broder | bdrung: what's the issue with debug-symbols-directly-in-usr-lib-debug? i'm having a hard time understanding it from reading the bug | 01:14 |
---|---|---|
broder | cjwatson: do you know what uid/gid was showing up in the broken builds? lintian doesn't seem to have an explicit tag for this, but the lintian lab has enough information for me to scrape out a report manually | 02:22 |
broder | also, does anybody remember when the buildd sudo issue was, even roughly? | 02:29 |
broder | wc -l ~/non-root-files -> 11534 | 02:45 |
broder | might require a little more pruning | 02:45 |
wgrant | broder: https://wiki.ubuntu.com/IncidentReports/2011-02-25-Permissions-build-failures? | 02:50 |
broder | wgrant: excellent. thanks | 02:53 |
broder | actually, bdrung's example of audacity was built 4 weeks ago, so it's clearly not fallout from the sudo incident | 02:57 |
wgrant | broder: Link? | 02:58 |
broder | wgrant: https://launchpad.net/ubuntu/+source/audacity/1.3.13-5 is triggering non-standard-file-perm from lintian - http://lintian.ubuntuwire.org/full/pkg-multimedia-maintainers@lists.alioth.debian.org.html#audacity | 03:00 |
broder | there are about 21000 warnings from cases where lintian expects the file permissions to be 0644 but they are actually 0664 | 03:01 |
broder | http://web.mit.edu/broder/Public/non-standard-file-perm is the full list | 03:01 |
broder | (that last link is just instances of the lintian warning that are not present in debian's lintian instance) | 03:02 |
broder | wgrant: looks like just about all of those broken uploads happened after precise opened (http://paste.ubuntu.com/736896/), though it's certainly not universal to all builds | 03:40 |
wgrant | broder: Which arches? | 03:42 |
broder | wgrant: i'm just looking at i386 | 03:42 |
wgrant | broder: Ah, so most of this are stuff that would have been done by pkgbinarymangler. | 03:44 |
wgrant | pngs/svgs mostly | 03:44 |
broder | oh, good point. i wonder if optipng started screwing up permissions | 03:45 |
micahg | umm, there was a bug last cycle about images having issues with permissions as well... | 03:45 |
wgrant | That's normally a bad umask issue. | 03:45 |
wgrant | So we may have some non-virt builders with the sudo issue still/again. | 03:45 |
wgrant | (LP people have very little visibility into our builders, unfortunately :/) | 03:45 |
broder | fyi http://paste.ubuntu.com/736898/ is just non-png and non-svg | 03:46 |
wgrant | In fact, they're almost all pngs and svgs. | 03:46 |
wgrant | Thanks. | 03:46 |
wgrant | Do those appear in Debian too? | 03:46 |
broder | no, all of these should be unique to us | 03:46 |
wgrant | k | 03:46 |
broder | (assuming i didn't screw up the queries) | 03:46 |
wgrant | Although I think some of these packages are Ubuntu-specific. | 03:46 |
broder | the brother-* ones were uploaded in february, so could be early uncaught fallout from the sudo thing | 03:46 |
wgrant | I thought we caught everything from that, but I was working on second-hand information on which buildds were affected and when. | 03:47 |
broder | right, the incident report says it was first caught 2/25, but the brother-* uploads were 2/20 | 03:47 |
wgrant | Hmm. | 03:49 |
wgrant | At least vernadsky and rothera have been bad in the last week. | 03:49 |
wgrant | But there are so few affected builds that it can't really be a general issue. | 03:49 |
broder | ok. you're on your own for cross-referencing builds. the ultimate debian database doesn't have that | 03:56 |
wgrant | I figured :) | 03:56 |
wgrant | Hmm | 03:57 |
wgrant | readline-common is interesting | 03:57 |
wgrant | debian/readline-common.overrides is 664 in both Ubuntu and Debian. | 03:57 |
broder | i...probably didn't filter out overridden tags | 03:57 |
broder | one moment... | 03:57 |
wgrant | No, it's correct. | 03:58 |
wgrant | In the Ubuntu binary it's 664 | 03:58 |
wgrant | But in Debian it's 644 | 03:58 |
broder | oh, i see | 03:58 |
wgrant | The source is 664 in both. | 03:58 |
broder | though i still think i included overridden tags, though i doubt we have any of those | 03:58 |
wgrant | Hmmm | 04:01 |
wgrant | cp -p debian/readline-common.overrides \ | 04:01 |
wgrant | $(d_comm)/usr/share/lintian/overrides/readline-common | 04:01 |
broder | hmm. do you think debian's binaryful upload of readline was for i386? | 04:04 |
wgrant | Hah | 04:04 |
wgrant | That is a good theory. | 04:04 |
broder | looking... | 04:04 |
wgrant | It was. | 04:04 |
wgrant | Architecture: source all i386 | 04:04 |
broder | awesome | 04:04 |
wgrant | Perhaps we should check powerpc or something instead. Or do you only have lintian results for i386? | 04:04 |
broder | i only have debian's results for i386 | 04:05 |
wgrant | Ah, but it's readline-common, so it didn't matter anyway. | 04:05 |
wgrant | In fact, an awful lot of them are arch-indep packages. | 04:06 |
wgrant | eg debdelta seems to be the same thing. | 04:07 |
wgrant | source.conf is 664 in the source | 04:07 |
wgrant | sources.conf, that is. | 04:07 |
broder | how is cp -p supposed to interact with umask? | 04:07 |
wgrant | Seems to ignore it. | 04:08 |
broder | huh - i'm surprised that we're still triggering issues on images. i see a patch in pkgbinarymangler to deal with that | 04:11 |
wgrant | Yeah, I'm just trying to track that down now. | 04:11 |
wgrant | It's odd. | 04:11 |
wgrant | There are certainly svgs from the last week. | 04:12 |
wgrant | But pkgbinarymangler doesn't do that. | 04:12 |
broder | and the pkgbinarymangler changelog says it was fixed in oneiric | 04:12 |
wgrant | Hmmm. | 04:12 |
wgrant | What is touching the svgs, if not pkgbinarymangler... | 04:12 |
broder | i thought pkgbinarymangler compressed both svgs and pngs | 04:12 |
wgrant | grepping for svg yields nothing. | 04:12 |
wgrant | But I thought the same. | 04:12 |
wgrant | HMm | 04:13 |
wgrant | But gbrainy is on the CDs. | 04:13 |
wgrant | Maybe it has a patch to do it. | 04:13 |
wgrant | Doesn't seem to. | 04:13 |
* broder goes and tries to hunt down old uds discussions | 04:13 | |
wgrant | Although it's cdbs. | 04:13 |
wgrant | So who knows what evil is going on there. | 04:14 |
broder | "scour" is the program you're looking for | 04:15 |
broder | ah - pitti created a dh_scour | 04:15 |
broder | https://blueprints.launchpad.net/ubuntu/+spec/performance-desktop-n-install-footprint | 04:15 |
broder | aha - and patched /usr/share/cdbs/1/rules/debhelper.mk | 04:16 |
broder | sneaky | 04:16 |
wgrant | Ahhh | 04:16 |
wgrant | That's why I couldn't find it. | 04:16 |
wgrant | Sneaky indeed :( | 04:16 |
broder | "if grep -q '^Component:[[:space:]]*main' /CurrentlyBuilding 2>/dev/null; then dh_scour -p$(cdbs_curpkg) $(DEB_DH_SCOUR_ARGS); fi" | 04:16 |
wgrant | scour respects umask | 04:20 |
broder | is that a good thing or a bad thing? | 04:20 |
wgrant | gbrainy built on zirconium, which is an ex-lpia builder so I don't trust it at all. | 04:20 |
wgrant | I'd prefer that it preserved the original mode, but I guess it's OK. | 04:20 |
wgrant | Hmmmmm. | 04:23 |
wgrant | Our sbuild has had umask(022) at the top since the incident. | 04:23 |
wgrant | So the sudo config shouldn't matter so much. | 04:23 |
wgrant | Oh | 04:23 |
wgrant | Except yes it does, because sbuild still calls sudo itself. | 04:24 |
wgrant | -rw-rw-r-- 1 wgrant wgrant 113 2011-11-13 15:23 ./deb-specific/pkginfo.inc.php | 04:24 |
wgrant | Looks like fusionforge may be another binary upload issue. | 04:25 |
* wgrant kicks Debian a bit. | 04:25 | |
broder | haha | 04:25 |
wgrant | zirconium and roseapple at least have had the scour issue in the last week. | 04:25 |
wgrant | I'll get their sudo checked out on Monday, hopefully. | 04:25 |
wgrant | See if we can blame that. | 04:26 |
wgrant | broder: Is it easy enough to produce http://web.mit.edu/broder/Public/non-standard-file-perm with dates as well? | 04:27 |
broder | wgrant: probably. you want package, filename, date? give me a few to remember how to write sql :) | 04:28 |
wgrant | broder: Version might be handy too, and is probably pretty trivial :) | 04:29 |
wgrant | But I don't know the UDD schema any more. | 04:30 |
broder | yeah, i think i can do this | 04:30 |
broder | wgrant: http://web.mit.edu/broder/Public/non-standard-file-perm-with-dates | 04:34 |
wgrant | broder: Thanks! | 04:35 |
broder | np | 04:35 |
wgrant | It looks like *all* the recent stuff is scour. | 04:36 |
wgrant | Except for two dirs in sip4, and the readline6 issue which probably affects Debian too. | 04:36 |
wgrant | However, what's our policy now on how builds have to cope with umasks? | 04:38 |
wgrant | Since we've changed the default... | 04:38 |
broder | i have no idea | 04:39 |
wgrant | broder: So, I'll get sudo configs checked out this week. I guess we also want a bug against scour (dh_scour seems like it really should preserve the mode), and probably against readline6 and fusionforge in Debian. | 04:46 |
wgrant | For now, at least. | 04:46 |
broder | that all sounds reasonable | 04:47 |
=== dendrobates is now known as dendro-afk | ||
ben_ | Does Ubuntu have a place like Debian's WNPP? | 06:00 |
=== ben_ is now known as ebnny | ||
broder | ben_: not generally. we have a needs-packaging tag for bugs for new packages, but since ubuntu doesn't have maintainers in the sense of debian, there's not as strong of a concept of orphaned, etc packages | 06:01 |
=== ebnny is now known as ben_ | ||
ben_ | broder: Ok, are there other places besides the WNPP site? | 06:01 |
broder | ben_: what do you mean? | 06:01 |
broder | what are you trying to figure out? | 06:01 |
alkisg | Hi, where can I find the final decision about a UDS session called "FreeRDP and Remmina to replace rdesktop, vinagre and tsclient"? | 09:50 |
alkisg | I'd like to know if remmina will replace vinagre, so that I change my app to take advantage of the default Precice packages | 09:51 |
iceroot | cjwatson: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/869502 can you have a look here? specially https://bugs.launchpad.net/ubuntu/+source/linux/+bug/869502/comments/94 the changelog says that you did the change | 10:13 |
ubottu | Launchpad bug 869502 in linux (Ubuntu) "Kernel-Panic with 3.0.0.12-generic on asus eee pcs and msi wind (both using rt2800 wifi chipset)" [High,Confirmed] | 10:13 |
=== yofel_ is now known as yofel | ||
cjwatson | iceroot: You are confused. The only change I made was to deliver the existing firmware to the installer. | 11:28 |
cjwatson | iceroot: I am not a kernel hacker and cannot help you. | 11:28 |
cjwatson | iceroot: You'll need to ask somebody who actually works on the kernel ... | 11:30 |
Fusionite | Hey | 11:47 |
iceroot | cjwatson: and where did the firmware came before it was inside "linux-firmware" | 11:52 |
iceroot | so that i know where i have to look | 11:53 |
cjwatson | iceroot: I have no idea | 11:56 |
cjwatson | sorry, I really don't have the ability or knowledge to help you here | 11:56 |
iceroot | cjwatson: ok no problem | 11:58 |
cjwatson | the change I made had no effect on installed systems | 11:59 |
iceroot | cjwatson: ok, i was just looking where rt2860.bin is coming from and apt-file told me "linux-firmware", because the bug we are facing comes from rt2860.bin and i saw a changelog from you about rt-cards, so i asked you | 12:01 |
cjwatson | right, I'm sorry that the changelog was confusing, but my change is not the change you were looking for | 12:03 |
iceroot | cjwatson: ok, thank you for the info | 12:03 |
=== Quintasan_ is now known as Quintasan | ||
=== abhinav_ is now known as abhinav- | ||
=== maxb is now known as Guest38799 | ||
=== JanC is now known as Guest33428 | ||
=== funkyHat_ is now known as funkyHat | ||
dr3mro | hello can any one tell me how to merge bazzar branch into trunk ... i own the branch but i am not a amember of trunk | 15:08 |
tumbleweed | dr3mro: that's pretty OT for this channel. But if you don't have permission to write to the trunk, then you can't do the merge. You could create a proposal on launchpad, though | 15:10 |
dr3mro | create a proposal is what i look for but how ? | 15:10 |
tumbleweed | push the branch to LP, and click on the button on it's page, labelled Propose for mering | 15:12 |
=== Mario_ is now known as pygi | ||
=== Guest38799 is now known as maxb | ||
broder | hmm. the rsync daemon on one of the archive.ubuntu.com servers (91.189.92.170) seems to be down. who do i contact about that? | 19:39 |
tumbleweed | broder: you should be using rsync.archive.ubuntu.com (although both hosts in archive.ubuntu.com also appear in rsync.a.u.c) | 19:45 |
tumbleweed | generally #ubuntu-mirrors for this kind of issue | 19:45 |
=== Mario_ is now known as pygi | ||
lifeless | ok, semi random question tinme | 20:02 |
lifeless | is launchpadlib included on the Ubuntu installer CD these days ? | 20:02 |
JackyAlcine | lol | 20:02 |
lifeless | broder: #canonical-sysadmins if I remember the channel name correctly | 20:03 |
lifeless | broder: or #ubuntu-mirrors | 20:05 |
infinity | lifeless: Which lplib bits are you looking for? | 20:09 |
infinity | lifeless: But if you check apt-cache, you'll note that liblaunchpad-integration-3.0-1 (for instance) is in every *-desktop seed. | 20:10 |
lifeless | infinity: python-launchpadlib | 20:10 |
infinity | adconrad@cthulhu:~$ apt-cache show python-launchpadlib | grep ^Task | 20:10 |
infinity | Task: cloud-image, ubuntu-desktop, server, ubuntu-usb, kubuntu-desktop, kubuntu-mobile-desktop, edubuntu-desktop, edubuntu-usb, xubuntu-desktop, mythbuntu-backend-master, mythbuntu-backend-master, mythbuntu-backend-slave, mythbuntu-desktop, mythbuntu-frontend, lubuntu-desktop, ubuntustudio-desktop | 20:10 |
lifeless | thanks | 20:11 |
=== jelmer_ is now known as jelmer | ||
broder | tumbleweed: hmm, ok. debmirror defaults to archive.ubuntu.com (i'm not actually specifying it explicitly). i'll file a bug | 21:51 |
tumbleweed | broder: is rsync the default method? | 21:58 |
broder | hmm, no it's not | 21:58 |
tumbleweed | although I know it does always rsync some extra bits, unless you tell it not to | 21:59 |
tumbleweed | but those are small enough, that the load shouldn't be problematic on a.u.c | 21:59 |
broder | what's the difference between r.a.u.c and a.u.c? it looks like all the same IPs | 22:02 |
tumbleweed | broder: r.a.u.c is a much bigger pool | 22:03 |
broder | really? the pool looks exactly the same from where i am | 22:04 |
tumbleweed | I only see two IPs in a.u.c | 22:05 |
broder | yeah, i used to, but i just started seeing 6 | 22:05 |
tumbleweed | but then US, is problematic, because you don't have ab ig countrywide mirror | 22:05 |
tumbleweed | ok, yes 7 here now | 22:05 |
broder | from here, us.a.u.c, rsync.a.u.c, and a.u.c are all the same set of IPs | 22:05 |
* broder shrugs | 22:06 | |
tumbleweed | broder: hrm, found the mail where rsync.releases.ubuntu.com was announced and we were asked to use it, but that was private, not on a list. Can't find anything mentioning rsync.archive.ubuntu.com. Maybe I just discovered it existed and started using it | 22:09 |
lool | Hmm I'm getting an upload error to a PPA | 22:26 |
lool | Uploading powertop-1.13_1.13-1ubuntu2_source.changes: 1k/2k550 Changes file must be signed with a valid GPG signature: Verification failed 3 times: ['General error', 'General error', 'General error'] : Permission denied. | 22:26 |
lool | gpg --verify works on the .changes file | 22:26 |
lool | and on the .dsc | 22:26 |
lool | ah intermittent Launchpad bug apparently | 22:28 |
micahg_ | yeah, that's an old bug | 22:29 |
=== apachelogger_ is now known as apachelogger | ||
=== Fusionite is now known as Ultimate | ||
=== Ultimate is now known as Ultimation | ||
=== Ultimation is now known as Fusionite | ||
=== Fusionite is now known as Backdoor | ||
=== Backdoor is now known as Fusionite | ||
=== Fusionite is now known as Jarubyh_The_Seco | ||
=== Jarubyh_The_Seco is now known as Jarubyh_The_2nd | ||
=== Jarubyh_The_2nd is now known as Virus | ||
=== asac is now known as asaqweqasd | ||
=== asaqweqasd is now known as asac | ||
=== Virus is now known as BiosDestroyer | ||
BiosDestroyer | Take care all | 23:42 |
=== BiosDestroyer is now known as Fusionite|AFK |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!