=== cjwatson1 is now known as cjwatson === asac_ is now known as asac [03:00] Man, Freenode been glitchy [03:01] You say that like it's news. [03:05] ScottK, I remember freenode when it was steady as a rock [03:06] NCommander: Do you have some time to try and verify a security patch for me (It's C, so I'm hopeless)? [03:06] NCommander: yeah yeah and you walked in 15 feet of snow uphill and downhill everyday to school in the next city.... [03:07] jdong: You got it wrong, "Up hill both ways". [03:07] jdong, no, I took a bus [03:07] ScottK, define verify [03:07] NCommander: verify (v. trans. InfoSec): "Yeah, I think most of ~/ is still there AFAICT..." [03:07] *ducks* [03:07] NCommander: Upstream produced a new upstream release that has both a working exploit test and a very invasive fix. [03:08] Ok [03:08] NCommander: Upstream gave a hint on how to shortcut avoiding the problem in a less invasive way [03:08] NCommander: Debian maintainer has produced a patch that he thinks is good, but upstream doesn't have time to deal with confirming it works. [03:09] I'm looking for some independent verification. [03:09] So, why not install package, run verification test? [03:09] You've looked at the code before so I thought of you. [03:09] I'd have to understand Perl to run the test. [03:10] That and I'm currently busy trying to understand why cmake seems broken in Jaunty. [03:18] I guess that's a no then. [03:18] Tries again. [03:18] I guess that's a no then. [03:20] ScottK, sorry, my laptop froze [03:20] ScottK, sure, I can be convinced to test it [03:20] OK. I'll email you some stuff. You'll want libspf2 from Jaunty and Hardy. What address. [03:22] NCommander: ^^ [03:23] sonicmctails@gmail.com [03:24] NCommander: Forwarded you the upstream hint and the proposed patch. [03:24] is the patch in jaunty or the new upstream? === deejoe is now known as spudnoggin [03:25] The patch is against 1.2.5 (in Dapper -> Hardy). [03:25] The new upstream is in Jaunty/Intrepid-security. === spudnoggin is now known as deejoe [03:25] The test is in the new upstream. [03:25] * NCommander grabs the jaunty source package [03:26] If the package is vulnerable, the test will cause a segfault. [03:26] NCommander: Be sure to grab the -security version. [03:26] ScottK, libspf2-2: Depends: glibc-private but it is not installable [03:26] That adds the patch that made it possible to get to this problem (fix one vulnerability and expose another latent one |o/) [03:26] *ahem* [03:27] ? [03:27] Trying to install on jaunty [03:27] Package is broken [03:28] Weird. Installs on Intrepid. [03:28] Same package. [03:28] :-P [03:28] Yeah [03:28] its broken on jaunty [03:28] OK. Separate issue. I'll look at that. [03:29] ok, I figured out how to run the test suite against various versions of the package :-) [03:29] Great. [03:29] I'll test the jaunty package (assuming it compiles) [03:29] Try the hardy-security version against it and see. [03:30] I need to have a known working negetive [03:30] Right. [03:30] argh [03:30] damn it [03:30] I actually have to install the library [03:30] * NCommander breaks pbuilder out [03:31] argh [03:31] the test suite needs a non-packaged perl module it seems [03:34] Upstream runs Gentoo, so who knows. [03:34] * NCommander whacks head [03:34] bah [03:34] Getting this test suite to run is a **** [03:35] * NCommander upgrades perl base [03:35] ok [03:35] confirmed that the version in jaunty does not suffer the buffer overflow [03:38] argh [03:40] t/11_overflows.......dubious [03:40] Test returned status 0 (wstat 11, 0xb) [03:40] DIED. FAILED tests 3-5 [03:40] Failed 3/5 tests, 40.00% okay [03:41] ScottK, ^ [03:41] (that being said, the test suite might be a false postive, I had to kick it to actually compile ) [03:41] That's Jaunty? [03:41] Hardy security [03:42] OK. I'm fairly certain that one isn't vulnerable. More than one person who knows the code verified that. [03:42] Ah. Hardy Security. Sorry [03:42] yeah [03:42] it segfaulted [03:43] I'd expect Hardy Security to fail. [03:43] That's the indication it's vulnerable. [03:43] it got through 2/3 tests however [03:44] The patch in Hardy security I think is tested for in that same test. [03:45] root@blacksteel:~/libspf2-1.2.5.dfsg/perl# perl t/11_overflows.t [03:45] 1..5 [03:45] ok 1 - use Mail::SPF_XS; [03:45] ok 2 - parse_cidr did not run off start of data [03:45] Segmentation fault [03:46] I think that's right. I suspect the one in hardy-release would fail 2 [03:48] so now what? [03:48] libspf has two rdepends [03:48] It might just be worth doing a backport [03:50] So the question is does the patch I emailed you fix it? [03:50] oh [03:50] I thought the patch was applied already [03:52] No. [03:52] I emailed you patch #2 that hopefully finishes fixing it. [03:52] Lets find out [03:53] t/11_overflows.......ok [03:53] WOOOO! [03:53] NCommander, chillax :P [03:53] chillax? [03:55] NCommander: So it works. Would you view it as sane? [03:56] Yes, expect it makes a blank patch file as well [03:56] 51_actually-keep-track-of-max_var_len.dpatch - empty [03:56] 52 - has the real patch [03:57] expect/except? [03:57] *except :-P [03:58] OK. [03:58] NCommander: Thanks. I'll work on -security debdiffs then. [03:58] \o/ [03:58] NCommander: Any thoughts on where glibc-private in Jaunty comes from? [03:59] no idea [03:59] it might just need a binary rebuild [03:59] Tried that. No luck. [03:59] It picks that up from somewhere. [04:01] apt-cache doesn't know what it is [04:01] It doesn't exist. [04:01] its in none of the shlibs files I have [04:03] * NCommander pokes LP [04:18] ScottK, I'm semi-stumped [04:18] NCommander: I dunno the relevancy, but glibc in Intrepid defines a large number of symbols in GLIBC_PRIVATE that seem to be gone in Jaunty. [04:18] bug in glibc? [04:19] * NCommander debates downgrading to intrepid [04:20] NCommander: I was thinking more planned change we need to figure out how to work with. [04:20] ?? [04:21] Looking at debian/changelog this symbol thing was on purpose. [04:21] Not sure if it's relevant to the problem at hand or not. [04:21] where is doko when you need him === jamesh_ is now known as jamesh [05:19] hello [05:38] NCommander: Thanks. -security debdiffs are done. [05:38] yay [05:38] * ScottK goes and sleeps. [06:23] <\sh> moins [06:24] <\sh> NCommander: welcome on board of the MOTU ship :) [06:24] \sh, thanks :-) [06:24] \sh, do you know who the resident grub guy is on Ubuntu? [06:25] <\sh> NCommander: nope...sorry... [06:57] cjwatson, I'd like to talk to you on adding splash images for grub [07:05] NCommander: Not by default, I hope. [07:06] wgrant, ? [07:07] NCommander: It was, I believe, decided that yet another mode change wasn't necessary. [07:07] Good morning [07:07] wgrant, mode change? [07:07] * NCommander notes our current grub does properly support splashimages, and those work just fine [07:08] NCommander: Splash image == graphical mode == mode switch == flickering == baaad [07:08] Morning pitti. [07:08] fair enough [07:08] morning Pici [07:08] * NCommander hit head [07:08] pitti, [07:08] Flickering and slowness and other stuff, probably. [07:08] wgrant: pong [07:08] It's not accidental that we lack a splash image, at any rate. [07:08] pitti: See bug #293318/ [07:08] Launchpad bug 293318 in gnome-settings-daemon "gnome-settings-daemon leaks memory" [High,In progress] https://launchpad.net/bugs/293318 [07:08] I think it's that one. [07:08] Yes it is. [07:09] fedora removed their grub splash image [07:09] They did? [07:09] for the same reason [07:09] * NCommander notes Debian still has one [07:09] wgrant: I got the bug mail, yes [07:09] wgrant, I can find that leak [07:09] NCommander: Huh? [07:09] now they don't even show the menu anymore, unless there's another boot option available (like windows) [07:09] the memory leak [07:09] tjaalton: That's what we do, isn't it? [07:09] wgrant: yes [07:09] NCommander: I've even fixed it upstream... [07:09] oh [07:10] nm [07:10] * NCommander hadn't finished reading the bug :-) [07:11] NCommander: Why do you want a GRUB splash? [07:11] Brainstorm #21, and my mom freaked when she saw it when I installed Ubuntu on her PC [07:12] Surely it's no more scary than watching a POST. [07:12] At least no computer I've owned in a few years shows a text based PoST [07:12] ... PSOT [07:12] ... POST! [07:12] on mine, I get Sony VIAO -> grub -> usplash [07:12] just get rid of windows and you're saved [07:13] * NCommander gets the grub countdown [07:13] no grub menu anymore [07:13] yes, I know [07:13] On my eee I get grub (very short) -> usplash [07:13] \o/ bios quickboot [07:14] NCommander: oh, yet another "idea" I can close?-) [07:15] tjaalton, if you must ;-) [07:19] ah, fedora didn't actually remove the splash, but it's not shown unless the menu is [07:21] * NCommander notes thats how itlled is on Ubuntu as well if a splash is insta [07:22] wgrant: did you sent the two g-s-d input patches upstream? There aren't any patch headers or gnome bz refs [07:22] pitti: The upstream-relevant parts are upstream. [07:22] wgrant: ah, nice; thanks [07:22] I didn't add patch headers because the patches were already there. [07:24] wgrant: I added the bug ref for 293318 to the changelog; does 08_extra_touchpad_options.patch correspond to a bug report, too? [07:25] pitti: They're both for that bug, though 08_extra_touchpad_options.patch is more of an additional defense. [07:25] ok, thanks! [07:26] No, thank you. [07:26] tjaalton: Are any other distros going to be sharing our pain in the immediate future? [07:26] fedora isn't yet? [07:27] wgrant: I assume "pain" == "input hotplug"? :-) [07:27] pitti: Input hotplug for everything, yes. [07:30] good morning [07:31] wgrant: fedora is [07:31] at least [07:31] and when lenny is released I bet debian will follow [07:31] tjaalton: Hmm, I wonder how they dealt with g-s-d... [07:32] From what I could see they hadn't. [07:32] wgrant: nothing on http://daniel.holba.ch/harvest/handler.py?pkg=gnome-settings-daemon ? [07:33] wgrant: what in particular? [07:34] pitti: No... [07:34] tjaalton: Mouse settings (most annoyingly right/left-handedness) don't persist when you hotplug devices. [07:34] Some keyboard stuff doesn't work at all, but that's another less important story. [07:36] wgrant: okei [07:36] uh [07:36] "okay" [07:44] slangasek: did you see the response in bug 236830? seems there is still a bug there? [07:44] Launchpad bug 236830 in samba "cifs does not support kerberos authentication" [Medium,Fix committed] https://launchpad.net/bugs/236830 === anil1 is now known as anilg [08:10] 谁知道有64的JAVA 那里下载 [08:11] Indeed. [08:12] where [08:13] I agree with RAOF. [08:14] Aha. Sorry. [08:14] Google tells me that you want to download Java. [08:15] SYDN: You want the "sun-java6-jre" package. [08:16] oh [08:17] There is a #ubuntu-cn channel though. [08:17] * iulian wonders why he didn't join there. [08:42] mvo: good morning [08:42] mvo: I checked the hardy SRUs, and it seems that bug 220890 is the only one which holds up -updates propagation of 3 hardy-proposed uploads [08:42] Launchpad bug 220890 in python-apt "[hardy] software-properties-gtk doesn't recognize (nor know about) ports.ubuntu.com" [High,In progress] https://launchpad.net/bugs/220890 [08:42] mvo: anything we can do about it? or is it not a regression, and we can copy and keep the bug open? [08:45] pitti: I look into it now, sorry that it got put on hold during intrepid [08:47] Ugh, that one [08:47] * NCommander puts that on the Ports todo tracker [08:50] superm1: ok, changed in bzr, thanks [08:51] pitti: 236830> the only issue I see at the end of the bug log was user error, which EtienneG addressed? [08:52] slangasek: ahh, I read it as "that's what the package should ship" [08:52] NCommander: what is the url for this? [08:52] slangasek: ok, so that's basically "didn't get any feedback" then [08:52] mvo, right now, I'm just going to note it on the kernelportspage [08:52] * NCommander kicks the spacebar [08:53] pitti: ah, so it is; I guess we should kick people for testing :/ [08:53] But at least w.r.t. to powerpc, I've been working w/ a bunch of other people to make sure PowerPC is a top-notch port again [08:55] good morning [08:57] mvo, I think I can probably fix this bug in Jaunty [08:57] (the ports one) [08:59] NCommander: I check it out now and let you know what I find, I thought I fixed it already :/ [09:00] I thought it was broken on Intrepid [09:00] mvo, are you a ports user? [09:00] no, that i part of the problem [09:00] install lpia on something w/ a CD then [09:01] Its considered a port on LP [09:02] right, good idea [09:02] I have those occasionally :-) [09:03] mvo: I'm ready to accept update-manager for intrepid-proposed, unless you have some more changes stacked up? [09:05] pitti: is the other one in -updates now? [09:05] mvo: yes, I copied it this morning [09:06] excellent! [09:06] nothing pending currently [09:06] mvo: good; accepted then [09:07] mvo: pitti: I have a debdiff coming shortly for update-notifier for Jaunty; adds an update-motd hook such that update notifications are appended to the MOTD [09:07] pitti, what is binary NEWed? [09:09] kirkland: nice, I will merge it when its ready [09:11] Is anyone else unfortunate enough to have to deal with the wl wifi driver? [09:12] soren, yes, I have [09:12] I find there's a huuuuge difference between the latency of my network connections when I use that and when I use ndiswrapper (which used to be my only option). [09:13] pitti, https://bugs.edge.launchpad.net/ubuntu/intrepid/+source/mtd-utils/+bug/294428 [09:13] * soren runs a simple benchmark [09:13] Launchpad bug 294428 in mtd-utils "mtd-utils build fail" [Undecided,Fix committed] [09:14] NCommander: Is it horribly laggy for you as well? [09:19] NCommander: a package which hasn't been built before must be checked by the archive admins, and is held in a queue for that before being published to the archive [09:19] soren, my current machine doesn't have wl, but in my experience, yes [09:20] pitti, ah, I see. I enabled proposed on my intrepid box, and checked it, so now what :-)? [09:21] NCommander: if it works, then it's good for copying to intrepid-updates, and closing the bug [09:21] pitti: Hi. Please give-back vim on i386. Thanks. [09:21] It works to the extent I can test it (I don't have anythign attached to this PC I can forgot jffs2, but the tools, and the help for the tools work) [09:22] NCommander: http://people.ubuntu.com/~soren/wlstats.txt <---- No fun at all [09:22] geser: it's built already [09:23] pitti: https://launchpad.net/ubuntu/jaunty/+source/vim/+builds shows a FTBFS for i386 [09:24] geser: indeed, seems my script is broken [09:25] * pitti fixes it for s/intrepid/jaunty/ [09:25] given back [09:25] * soren wonders why he didn't see the e-mail about that build failure... :/ [09:36] How does one get posting privileges revoked from -devel-discuss? [09:37] mail -devel-discuss-owner, I guess? [09:37] wgrant, is there a particular reason you ask? :P [09:38] They're not doing a very good job of making developers want to read it. [09:42] that's why it's called -discuss :) [09:42] for actual development there is ubuntu-devel [09:43] I mean, I thought I directed some serious torrents of hate at Launchpad, but that just doesn't compare... [10:02] wgrant: hmm, I wonder who has powers over that list [10:03] oh, wow. It's getting even better. [10:04] What list? [10:04] -devel-discuss [10:05] NCommander: https://lists.ubuntu.com/archives/ubuntu-devel-discuss/2008-November/006221.html and that general thread (and other useless wastes of bytes on that list) [10:05] NCommander: oh, and those sorts of posts are normal for that guy, too. [10:05] WTF is the difference between devel-discuss and -devel [10:05] One has ******** on it. [10:05] NCommander: devel == sane. has developers on it, and is moderated so you don't get that kind of rubbish. [10:05] NCommander, one is discussion about developement, one is discussion about discussion about development? [10:06] Hobbsee, semi-moderated, developers are whitelisted [10:06] NCommander: -devel-discuss: users speaking to developers...Or, now, as the case may be, lots of hate mail, which most people ignore, and occasional developer discussions, which a few people follow. [10:06] who do I get in contact if I want a mailing list [10:06] NCommander: well, yeah. But the developers aren't expected to behave badly. [10:06] NCommander: depends. What for? [10:06] * NCommander would like an ubuntu-ports list [10:06] NCommander: do you want to use listadmin on it? [10:06] listadmin? [10:07] * NCommander knows listsrv, but not listadmin [10:07] NCommander: http://debaday.debian.net/2007/09/12/listadmin-command-line-mailman-moderator-queue-manipulation/ [10:07] Probably [10:07] * NCommander would whitelist developers of course [10:07] NCommander: if you do, you can't use the launchpad lists, you'll need to get a list on lists.ubuntu.com (and I think you'll need to speak to jcastro about it) [10:08] jcastro, ^ [10:08] what program is this volume OSD of ubuntu part of? [10:08] it's probably night for jcastro [10:09] Heller_Barde: gnome-settings-daemon, I believe. [10:09] ah. and where can i find out how the shortcut is bound to it? [10:10] Heller_Barde: What do you mean? [10:10] or does gnome-settings-daemon capture shortcuts itself? [10:10] It does. [10:10] ah [10:10] i don't use gnome and actually dont want to but nothing else is able to look appropriately aesthetic :( [10:11] oh well [10:12] bah the two-stage rsync mirroring script from ubuntu is useless on slow internet connections [10:13] rsync the debs first, then the indexes. sounds like the perfect idea. but then it takes so goddamn long to rsync the debs that the indexes have updated by then =.= [10:13] hyperair: You're not meant to mirror the archive on dialup. [10:13] wgrant: i'm not on dialup damnit [10:14] wgrant: this is a university connection [10:18] pitti: I looked into #220890 and AFAICT the remaining issue is that source code is not recognized as "offical" on powerpc but binary deb lines are. so it should be ok for -update (and I will push another one afterwards that fixes the source lines) [10:31] mvo: ok, thank you [10:32] pitti: is your list for coping stuff from -proposed to -updated already empty (for intrepid)? [10:32] sebner: yes; see http://people.ubuntu.com/~ubuntu-archive/pending-sru.html [10:33] pitti: intrepid has 2 packages with 7 days ;D [10:33] sebner: but not verified one [10:33] s [10:34] pitti: what has to happen to make it verified? [10:34] sebner: test that the package in -proposed still works as usual, and that it fixes the bug, and give written confirmation of that in the bug [10:35] pitti: my package (audacious) has over 5 users can confirm that the update in -proposed fixes the issue ;) [10:35] sebner: ah, then nobody from motu-sru set the verification-done tag apparently; please do so then [10:37] pitti: done [10:37] thanks [10:37] pitti: np, thank you your copying then :D [10:40] pitti, hi [10:41] what is missing to get the CUPS SRU released? [10:41] tkamppeter: testing and being 7 days in -proposed [10:49] pitti, did not know that there is a "minimum 7 days" condition. All bugs related to pstopdf are answered and the answers suggest that the problems which I have addressed with pstopdf are really solved. [10:49] tkamppeter: right, except for the 'prints garbage on samsung' for splix [10:49] but we can keep that open and still move to -updates [10:50] * ogra grumbles at asac ... my mobile phone is suddenly not detected as USB HDD anymore ... grrr [10:51] * mvo thinks that asac probably sneaked int ogra house just to play that trick on him [10:51] * ogra tries to get the pictures off the camera with bluetooth instead [10:52] mvo, he doesnt go that far south for playing tricks i bet :) [10:52] ogra, perhaps he has agents everywhere? [10:52] haha [10:52] he is, after all, a manager ;) [10:53] ogra: you never know, he is on vac, maybe that is part of his vacationfun ;) [10:53] * mvo stops being silly [10:53] pitti, and the AppArmor thingy with the network protocols I can confirm. The problem appears (I have 6 network printers) on the machine where I did not update to -proposed and does not appear with the -proposed package. [10:54] tkamppeter: can you please say so in the bug, for having a written track of testing? [10:56] pitti, done. [11:02] pitti, the garbarge printing which you observer with SpliX now is most probably a problem of SpliX itself. Therefore I need you to do tests with the LSB packages from OpenPrinting. The test on 2.0.0rc2 can reveal whether there is perhaps a bug in the Ubuntu packaging. The test of the 1.1.1 version can show whether the 2.x generation of SpliX introduced a new bug. [11:02] tkamppeter: ah, ok; I'll do that [11:03] pitti, I have modified the OpenPrinting packaging last week so that these drivers can coexist with the ones coming from the distros. [11:03] So you do not need to uninstall any package. [11:20] mvo: Does do-release-upgrade do the xorg.conf inputdevice mangling too? [11:21] wgrant, what mangling ? [11:22] ogra: We comment out inputdevices on upgrades to let input-hotplug work its magic. [11:22] xorg will ignore keyboard and mose settings from the file if not forced [11:22] No it won't... [11:22] thats inside xorg afaik [11:22] Really new xservers will ignore kbd and mouse entries, as evdev needs to grab them. [11:23] there was a setting i dont remember you have to add to the xorg.conf to make it use these entries [11:23] AutoAddDevices, yes. [11:23] right [11:23] But that's not what I'm talking about here. [11:23] else it will use evdev by default [11:23] and ignore the xorg.conf entries for keyboard and mouse [11:24] I know, but this is somewhat separate. [11:24] for other input devices it depends on the driever ... [11:25] i.e. wacom will only use hal if yu copy the .fdi in place, else xorg.conf should be used [11:25] And lots of them will break horribly, which is why we now comment them out on upgrades. [11:25] really? [11:25] * ogra wasnt aware we touch xorg.conf at all [11:25] As of late Intrepid we do. [11:26] hmm [11:31] pitti, I have investigated bug 293883 now and it is an evince bug, so it does not block the SRU release of CUPS. pstopdf was not the culprit there. [11:31] Launchpad bug 293883 in evince "8.10 Printed PDF missing parts / corrupt" [Medium,Confirmed] https://launchpad.net/bugs/293883 [11:38] NCommander: we tried it before and it broke some people's hardware, so I'm not all that interested in trying it again TBH [11:39] Works for me [11:39] * NCommander deletes the idea from his brain [11:39] rm -rf /sys/brain/idea [11:39] wgrant: yes [11:39] mok0, file not found [11:39] mvo: Hmm, OK, I have a report from somebody that it didn't. [11:40] hehe [11:40] But perhaps not fully trustworthy. [11:40] wgrant: it writes out a log what its doing [11:40] wgrant: please ask the reporter about the stuff in /var/log/dist-uprade/* [11:40] wgrant: is there a LP bug for it? [11:40] mvo: There isn't. I'll ask for logs. [11:40] thanks [11:42] late intrepid touches xorg.conf? [11:42] hyperair: Since late in the Intrepid release cycle, the dist-upgrader touches xorg.conf. [11:42] oh [11:42] what does it do [11:43] Comments out inputdevices. [12:00] slangasek: I ported your hal patch 99_new-kernel-rfkill-interface.patch as best as I could to new hal upstream; but since I don't get any observable lshal change when flipping the killswitch either with your old nor with my ported patch, I can't really test it; can you? [12:04] pitti: I guess that's for jaunty? I'm afraid I can't test that very easily; but the main point would be whether NM detects the status change, I guess? [12:05] slangasek: yes, for jaunty [12:05] slangasek: well, NM doesn't do anything if I change the killswitch [12:05] hmm [12:05] not with either version? [12:05] but didn't do that in intrepid either [12:05] that's why I can't test the ported patch [12:05] hrm [12:05] slangasek: Bit early, isn't it? [12:05] pitti: this is on a system with Intel wireless? [12:05] StevenK: ? [12:06] slangasek: It's 4am in LA [12:06] slangasek: yes, iwl3945 with dell killswitch [12:06] pitti: do you have /sys/class/net/$dev/device/rfkill? [12:06] StevenK: yep, right in the middle of my Monday shift [12:07] pitti: (you're running the Intrepid kernel, right?) [12:08] slangasek: intrepid kernel> yes [12:08] slangasek: /sys/class/net/wlan0/device/rfkill/rfkill0/state [12:09] slangasek: I can put the new consolekit and hal into my intrepid PPA, if that would help you [12:09] pitti: that would be better [12:09] I'm still puzzled that it doesn't work for you, though [12:09]