in our list," which turns into "spoof the default gateway"
[10:19] <bluefoxicy> fabbione:  how is it bad
[10:20] <bluefoxicy> my understanding is that this isn't generally used except in very specific environments
[10:20] <fabbione> bluefoxicy: if you disable it, you can cause a DoS on the network
[10:20] <fabbione> bluefoxicy: it's always used
[10:20] <bluefoxicy> how can you cause a DoS
[10:20] <bluefoxicy> the only nodes I can imagine caring would be ... well... routers
[10:20] <fabbione> host A wants to talk to host B
[10:20] <bluefoxicy> and they typically use a routing information protocol
[10:20] <fabbione> host A default gw C
[10:21] <fabbione> C is a router on the same subnet as D
[10:21] <fabbione> B is connected behind D
[10:21] <bluefoxicy> A -- C -- D -- B ?
[10:21] <fabbione> when i first attempt from A to talk to B, C tells me that D is behind D
[10:21] <fabbione> (with a redirect)
[10:21] <Kamion> infinity: ta. *gone*
[10:21] <fabbione> A C D are on the same subnet
[10:21] <fabbione> BD are on another subnet
[10:21] <bluefoxicy> ah.
[10:22] <fabbione> C is default gw of A
[10:22] <fabbione> C knows where B is due routing table
[10:22] <bluefoxicy> so you're saying
[10:22] <fabbione> ok?
[10:22] <bluefoxicy> A's routing table is misconfigured.
[10:22] <fabbione> no
[10:22] <bluefoxicy> and you want it to magically fix itself.
[10:22] <fabbione> A is perfectly configured host with one default gw
[10:22] <fabbione> A default gw is C
[10:22] <fabbione> C knows that B is behind D
[10:23] <fabbione> ACD are on the same subnet
[10:23] <fabbione> when A tries to talk to B
[10:23] <fabbione>  C will tell A (with a redirect) that B is behind D
[10:23] <bluefoxicy> at the expense that E on the same subnet as A C D can tell A (via spoofed ICMP packet from IP address C) that IT is the choice router for 0.0.0.0/0 (default gw) and thus now have a man-in-the-middle position between A and *
[10:23] <fabbione> becuase it's totally pointless for the traffic to go trough A
[10:23] <bluefoxicy> also
[10:23] <bluefoxicy> subnets were created for a reason.
[10:24] <bluefoxicy> Each subnet should exist in exactly one place; and all routers attached to it should be on its subnet on attached interface.
[10:24] <fabbione> bluefoxicy: if E has access to the physical network, it can do much worst than that.
[10:24] <fabbione> your point is void
[10:24] <fabbione> that setting is bad set to 0
[10:24] <bluefoxicy> you're describing a condition where (A) has everything going through (C), (B) is on another subnet, and apparently nobody told (A) that that subnet is behind (D)
[10:25] <bluefoxicy> which is the proper way to set up the routing table
[10:25] <fabbione> no
[10:25] <bluefoxicy> the subnet (B) is on should not be inaccessible directly by (D) at any point
[10:25] <fabbione> you don't understand the picture i am giving to you
[10:25] <bluefoxicy> if it is your network is wrong.
[10:25] <fabbione> the perfect configured network
[10:25] <fabbione> no errors whatsoever
[10:25] <fabbione> A 192.168.1.100
[10:25] <fabbione> C 192.168.1.1
[10:26] <fabbione> D 192.168.1.2
[10:26] <bluefoxicy> what subnet is behind C?
[10:26] <fabbione> D has also a 10.0.0.2 interface
[10:26] <bluefoxicy> routers have at least 2 interfaces (by definition)
[10:26] <fabbione> C 172.... whatever..
[10:26] <bluefoxicy> okay
[10:26] <fabbione> B is 10.0.0.100
[10:26] <fabbione> ok?
[10:26] <bluefoxicy> then A should know to send 172.* to c and 10.* to D
[10:27] <fabbione> right they do
[10:27] <fabbione> we assume they are talking somekind of IGP
[10:27] <bluefoxicy> and then traffic will never be routed to the wrong router
[10:27] <fabbione> A is a host... remember.. one default gw
[10:27] <fabbione> exactly
[10:27] <bluefoxicy> yes
[10:27] <fabbione> A sends a packets to B
[10:27] <fabbione> the packets hits C
[10:28] <bluefoxicy> normally you don't really have two routers on the same network segment
[10:28] <fabbione> wrong
[10:28] <fabbione> you can have N routers on a network segment
[10:28] <bluefoxicy> wait, yeah... I did.
[10:28] <fabbione> C recog that packets from A can travel trough D directly to reach B
[10:28] <fabbione> C sends a redirect back to A
[10:29] <fabbione> A sends the packets via D
[10:29] <bluefoxicy> but the hosts on that segment had their default routes pointing in whatever direction The Internet was in; and routes to other things (the Wifi AP for one; also the Active Directory server) pointing towards the nearest router in that direction
[10:29] <fabbione> C forgets of that traffic 100%
[10:29] <fabbione> why can't you let me finish before doing other assumptions?
[10:29] <fabbione> the default gw does NOT necessarely points to internet
[10:29] <bluefoxicy> yeah
[10:29] <bluefoxicy> no, but mine did.  I'm saying, my HOSTS each had 2-3 ROUTES
[10:30] <fabbione> ok you are not listening
[10:30] <fabbione> pointless to discuss
[10:30] <bluefoxicy> each host knew about the routers around it, and the routes into the network; they didn't need the routers to tell them about it
[10:31] <bluefoxicy> you're saying that each host is only going to have a default route; no other routes are possible
[10:31] <bluefoxicy> and that routers are going to tell them how to route packets to other routers near them
[10:31] <fabbione> you are putting words in my mounths i didn't say
[10:31] <fabbione> a host needs only a default gw
 A is a host... remember.. one default gw
[10:31] <fabbione> the router will tell the host where to find the other stuff via icmp_redirect
[10:31] <fabbione> it doesn't need
[10:32] <fabbione> more than one route
[10:32] <fabbione> if you disable the icmp_redirect you will see default GW DoSed by ubuntu machines
[10:33] <bluefoxicy> only in network topologies that bank on that set-up
[10:33] <fabbione> and how uncommon do you think they are?
[10:33] <bluefoxicy> bah, I'll have to research it more.
[10:33] <fabbione> you better do
[10:35] <bluefoxicy> s/C/R1/ s/D/R2
[10:35] <bluefoxicy> and you just made word for word Microsoft's same explanation of the issue
[10:37] <fabbione> of course,  you didn't let me finish the explanation without coming up with other stuff
[10:37] <fabbione> the final issue is simple
[10:38] <bluefoxicy> it's a two way channel, I can talk and listen at the same time
[10:38] <fabbione> if host A block redirects from R1 to talk to to B via R2
[10:38] <fabbione> you will endup routing all the traffic from A to B via R1 and R2
[10:38] <fabbione> when there is no real need to involve R1
[10:38] <fabbione> now
[10:38] <bluefoxicy> in fact I can hold multiple lines of conversation with the same person or multiple people at the same time; although when with the same person they consistently have trouble demuxing the conversations from eachother
[10:39] <fabbione> if you want a practical example
[10:39] <fabbione> A is your workstation
[10:39] <fabbione> R1 is your 10Mb asdl router
[10:39] <bluefoxicy> fabbione:  and if your network is designed well in the first place you won't have that problem
[10:39] <fabbione> bluefoxicy: wrong.. again..
[10:40] <quail> Mark Shuttleworth on More4 >>> http://www.channel4.com/more4/news/news-opinion-feature.jsp?id=350
[10:40] <fabbione> StevenK: well ADSL ;)
[10:40] <fabbione> R2 is your gigabit server that act also as rotuer to your 100Mb printer
[10:40] <fabbione> B is the printer
[10:41] <fabbione> disable ICMP_Redirect and print
[10:41] <fabbione> and tell me if you can actually get to ping the adsl router
[10:41] <fabbione> (assuming you print a doc big enough.. like 100MB ps file)
[10:41] <bluefoxicy> lol this is cute
[10:41] <bluefoxicy> ping host B
[10:41] <bluefoxicy> and let it run
[10:41] <fabbione> repeat the above with icmp_redirect turned on
[10:42] <fabbione> (accepting icmp_redirect of course)
[10:42] <bluefoxicy> and it will keep firing pings through router C, and it will keep sending redirects to host A
[10:43] <fabbione> that too
[10:43] <bluefoxicy> fabbione:  that's with redirects on
[10:43] <fabbione> off
[10:43] <bluefoxicy> no, on.
[10:43] <bluefoxicy> Ping doesn't re-initialize its route from the routing table until you restart the program
[10:43] <fabbione> if the redirect is accepted than you won't see all the redirects from C
[10:43] <fabbione> what kernel are you running? 2.0?
[10:44] <bluefoxicy> no, i'm reading a paper
[10:44] <fabbione> the last kernel i saw with such problem was 2.2.0 from RH
[10:44] <fabbione> oh clearly.. test it
[10:44] <fabbione> 2.4 or higher don't behave that way
[10:44] <fabbione> route lookup is performed on a pkt base
[10:44] <bluefoxicy> my network layouts look more like HOST --- ROUTER -- 2 ROUTERS -- OTHER HOSTS
[10:44] <bluefoxicy> or a router with 15 ports on it, etc.
[10:45] <fabbione> that's not the same setup i did describe to you
[10:45] <bluefoxicy> instead of trying to daisy chain routers through switches or hubs
[10:45] <bluefoxicy> the setup you descibed to me was "day 2 of CISCO CCNA academy:  Don't Ever Do This, It's Stupid"
[10:46] <fabbione> ahhaha
[10:46] <fabbione> yeah right
[10:46] <fabbione> ok
[10:46] <bluefoxicy> I wanted to kill justin for his braindead network layout he gave us to implement, it took us 1.5 months to get it half-functional.
[10:46] <fabbione> whatever
[10:47] <bluefoxicy> if he used a design where the routing was actually a backbone instead of a distributed mess mixed in with the rest of the network, it would'a been 3 days.
[10:47] <fabbione> according to your view there can't be more than one router in a subnet with hosts...
[10:47] <bluefoxicy> we kept runnining into issues like (amazingly) hosts having to have 2-3 routing table entries
[10:47] <fabbione> i don't care what you did at your CCNA
[10:48] <bluefoxicy> actually that network was in my implementing and maintaining network intrusion defenses class
[10:48] <bluefoxicy> and my view isn't that you can't do it
[10:48] <bluefoxicy> it's that it's rather close to pouring whiskey into your gas tank
[10:48] <bluefoxicy> it'll work
[10:48] <bluefoxicy> it's still not the best idea.
[10:49] <bluefoxicy> aww... this guy doesn't even address any security issues :(
[10:49] <bluefoxicy> he just talks about how networks reliant on redirects are inefficient and poorly designed.  That's no fun.
[10:49] <bluefoxicy> it's 5am
[10:50] <bluefoxicy> I need sleep.
[10:50] <bluefoxicy> fabbione:  I am pretty certain that any network you give me lain out like that I can do better.
[10:50] <bluefoxicy> and yeah
[10:50] <bluefoxicy> if you have routers forwarding to routers on the same interface you're doubling the bandwidth, which halves your capacity and can lead to network storms
[10:51] <bluefoxicy> but fixing the design will fix that
[10:51] <fabbione> anyway
[10:51] <bluefoxicy> sleep0rz
[10:51] <fabbione> i don't care how you design the net
[10:51] <fabbione> that kind of networks are in use
[10:51] <fabbione> by who.. i don't care
[10:51] <fabbione> it's a real use case
[10:52] <bluefoxicy> there are a lot of 'real use cases' that are both stupid and fixable for a low but significant cost barrier
[10:52] <bluefoxicy> unfortunately i'd be lying if I said I didn't think they mattered
[10:58] <jdub> Kamion: WOOHOO! congrats :-)
[11:23] <carlos> doko: hi
[11:23] <carlos> doko: around ?
[11:23] <doko> carlos: yes
[11:23] <carlos> doko: http://people.ubuntu.com/~carlos/oo-translations-20060808.tar.gz
[11:23] <doko> carlos: ohh, nice.
[11:55] <nags> seb128, hi
[11:55] <nags> seb128, asked Prashanth Mohan to upload
[11:55] <nags> seb128, evolution scripts
[11:55] <seb128> re
[11:55] <seb128> cool
[11:55] <nags> seb128, now they are available with a ready me
[11:55] <nags> seb128, me getting the link
[11:55] <seb128> nice
[11:56] <nags> seb128, Evolution automation scripts are now uploaded here - http://people.freedesktop.org/~prashmohan/evolution/
[11:57] <seb128> nags: thank you for the pointer
[11:57] <nags> seb128, can I convert that avi to ogg - screencast of Evolution automation and Tinderbox integration ? Many people were unable to view the contents properly. They all reported the screen is blank
[11:57] <seb128> nags: probably but I'm not sure of how
[11:58] <nags> seb128, If possible can somebody verify those Evolution scripts and give their feedback, so that we can incorporate them back ? and over a period of time, we can use that in regression ?
[11:58] <nags> seb128, okay, no problem :)
[11:58] <seb128> nags: yeah, that would be nice to get them integrated
[11:58] <nags> seb128, also, I'm working with cr3, I will let him know about the scripts
[11:59] <madduck> thom++ on your blog post
[11:59] <nags> seb128, we are also working on gedit scripts, more than 70 scenarios we have automated, by next week it will be on your table :)
[12:00] <nags> brb
[12:01] <seb128> rock!
[12:27] <siretart> seb128: pong re bug #48138
[12:27] <Ubugtu> Malone bug 48138 in totem "totem-xine crashes when started with no argument" [Unknown,Fix released]  http://launchpad.net/bugs/48138
[12:27] <siretart> seb128: I reassigned it because following the bugtrail, and after looking at the referenced gnome bug, I came to the conclusion that the bug is rather somewhere in totem-xine than libxine
[12:28] <seb128> siretart: just wanted to point that the commit message is a libxine one, it has been reassigned back to xine-lib since ;)
[12:28] <seb128> siretart: I've not copied the first line of the comment from bugzilla which is "Fixed in xine-lib HEAD"
[12:29] <seb128> follow by the "revision 1.30"
[12:29] <siretart> oh. I see. my bad
[12:29] <siretart> checking xine-lib cvs..
[12:32] <ajmitch> hi
[12:35] <siretart> seb128: you're right, my bad. I identified the patch in upstream cvs, I can try to cherry pick that patch to our package. 
[12:35] <seb128> siretart: np, thank you for looking at it, that would be nice ;)
[12:49] <ozamosi> Does anyone here know where I can find documentation about the way ubiquity works?
[12:49] <pitti> RTFS?
[12:49] <Hobbsee> heh
[12:52] <pitti> ozamosi: seriously, you can look into the various specifications on wiki.ubuntu.com
[12:52] <ozamosi> If S means spec, then yes, that's what I'm looking for.
[12:52] <pitti> ozamosi: 'S' usually means 'source', but the specs might be helpful as well, depending on how much detail you need
[12:55] <ozamosi> Of course... When the spec was written, it was called Ubuntu Express... Found it! :)
[01:01] <tseng> ubuntu express was renamed to espresso and then to ubiquity
[01:01] <tseng> to aid in your searching
[01:23] <pygi> hey Hobbsee 
[01:23] <Hobbsee> hi pygi 
[01:24] <rodarvus> guys, I have a pending update for xorg-server (about 16 new patches applied). I'd like to have as much testing as possible, on this new release
[01:24] <rodarvus> it is available on http://people.ubuntu.com/~rodarvus/packages/xorg-server/ (for i386 and amd64)
[01:24] <ajmitch> rodarvus: will test
[01:24] <pitti> rodarvus: grabbing
[01:24] <rodarvus> ajmitch, thanks!
[01:25] <rodarvus> pitti, thanks!
[01:26] <StevenK> Hum.
[01:26] <StevenK> Launchpad is on crack.
[01:26] <ajmitch> StevenK: this is news?
[01:26] <StevenK> The buildd log says a build took 48 seconds, and Launchpad says it took 16 minutes.
[01:27] <StevenK> ajmitch: :-P
[01:27] <rodarvus> StevenK, I'm sure the LaunchPad developers will be thankful to receive your bug report :)
[01:34] <pitti> can anyone please bzr get http://bazaar.launchpad.net/~ubuntu-core-dev/pkg-create-dbgsym/ubuntu , cd dhtest.dbg, and debuild -us -uc -b? (on very latest edgy)
[01:35] <pitti> (or apt-get source pkg-create-dbgsym, works as well)
[01:35] <pitti> building zsh breaks for me, too
[01:38] <rodarvus> pitti, it works - do you want me to upload the resulting .deb somewhere?
[01:38] <pitti> rodarvus: no, not necessary
[01:38] <pitti> rodarvus: suddenly, all packages which build a -dbg package break for me
[01:38] <pitti> rodarvus: since they already have a gnu debuglink, so when dh_strip tries to add its own, it breaks
[01:38] <rodarvus> time to reinstall? :)
[01:39] <pitti> rodarvus: bwah, I reinstalled two days ago :/
[01:39] <rodarvus> ouch
[01:39] <pitti> and it still worked yesterday
[01:39] <pitti> it broke after today's dist-upgrade
[01:39] <rodarvus> I reinstalled my main desktop yesterday
[01:39] <rodarvus> too much breakage on it after building/testing X.Org 7.1
[01:39] <pitti> rodarvus: do you have libc 2.4-1ubuntu7?
[01:40] <pitti> it's the only package that is remotely toolchanin
[01:40] <rodarvus> pitti, yes
[01:40] <pitti> although I don't see how libc6 would have influence on the debuglink
[01:40] <pitti> hm
[01:40] <pitti> my box is clearly screwed then
[01:40] <pitti> rodarvus: thanks!
[01:40] <rodarvus> np
[01:40] <Hobbsee> pitti: i killed it last night.  sorry about that :P
[01:42] <Hobbsee> pitti: the plan of world domination is succeeding.
[01:44] <rodarvus> pitti, wait
[01:44] <pitti> rodarvus: please tell me I'm not on crack
[01:44] <rodarvus> building the code from bzr fails for me too, while building from the source in the archives works
[01:45] <pitti> rodarvus: are you sure? they are identical
[01:45] <rodarvus> yup
[01:45] <Hobbsee> good pitti 
[01:45] <rodarvus> pitti, pasted on /msg window
[01:45] <pitti> rodarvus: right, I get the same error here
[01:45] <pitti> rodarvus: bug diff -Nur' in the two directories shouldn't give any difference, doesn it?
[01:46] <rodarvus> let me check
[01:46] <pitti> can anyone else on *latest* dapper please try to build zsh and see whether dh_strip aborts with an "Invalid operation"?
[01:47] <mvo> pitti: I can do this, give me a bit
[01:47] <pitti> rodarvus: I get the same fault from the archive version
[01:47] <pitti> mvo: or, any package that builds a -dbg package, doesn't matter
[01:47] <pitti> zsh just was the smallest example I could find in 30 seconds
[01:48] <mvo> pitti: I have some outstanding upgrades (~10), I do them now and will build then
[01:48] <rodarvus> pitti, same stuff on both
[01:48] <pitti> mvo: can you please try both before and after?
[01:48] <pitti> mvo: is libc amongst the outstanding ones?
[01:49] <mvo> pitti: no, looks like its only gstreamer stuff
[01:49] <pitti> mvo: just build dhtest.dbg in pkg-create-dbgsym then, it's faster
[01:49] <pitti> mvo: oh, ok
[01:49] <ajmitch> rodarvus: this xserver appears to work ok on i386 - at least no issues in the limited testing I've done
[01:50] <rodarvus> ajmitch, thanks
[01:50] <rodarvus> ajmitch, switching to vt1 works as usual (considering vt switching is working for you currently)
[01:50] <ajmitch> yes
[01:50] <rodarvus> hooray
[01:52] <rodarvus> pitti, the machine you are going to test xorg-server on is an amd64?
[01:52] <pitti> rodarvus, mvo: bingo! after downgrading libc6 it works again
[01:52] <pitti> rodarvus: yes
[01:53] <rodarvus> nice
[01:54] <mvo> pitti: that means I can stop building zsh?
[01:54] <pitti> mvo: yes, I think so
[01:57] <mvo> pitti: it seems to have build for me, what version of glibc do you have installed?
[02:00] <pitti> mvo: 2.4-1ubuntu7
[02:00] <pitti> mvo: it works with ubuntu6 and fails with ubuntu7
 can anyone else on *latest* dapper please try to build zsh and see whether dh_strip aborts with an "Invalid operation"? <- you meant latest edgy?
[02:02] <pitti> erm, of course
[02:02] <pitti> sorry
[02:02] <mvo> :)
[02:02] <mvo> ok, that explains why it build here :)
[02:03] <pitti> mvo, rodarvus: bug 55880, in case you can add some confirmations
[02:03] <Ubugtu> Malone bug 55880 in glibc "libc6 2.4-1ubuntu7 adds gnu debuglinks by default" [Critical,Unconfirmed]  http://launchpad.net/bugs/55880
[02:08] <pitti> rodarvus: no noticeable difference with the new x server
[02:08] <rodarvus> yaya
[02:08] <rodarvus> must mean this version is perfect
[02:08] <rodarvus> I'll upload it right away
[02:08] <rodarvus> pitti, thanks!
[02:08] <ogra> bugfree and all ?
[02:08] <rodarvus> ogra, X never has bugs
[02:09] <rodarvus> its always fault of the kernel, gnome/kde, etc
[02:09] <ogra> yeah, tell that to malone :)
[02:09] <rodarvus> ogra, all X bugs are there because I'm lazy to properly redirect them to their real owners :P
[02:10] <seb128> what should be done with bugs like bug #55865 ?
[02:10] <Ubugtu> Malone bug 55865 in Ubuntu "[Edgy]  Please remove obsolete libnautilus-burn3 from archive" [Untriaged,Unconfirmed]  http://launchpad.net/bugs/55865
[02:10] <ogra> its like ltsp ... there are only network hardware probs, but no bugs :P 
[02:10] <seb128> reassigning to ubuntu-archive?
[02:11] <Hobbsee> seb128: can it.  user was wrong.
[02:11] <Hobbsee> s/can/traschcan/
[02:11] <slomo> seb128: ubuntu-archive only wants to be subscribed iirc... but this one can maybe even be rejected... they will remove the old binaries anyway next time
[02:12] <pitti> seb128: not reassign, subscribe it
[02:15] <seb128> pitti: it's already subscribed
[02:15] <pitti> rodarvus: hm, another nice test - the build should fail since it builds a -dbg :)
[02:15] <seb128> pitti: but I don't want to take the bug and I don't want to let it unassigned
[02:15] <rodarvus> heh
[02:15] <pitti> seb128: unassigning should work as long as -archive is sub'ed
[02:16] <seb128> pitti: right, but the bug stays as noise for bug triagers then
[02:17] <pitti> seb128: well, it should disappear automatically anyway...
[02:17] <pitti> it's the standard NBS case
[02:24] <Riddell> pitti: builds with -dbg fail now?
[02:25] <pitti> Riddell: yes, see bug 55880
[02:25] <Ubugtu> Malone bug 55880 in glibc "libc6 2.4-1ubuntu7 adds gnu debuglinks by default" [Critical,Unconfirmed]  http://launchpad.net/bugs/55880
[02:25] <Riddell> pitti: ok, that'll be all the KDE packages needing changed then :)
[02:26] <pitti> Riddell: no, don't
[02:26] <pitti> Riddell: either we'll fix it in glibc or in debhelper/pkg-create-dbgsym
[02:26] <Riddell> yeah, just read the bug
[02:28] <pitti> Riddell: I wait until jbailey wakes up to discuss it with him
[02:28] <pitti> I didn't find anything in the glibc changelog
[02:50] <cprov> who is Gauvain Pocentek <gauvainpocentek@ubuntu.com> ?
[02:52] <ogra> Gloubiboulga
[02:52] <cprov> ogra: tks
[02:53] <ogra> btw, using your own tools helps ;) https://launchpad.net/people/gauvainpocentek
[02:53] <Gloubiboulga> bug 55795
[02:53] <Ubugtu> Malone bug 55795 in launchpad "replaces Debian maintainer by Ubuntu maintainer in changelog" [Untriaged,Unconfirmed]  http://launchpad.net/bugs/55795
[02:53] <ogra> Gloubiboulga, thats a duplicate 
[02:53] <ogra> err, oh
[02:54] <ogra> ignore me ...
[02:54] <StevenK> How the hell is that a bug?
[02:55] <ogra> thats no bug
[02:55] <ogra> the one who made the last change has to show up in the changelog
[02:55] <StevenK> Like I said.
[02:56] <desrt> erm
[02:56] <desrt> that probably is a bug.
[02:56] <ogra> desrt, ?
[02:56] <desrt> he took the upstream changelog entry directly and changed only the name
[02:56] <desrt> if he changed other stuff to he should have done
[02:56] <desrt>  [ Changes by _____ ] 
[02:57] <desrt> then added his own changes
[02:57] <desrt> *too
[02:57] <ogra> * Resynchronise with Debian. 
[02:57] <ogra> thats enough if a package has ubuntuX versions before and after the merge
[02:57] <desrt> that's very old
[02:58] <desrt> there shouldn't be an ubuntu maintainer's name on a non ubuntu version, though
[02:58] <ogra> stillflame, it has ubuntu specific changes, so the name of the person making these changes is the right one for the changelog
[02:58] <desrt> they should do the synch, preserving the debian changelog and then have their name on the change -after- the synch
[02:58] <ogra> right, if it would have been synced that wouldnt happen at all
[02:58] <ogra> but it was maerged
[02:58] <ogra> *merged
[02:59] <ogra> so thats perfectly right
[03:01] <desrt> well.  seems fishy to me.
[03:03] <ogra> it names the person who was responsibe for the manual merging between the last and the current ubuntuX version ... 
[03:03] <desrt> the merging occured as work going into the ubuntu1 version and should be listed there
[03:03] <desrt> it wasn't done as work going into the upstream version (2.0.7-3) and therefore should _not_ be listed there
[03:04] <desrt> (and, in fact, it isn't listed there at all -- only the name has changed)
[03:04] <ogra> you mean it should have two different entries ? 
[03:04] <desrt> i would imagine so
[03:04] <desrt> considering it already has what appears to be two somewhat broken entries
[03:04] <desrt> discover (2.0.7-3ubuntu1) edgy; urgency=low
[03:04] <desrt> discover (2.0.7-3) unstable; urgency=low
[03:05] <desrt> to most people this doesn't look like a single changelog entry.  i think that's the problem.
[03:06] <desrt> if he was doing that it should read like
[03:06] <desrt> discover -3ubuntu1
[03:06] <desrt>   *merge
[03:06] <desrt>     -list of
[03:06] <desrt>     -changes from
[03:06] <desrt>     -debian
[03:06] <jdub> rodarvus: 003_fedora_root_window_black_pattern.patch -> we already get that by doing -br in gdm.conf
[03:06] <ogra> Gloubiboulga, does xubuntu have no pointrelease ? http://cdimage.ubuntu.com/xubuntu/re...6.1/release.1/ is mentioned in the announcement, but doesnt seem to exists
[03:06] <desrt> -- guy <foo@ubuntu.com>
[03:06] <jdub> rodarvus: better to use configuration than patching :-)
[03:07] <rodarvus> jdub, nice, thanks for the heads up :)
[03:07] <rodarvus> I'll remove it in the next package upload
[03:07] <rodarvus> jdub, xubuntu and kubuntu do it in their windowmanagers too?
[03:07] <ogra> desrt, that would extend the merge work a lot ... we rely on dpkg-genchanges -v doing the right thing usually
[03:07] <rodarvus> (... do you know if...) :)
[03:07] <jdub> rodarvus: no idea
[03:08] <Gloubiboulga> ogra, http://cdimage.ubuntu.com/xubuntu/releases/dapper/release.1/
[03:08] <ogra> instead of reformatting the whole changelog :)
[03:08] <jdub> what does xubuntu use for it's display manager anyway?
[03:08] <desrt> jdub; gdm
[03:08] <ogra> Gloubiboulga, hmm, the link in the announcement is wrong then
[03:08] <desrt> jdub; with a nice blue theme
[03:08] <ogra> http://cdimage.ubuntu.com/xubuntu/releases/6.06.1/release.1/
[03:08] <jdub> rodarvus: so there's your answer for xubuntu; check kdm for kubutu ;-)
[03:08] <jdub> desrt: thanks ;)
[03:08] <rodarvus> doesn't gdm requires the gnome stack to be installed?
[03:09] <jdub> YAYAYAY CAIRO!
[03:09] <rodarvus> I should install xubuntu one of these days
[03:09] <Riddell> rodarvus: yes (kubuntu does)
[03:09] <jdub> rodarvus: yeah, care of gnome-canvas
[03:09] <robertj> can anyone suggest a way to be more...spontanious? my /random is running short :(
[03:10] <robertj> and the mouse isn't in easy reach
[03:10] <Spads> lrwxrwxrwx 1 root root 7 2006-08-10 14:09 /dev/ass -> urandom
[03:10] <rodarvus> jdub, desrt, Riddell: thanks again!
[03:11] <Gloubiboulga> rodarvus, gdm only depends on gconf IIRC
[03:11] <Gloubiboulga> but not on all the gnome libs
[03:12] <ogra> gdm on gconf ? 
[03:12] <ogra> what for ? 
[03:13] <sladen> jdub: sadly cairo is not Yeah, Yeah, Yeah, good.  If it were, people like Inkscape would be wanting to use it in preference to stucking with libart.  See: http://www.xaraxtreme.org/about/performance.html for how much, er, "potential" and loving cairo needs to receive.
[03:13] <jdub> sladen: i know all about cairo's issues; i am merely celebrating 1.2.2 being in the repos
[03:13] <ogra> it depends on libgnomecanvas2-0 and thus on gnome ...
[03:14] <jdub> ogra: though i see now that gnomecanvas has tidied up depends now
[03:14] <Gloubiboulga> ogra, err, you're right
[03:15] <ogra> jdub, right ... but enough to not dep on gnome anymore ? 
[03:15] <jdub> ogra: look at libgnomecanvas2-0 depends
[03:15] <ogra> yeah, i'm just doing that ... nice ! :)
[03:15] <jdub> ogra: schmick compared to what it used to be like
[03:24] <lool> hmm is dholbach on holidays?
[03:24] <lool> looks like it
[03:25] <pitti> hi lool 
[03:27] <lool> hi pitti!
[03:49] <seb128> lool: yep, until the wiesbaden sprint in 10 days
[03:50] <lool> seb128: thanks
[03:50] <Hobbsee> hi heno 
[03:50] <spike> hi
[03:50] <seb128> np
[03:50] <spike> I was in the need of lndir, and it's not in the xutils pkgs, is there a reason for that?
[03:50] <spike> it's in the debian one
[03:51] <spike> and an apt-file search lndir will find some manpages references
[03:52] <spike> any idea?
[03:52] <azeem> spike: it got lost during the X modularization
[03:54] <spike> azeem: I see. and it's definitely lost or what? should I raise a bug if one doesnt exist already?
[03:55] <azeem> spike: I suggest filing a wishlist for moreutils in Debian to include it (and point them to the source or attach it)
[03:55] <rodarvus> spike, lndir is part of xutils-dev
[03:55] <rodarvus> which is present on Ubuntu Edgy.
[03:55] <rodarvus> rodarvus@wakko:~$ dpkg -L xutils-dev | grep lndir
[03:55] <rodarvus> /usr/share/man/man1/lndir.1x.gz
[03:55] <rodarvus> /usr/bin/lndir
[03:55] <rodarvus> rodarvus@wakko:~$
[03:57] <azeem> seems packages.u.c is broken WRT contents and edgy
[03:57] <spike> rodarvus: uhm, how does it come apt-file search wont find it? nor http://packages.ubuntu.com/
[03:57] <azeem> "You have searched for the contents of xutils-dev in edgy, architecture i386.
[03:57] <azeem> Can't find that package, at least not in that distribution and on that architecture"
[03:57] <rodarvus> spike, *shrugs*
[03:57] <rodarvus> maybe it hasn't been updated for new packages in edgy?
[03:58] <rodarvus> xutils-dev is a new package
[03:58] <spike> oh, nm, my bad, I was thinking of dapper
[03:58] <spike> indeed, sorry
[04:02] <heno> Hobbsee: hello :)
[04:02] <Hobbsee> :)
[04:05] <nags> seb128, one help regarding libwnck, can I ask you ?
[04:05] <seb128> nags: sure
[04:05] <icecrash> moin
[04:06] <nags> seb128, copied gok-windowlister.c and trying to write a small program, to bring a bg window to fg
[04:06] <nags> seb128, Casanova tried this - http://pastebin.ca/125165
[04:07] <nags> seb128, we both are getting segfault in screen.c
[04:07] <nags> seb128, #0  0xb7ecd4ba in wnck_screen_get_default () at screen.c:468
[04:08] <seb128> nags: why do you fflush(stdout); every time you printf something? just curious
[04:08] <nags> wnck_screen_get_default () actually calls _wnck_event_filter_init ();
[04:08] <nags> Casanova, ^^^^ ?
[04:08] <Casanova> seb128: i just wanted to know where it segfaulted :D
[04:09] <seb128> Casanova: if you printf("message\n"); that should be enough, no need to fflush
[04:09] <Casanova> seb128: hmm well it wasnt printing :)
[04:13] <seb128> Casanova: run gtk_init before calling refreshList
[04:13] <seb128> nags:
[04:13] <Casanova> oh ok
[04:14] <nags> seb128, okay
[04:15] <seb128> or gdk_init
[04:15] <seb128> should be enough for that
[04:18] <nags> seb128, okay
[04:22] <nags> seb128, Thanks, it works :)
[04:22] <seb128> np
[04:26] <Casanova> seb128: it still crashed on wnck_window_get_workspace (window_list_entry->data);
[04:26] <Casanova> any idea why?
[04:27] <Casanova> oops
[04:27] <seb128> Casanova: doesn't for me
[04:27] <Casanova> seb128: sorry my fault :)
[04:28] <seb128> Casanova: are you sure window_list_entry != NULL?
[04:28] <seb128> k
[04:33] <LisaH> Hi....My name's Lisa and I'm writing an article for NewsForge.com about today's 6.06.1 release. I have a couple of questions for the developers if one of you has a minute.
[04:35] <Hobbsee> hi LisaH 
[04:39] <iwj> Feh, it trashes the homepage setting.  Why ?
[04:39] <Kamion> ogra: fixed the xubuntu point release link to work; thanks
[04:40] <ogra> :)
[04:40] <ogra> Kamion, that really shouldnt have been your task
[04:40] <Kamion> (waiting for mirrors to sync; mirnyy's being slow)
[04:40] <ogra> go away from IRC !!
[04:40] <Kamion> *shrug* it was just a missing symlink
[04:40] <ogra> :)
[04:41] <Hobbsee> Kamion: excuses excuses.  now take a holiday :P
[04:41] <Kamion> don't worry, I am
[04:41] <Kamion> IRC != work
[04:41] <Hobbsee> Kamion: surely that depends as to whether it's on or offtopic?
[04:50] <infinity> LisaH: Are you looking for an interview with someone, PR fluff, or just to clear up a few random points?
[04:56] <pitti> LisaH: might be best to just ask the questions here
[04:57] <LisaH> Thanks...I've got a few chats going on at the moment but I may be back here for some further questions. 
[05:15] <pitti> BenC: ah, cool; from looking at the git commits, I only saw pid and signal number passed to the crash dump helper
[05:16] <BenC> pitti: You probably saw the case where the core size is greater than the apport temp size, there's a second part where the helper gets called with the core file name
[05:18] <pitti> BenC: hm, I looked at both in http://www.kernel.org/git/?p=linux/kernel/git/bcollins/ubuntu-2.6.git;a=blob;h=45b34f79920db4bbe3ae05711dc18a9c5161fe0b;hb=6955793faeb05f4a73ec25b7b59e9d5e036a7793;f=fs/exec.c
[05:18] <pitti> BenC: lines 1514 and 1572
[05:18] <pitti> BenC: aaah, now I see
[05:18] <pitti> must have been bling
[05:18] <pitti> blind, too
[05:18] <pitti> BenC: sorry for the fuss!
[05:19] <BenC> np :)
[06:08] <Keybuk> so, what methods of intelligent design have people followed that worked for them?
[06:08] <Keybuk> sfllaw: you mentioned that there are sane ways?
[06:09] <doko> Keybuk, infinity, Kamion, mdz: please do not process the new binaries from the gcj-4.1 source until it's built on all architectures
[06:09] <pitti> Keybuk: the only method that really worked for me was iterative design while comparing to a wide range of use cases
[06:09] <Keybuk> doko: normally we don't anyway
[06:09] <Keybuk> due to bugs in Soyuz
[06:10] <infinity> \o/
[06:10] <infinity> That's going to be a painful week.
[06:10] <Keybuk> infinity: let Lamont do it <g>
[06:10] <sfllaw> Keybuk: The parts of upfront design that are useful are those that define the problem.
[06:10] <Keybuk> "your toy architecture, your flaming beige plastic anal violation"
[06:10] <infinity> Nah, then I wouldn't have anything to complain about.
[06:10] <infinity> Besides, it's my toy arch too.  Or one of them.
[06:11] <sfllaw> So requirements specs, like "what problems need to be solved".
[06:11] <sfllaw> That's actually really hard.
[06:11] <sfllaw> Because most people think of solutions before problems.
[06:11] <sfllaw> Which is where you get into trouble.
[06:11] <Keybuk> sfllaw: yeah, that's the part I really enjoy
[06:11] <Keybuk> my problem comes when it comes to the actual coding
[06:11] <Keybuk> I find if I've spent too long designing it, especially at the code level, I get a kind of "coder's block" and can't actually write it
[06:11] <sfllaw> When it comes to coding, if you've thought which problems need solving, you don't need to specify the implementation too much.
[06:12] <sfllaw> Lots of what is fun about coding is encountering a problem and then solving it on the fly.
[06:12] <Keybuk> Matt was advocating a detailed advanced planning thing this time around, so that time estimates would be possible
[06:12] <Keybuk> where you break the code down into very small tasks, and work them out by the hour how long they take
[06:12] <sfllaw> If you pull the problems into the specification process, you can hold them all in your head at the same time.
[06:12] <sfllaw> Oh.
[06:12] <sfllaw> That's really hard.
[06:13] <sfllaw> The best I can find for scheduling is Joel's Painless Software Scheduling.
[06:13] <sfllaw> But even that gives you slack.
[06:13] <mjg59> Keybuk: What's responsible for generating the /dev/disk/by-label directory?
[06:13] <Keybuk> mjg59: udev rules (65-persistent-storage.rules)
[06:13] <sfllaw> I don't like implementation specs too much.
[06:13] <Keybuk> mjg59: it uses vol_id to find the label of the disk
[06:14] <sfllaw> I don't even like functional specs, except for specifying infrastructure and API/ABIs.
[06:14] <mjg59> Keybuk: Ok. If I create a partition, how do I provoke it into appearing?
[06:14] <sfllaw> That's because interoperability is sometimes valuable.
[06:14] <Keybuk> mjg59: normally it should just appear when the kernel sends out the uevent
[06:14] <Keybuk> though the kernel is not always good at this
[06:14] <mjg59> Keybuk: I've done a mkswap /dev/hda3 -L swap
[06:14] <Keybuk> ahh
[06:14] <Keybuk> right
[06:14] <mjg59> And blkid
[06:14] <Keybuk> block device != partition
[06:14] <mjg59> But no by-label
[06:14] <Keybuk> you mean filesystem
[06:14] <mjg59> Yes, sorry
[06:15] <Keybuk> echo -n add > /sys/block/hda/hda3/uevent
[06:15] <mjg59> Ok, that works
[06:15] <mjg59> Next issue
[06:15] <Keybuk> it's a long-standing bug of mine that the kernel deals with block devices not partitions ;)
[06:15] <Keybuk> gah, FILESYSTEMS
[06:15] <mjg59> If I have a vfat partition, and then mkswap it, libblkid seems to find the fat UUID rather than the swap one
[06:15] <Keybuk> yes
[06:16] <Keybuk> vol_id has a similar problem with fat/ext3 I believe
[06:16] <mjg59> I'm not sure if that's helpful, though I guess it doesn't break anything
[06:16] <Keybuk> partman deliberately zeros the partition bits first
[06:16] <Keybuk> right, it won't break anything, it just means things are ... wrong ;)
[06:16] <mjg59> And thirdly:
[06:16] <mjg59> blkid doesn't seem to find my swap uuid/label when it contains a suspended image
[06:17] <Keybuk> indeed
[06:17] <Keybuk> blkid is rubbish
[06:17] <mjg59> So resume from disk is broken right now
[06:17] <Keybuk> at the sprint, I'm going to sit down with tollef and get mount to use vol_id instead
[06:17] <mjg59> Presumably it just needs teaching about where they live?
[06:17] <Keybuk> yeah, but it's less effort to just do the vol_id change
[06:17] <mjg59> Mount's not the issue here
[06:17] <Keybuk> SuSE already patched it
[06:17] <Keybuk> oh, what is the issue?
[06:17] <mjg59> It's a swap partitoin
[06:17] <Keybuk> s/mount/util-linux/
[06:18] <mjg59> Ah, right
[06:18] <Keybuk> they all use the common code
[06:18] <mjg59> Yes, that makes more sense
[06:18] <Keybuk> sorry
[06:18] <Keybuk> there are other issues anyway, like blkid doesn't seem to support fat uuids
[06:18] <mjg59> I'll look at blkid anyway, since I'd like this to work before then
[06:18] <mjg59> I'm playing with uswsusp. It seems to have the potential to be quite a bit faster than swsusp.
[06:18] <mjg59> (Back in a moment, need to change my buffers)
[06:21] <mjg59> Ok
[06:36] <zul> hi
[06:48] <mjg59> Keybuk: Turns out to be a trivial patch, by the looks of things
[06:49] <Keybuk> unsurprising
[06:49] <Keybuk> the vol_id patch was tiny also
[06:49] <mjg59> It actually deals with the standard image, just not the userspace one...
[06:49] <mjg59> Testingnow
[06:55] <nags> seb128, I have some queries with respect to libwnck, whom shall I ask ?
[07:02] <mjg59> Keybuk: Yeah, fixed
[07:02] <mjg59> 4 extra lines
[07:03] <mjg59> I'll upload now
[07:03] <elmo> hey, my gnome thinks my machine is at 2% battery when it has 2 hours remaining - how do I kick it to uncache those values
[07:03] <elmo> +?
[07:05] <pitti> elmo: try /etc/dbus-1/event.d/20hal restart
[07:05] <elmo> pitti: it's been like this for months
[07:05] <elmo> pitti: so I can't imagine restarting anything'll help
[07:06] <elmo> the problem is when I first installed ubuntu on the machine, the battery was fried (I'd done a partial 'drain-recharge' cycle in the BIOS), so back then, it really did only last a couple of minutes
[07:06] <pitti> oh
[07:12] <mjg59> Keybuk: Hm. So udev uses libvolid to find swap labels and stuff?
[07:13] <Keybuk> mjg59: right
[07:13] <mjg59> Ok
[07:13] <Keybuk> fix which?
[07:13] <Keybuk> I may have already fixed whatever you're planning to fix there
[07:14] <mjg59> Keybuk: volume_id doesn't appear to detect my suspended swap partition
[07:14] <Keybuk> mjg59: the edgy one?
[07:14] <Keybuk> should do
[07:14] <mjg59> Yeah
[07:14] <mjg59> Hm
[07:14] <Keybuk> there's definitely a patch for that in there, and I tested it a bit
[07:15] <mjg59> Keybuk: It needs to recognise ULSUSPEND as well as S1SUSPEND
[07:15] <Keybuk> aha!
[07:15] <Keybuk> yes, fix that
[07:15] <Keybuk> just modify the existing patch
[07:15] <mjg59> Oh god what's the right way to do this?
[07:16] <Keybuk> debian/patches/60-vol_id-swap.patch
[07:16] <Keybuk> add
[07:16] <mjg59> Patch it without that patch, add the patch, generate the diff, replace the original patch?
[07:16] <Keybuk> oh
[07:16] <Keybuk> yeah
[07:16] <Keybuk> I just apply the patch, tinker with it, and diff against the orig
[07:19] <Keybuk> so I do things the very hard way
[07:19] <mjg59> http://www.codon.org.uk/~mjg59/tmp/60-vol_id-swap.patch look ok?
[07:19] <Keybuk> missing a == 0) in there :p
[07:20] <Keybuk> the ) is especially important
[07:20] <mjg59> Oops
[07:20] <mjg59> Try again?
[07:20] <Keybuk> looks right
[07:20] <mjg59> Ok, testing now
[07:43] <sladen> rodarvus: can you remember what the current work around for "cannot find font fixed is" (bug #54809) ?
[07:43] <Ubugtu> Malone bug 54809 in Ubuntu "X server cannot find default font `fixed'" [Untriaged,Confirmed]  http://launchpad.net/bugs/54809
[07:46] <gnomefreak> yes ;)
[07:47] <gnomefreak> sladen: remove --purge xserver-xorg-core xorg and xserver-xorg  than install them it should fix it (atleast it did for a few of us)
[07:48] <ogra> sladen, run mkfontdir in the misc font directory fixes it
[07:49] <gnomefreak> of course :( you could take the easy way :(
[07:49] <sladen> ogra: no, that doesn't fix it here.
[07:50] <ogra> oh ? 
[07:50] <ogra> thats intresting then 
[07:50] <sladen> ogra: and even if I fix things to point to the /right/ directory it doesn't fix it
[07:51] <ogra> well, its not broken here currently ... i built a ltsp chroot today that starts X just fine ... hmm
[07:51] <sladen> gnomefreak: nope, apt-get --purge --reinstall install ...  didn't help
[07:51] <gnomefreak> sladen: no dont use --reinstall
[07:51] <sladen> gnomefreak: thanks for the pointer though
[07:51] <ogra> sladen, /usr/share/fonts/X11/misc/fonts.dir exists ?
[07:51] <gnomefreak> for some reason --reinstall doesnt fully do what you would like it to do
[07:52] <sladen> gnomefreak: well it doesn't remove the dependancies, which if it does work means it's one of those that is broken
[07:52] <gnomefreak> sladen: i couldnt get --reinstall to work with that i had to remove --purge than apt-get install and it fixed it
[07:53] <gnomefreak> it removes the 3 -desktop packages and something else (cant think of it offhand) with the X packages
[07:58] <treitter> what creates "md5sums" in control.tar.gz ?
[07:58] <treitter> dpkg, or do I do it?
[07:59] <pitti> treitter: dh_md5sums
[07:59] <pitti> treitter: it's not essential (only used by debsums), but still nice to have
[07:59] <treitter> pitti: invoked directly, or does something else call that?
[08:00] <pitti> treitter: cdbs calls it automatically, with just debhelper it's your job AFAIK
[08:00] <treitter> pitti: now I have to look up cdbs :)
[08:01] <treitter> ah. cdbs requires Makefiles, right?
[08:01] <treitter> my project doesn't use Makefiles (at least when I pick it up and package it)
[08:01] <pitti> treitter: not any more or less than plain debhelper packages
[08:02] <pitti> treitter: don't worry about it if you don't know cdbs or don't want to use it
[08:02] <treitter> pitti: cool. Thanks :)
[08:04] <treitter> ..so dpkg requires all the control files in ${PKG}/DEBIAN, while dh_md5sums requires them in ${PKG}/debian
[08:06] <treitter> hmm. I don't think dh_md5sums is ideal for what I'm doing. I'll just use md5sum directly
[08:23] <Burgwork> sabdfl, I love it when you fix my bugs and then blog about them
[08:25] <sladen> ogra: ...so, half the places refer to 'X11/fonts' and the other half to 'fonts/X11', meaning that the fonts end up in one, but the aliases that maps 'fixed' end up in the other...
[08:25] <mdke> Burgwork: that happens a lot?
[08:26] <Burgwork> mdke, no, not really, but the essential people one is my bug from UBZ, when we had people not being at spec bofs because of the lack of it
[08:27] <mdke> oh, that post, nice
[08:28] <doko> pitti: will you upload glibc tonight?
[08:28] <pitti> doko: yes, after pkg-create-dbgsym is in the archive
[08:39] <mjg59> Whee. That works.
[08:40] <mjg59> Now it just needs some usplash loving.
[08:41] <Surak> when there are more than one video board plugged on a machine, how does hwdata chooses which one it should use?
[08:48] <Surak> I have a particular case where Xorg.0.log says: "(II) Primary Device is: PCI 01:00:0", but the Live CD tries to use the other video card - which is disabled in bios, but still appears in device list. It appears first, but it's not necessairly the primary one....
[08:57] <Surak> sorry, I mean "discover", from discover1 package, not hwdata.
[09:04] <Keybuk> init:stop_my_job: Timer to stop job called
[09:04] <Keybuk> init: Goal change: stop test
[09:04] <Keybuk> init:job_kill_process: Sending TERM signal to test process (27481)
[09:04] <Keybuk> init:my_reaper: Child 27481 killed by Terminated
[09:04] <Keybuk> init: test process (27481) killed by signal 15
[09:04] <Keybuk> init: State change: test: running to stopping
[09:04] <Keybuk> \o/
[09:06] <sladen> Surak: is there a bug report for it, can you file one?
[09:07] <sladen> go Keybuk, now --fomg-optimize is so that the gentoo wannabe's quit moaning :)
[09:08] <Keybuk> heh
[09:09] <Surak> sladen: I was asking before, so I would post the bug more concisely. I found that /sbin/discover ignores which video board is the primary (as noted in Xorg.0.log), it only lists them in the same order as lspci does. However, would this be a bug from discover or from xserver-xorg's configure script?
[09:10] <Surak> g/as/that
[09:18] <mdz> mjg59: what's uswsusp?
[09:20] <sladen> mdz: userspace suspend
[09:21] <sladen> mdz: suspend for nice people, without the gash
[09:24] <mjg59> mdz: Suspend to disk with compression and potentially graphical niceness
[09:27] <Burgwork> mdz, just realized something about the .1 release announcement. We shoudl make it very clear that merely upgrading will achieive the same result
[09:36] <nate> Is the CD-Images rsync down?
[09:57] <mdz> Burgwork: it's already gone out
[10:01] <Burgwork> mdz, yep, I realize that. But for the .2 release notes
[10:01] <treitter> is there a way to have dpkg-deb ignore specific files (like .cvs or .svn dirs), or is it expected that you'll remove them yourself as necessary?
[10:03] <sladen> mdz: what's the best way of dealing with packages that have gone from '6.8.2.1-5' to version '1.0.0-6ubuntu3' ?  The downgrade is causing upgrade problems
[10:03] <zul> http://70.29.57.2/ubuntu/xen-edgy-within-edgy.png
[10:03] <Spads> sweet
[10:04] <sladen> zul: excellent news, shall we fridge it :)
[10:04] <crimsun> sladen: err, that should have an epoch. Which packages?
[10:04] <sladen> crimsun: xfonts-utils
[10:04] <zul> sladen: if you want maybe i should clean up my desktop first though :)
[10:05] <rodarvus> sladen, the right is 1:1.0.0-foo
[10:05] <rodarvus> (or, "use epochs" :) )
[10:05] <crimsun> that's odd, since it /does/ have an epoch
[10:05] <sladen> zul: I can't see anything dodgey on there, I don't /think/
[10:05] <Spads> zul: win32!
[10:05] <zul> i have a couple of torrents on my desktop i think
[10:06] <rodarvus> sladen, are you sure you didn't forgot to add the epoch to your last changelog entry?
[10:06] <rodarvus> most likely the case here
[10:06] <crimsun> rodarvus: (meaning to me? It's present.)
[10:06] <sladen> rodarvus: I haven't done an upload, this is the result of doing an dapper->edgy upgrade
[10:07] <rodarvus> crimsun, well, or you :) (/me didn't read the backlog)
[10:07] <crimsun> I'm quite sure, since I used dch -i and double-checked before uploading
[10:07] <rodarvus> xfonts-utils | 1:1.0.0-6ubuntu3 | http://archive.ubuntu.com edgy/main Sources
[10:07] <rodarvus> this is what I have here
[10:08] <crimsun> sladen: what is dpkg spitting out?
[10:09] <mdz> sladen: that's not a downgrade, that's a change in upstream numbering (hence the epoch)
[10:09] <mdz> if the upgrade isn't working properly, it's a bug
[10:09] <sladen> crimsun: the epooch appears to be there in your upload
[10:10] <crimsun> right. With what errors is dpkg bailing?
[10:11] <mdz> treitter: by the time you reach dpkg-deb, anything you don't want should have been excluded earlier
[10:11] <mdz> treitter: dpkg-deb just packs up the installed tree; if you don't want those files there, don't install them
[10:12] <treitter> mdz: yeah, that's what I thought. My situation is a little different, though. So I can handle it on my own :)
[10:12] <treitter> mdz: thanks
[10:15] <sladen> crimsun: I think the problem simple comes down to all the packages calling it with --new-option not depending on (>= 1:1.0.0-4) which is when the new option was introduced.  It's still a bit disgusting that X completely breaks from something so simple.  The lack of 'fixed' was the cause of something previously and having a rock-solid way of having at least one font available seems to be a bit like us providing busybox when your root partition can't 
[10:15] <libervisco> hey guys
[10:15] <sladen> hello there libervisco, is the weather nice where you are.  It's getting a bit dark here and earlier it was raining
[10:15] <libervisco> Where do you go if you badly need contributors for your software project?
[10:16] <libervisco> ehh yeah weather is ok.. kind of cloudy though :)
[10:16] <sladen> libervisco: who is "you" in this case, ubuntu, or yourself?
[10:16] <libervisco> any developer with a project in need
[10:17] <libervisco> I'm just trying to check if there is a site or service that enables devs to list their project as "in need of contributors"
[10:17] <libervisco> because I would like to open one
[10:17] <libervisco> on our site
[10:18] <libervisco> do you think it would be of help?
[10:24] <Surak> sladen: done. bug #55928 . I think I explained enough, but that's me :-) I can provide more information if needed
[10:24] <Ubugtu> Malone bug 55928 in xorg "xorg's configure script doesn't detect primary video adapter properly in multi-video environments." [Untriaged,Unconfirmed]  http://launchpad.net/bugs/55928
[10:27] <sladen> Surak: okay, many thanks for that.
[10:35] <Surak> regarding the broken bios, I've sent similar reports to intel and msi on several different modern motherboards that shows the issue. MSI is receptive to my reports lately.
[10:38] <sladen> Surak: remind me what the broken BIOS issue was?  ACPI DSDT?
[10:39] <sladen> excellent to know that MSI are listening
[10:39] <Surak> oh, it's on that bug. The fact that even when you disable integrated video, it still appears on pci device list
[11:04] <bluefoxicy> fabbione:  after consulting several network engineers it seems we've all concluded that attaching two routers to eachother through a switch housing another subnet will exponentially increase the load on the involved port on the default gateway as the two subnets grow; whereas connecting the two routers over a dedicated router-to-router link results in linear growth
[11:04] <bluefoxicy> s/exponentially/polynomially/g
[11:04] <bluefoxicy> so yeah, my point stands and has been validated:  that design is brain damage.
[11:20] <Riddell> jordi: ping
[11:51] <bluefoxicy> .... someone should figure out why Evolution needs 571 megs of memory (I just freed 400M of swap and 200M of RAM o_O) to manage around 400 messages, while Thunderbird (started at the same time) is using 185M to manage a few hundred thousand (other account is subscribed to about 30 mailing lists)
[11:53] <Burgwork> bluefoxicy, evo memory issues are being worked out
[11:53] <Surak> hum... how did you measure this?
[11:54] <bluefoxicy> Surak:  well, first off i looked at virtual and figured out how much it thinks is there; second, I closed evo and went fro 960M used (270 cached) and 860 swap to 660 used (330 cached) and 440 swap
[11:55] <bluefoxicy> so 690 + 860 -> 330 + 440 == 1550 -> 770, ... wtf.  That's more than 571 megs, how is that even possible.. oh.  I didn't count the evolution data server.
[11:55] <bluefoxicy> htf much was that reporting mapped into VMA, I didn't look :|
[11:55] <zyga> bluefoxicy: fragmentation probably
[11:56] <bluefoxicy> zyga:  nods
[11:56] <bluefoxicy> Surak:  point is I tend to open and close apps and see what the system thinks they're doing to memory.
[11:56] <Surak> hum, ok
[11:56] <bluefoxicy> I eventually got down to around 400 megs total
[11:57] <bluefoxicy> by closing Firefox as well (I'm running FF 2.0)
[11:57] <bluefoxicy> FF2 seems to be a RAM hog
[12:00] <bluefoxicy> another silly question would be
[12:00] <bluefoxicy> why is libc6-amd64 installed on i386
[12:01] <bluefoxicy>  /lib64/libc-2.4.so: cannot execute binary file
[12:06] <pitti> good night everyone