14:00:19 <licquia> #startmeeting LSB Bug Triage 2014 Apr 25
14:00:20 <lsbbot> Meeting started Fri Apr 25 14:00:19 2014 UTC.  The chair is licquia. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:20 <lsbbot> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:00:33 <licquia> good morning
14:01:58 * mwichmann waves
14:02:12 <mwichmann> and goes back to grumbling in his coffee about network test suites
14:02:27 <licquia> you've been having so much fun with those, haven't you? :-)
14:02:41 <mwichmann> it livens my day in between all the other silly stuff
14:03:05 <mwichmann> current amusement seems to be:
14:03:19 <mwichmann> if you do foo, it will fail, so do foo anyway but don't check the return
14:03:40 * mwichmann gets pedantic and checks the return, and gets the new-and-interesting fails
14:03:58 <licquia> #info lsb: 1 reopened, 2 pleasetest, 1 needinfo, 1 new
14:04:02 <mwichmann> which implies understanding "why"
14:04:08 <orc_fedo> sigh
14:04:25 <licquia> #info fhs: 27 new
14:04:48 <mwichmann> looks like another came up
14:04:54 * orc_fedo hides
14:05:05 <mwichmann> #info lsb: 1 reopened, 2 pleasetest, 1 needinfo, 2 new
14:05:09 <licquia> oops, saw that
14:05:16 * licquia didn't refresh in time
14:05:46 <mwichmann> new ones first?
14:05:54 <licquia> we can do that
14:06:05 <licquia> !lsbbug 3968
14:06:07 <lsbbot> licquia: 04Bug http://lsbbugs.linuxfoundation.org/show_bug.cgi?id=3968 normal, P2, ---, licquia, NEW , LSB4.1: Scribus fails because of lib incompatibility
14:06:19 * mwichmann hands the floor to orc_fedo
14:06:28 <mwichmann> careful - don't drop that
14:06:50 <licquia> bad things happen when the floor drops
14:07:08 <orc_fedo> terazzo floors, especially
14:07:45 <mwichmann> I think that we think that the tester is not doing things right
14:08:34 <licquia> another thought: they're linking directly against png at one version, and against cairo, which itself links against png at a different version
14:08:55 <mwichmann> well, "they"....
14:09:14 <mwichmann> the appbat app in question does link against both png and cairo
14:09:44 <licquia> right, that's what i meant
14:09:57 <licquia> anyway, wonder if the two libpngs are stepping on each other's toes somehow
14:10:21 <licquia> we've had that problem before
14:10:47 <mwichmann> can the dynamic linker handle this case?
14:11:03 <mwichmann> note the error message is about a png version that is not us...
14:11:10 <mwichmann> so there may be other things at play
14:11:22 <mwichmann> lsb png12 is 1.2.8
14:11:40 <licquia> i'm wondering if that error msg isn't coming from cairo
14:12:06 <licquia> in which case, it could indicate that cairo is trying to run with libpng12, which would be bad
14:12:31 <mwichmann> I have not installed lsb-scribus here though have a somewhat similar situation (cairo depends on png 1.5)
14:12:41 <orc_fedo> mwichmann: as to: can the dynamic linker handle  it seems one can 'blacklist paths on the fly' with a custom ldconfig --- I see the -C option to ldconfig and we could trim the cache in the testing
14:13:03 <orc_fedo> -C cache
14:13:03 <orc_fedo> Use cache instead of /etc/ld.so.cache.
14:13:04 <mwichmann> stewb dropped scribus from appbat in the 5.0 series, fwiw
14:13:13 <licquia> yup, remember that
14:13:20 <licquia> too many issues like this
14:13:59 <mwichmann> so I don't want to speculate on what causes his problem any more, just decide: how do we mark up the bug and give advice on how to proceed
14:14:19 <licquia> imho, one should decide to do one's image processing with cairo/qt, or with libpng/libjpeg/etc. directly, but not both
14:14:46 <licquia> sounds like we need to do our own testing on a recent fedora or something, and try to duplicate the bug
14:15:07 <mwichmann> makes sense
14:15:24 <licquia> as for markup: seems to be a 4.1 update issue
14:15:46 <mwichmann> +1
14:17:51 <licquia> #agreed target 3968 at stable updates
14:18:07 <licquia> #action test 3968 against recent distro with newer libpng
14:18:23 <licquia> !lsbbug 3969
14:18:25 <lsbbot> licquia: 04Bug http://lsbbugs.linuxfoundation.org/show_bug.cgi?id=3969 normal, P2, ---, licquia, NEW , lsb-dist-checker needs to enumerate needed locales
14:18:57 <licquia> i'm wondering if this isn't more about a problem report update
14:19:08 * licquia checks the prdb
14:19:52 <mwichmann> I thought dist-checker knew about such things?
14:20:27 <licquia> the specific error being reported looks more like a problem report thing
14:20:42 <licquia> post-test error msgs like that always come from the problem db
14:20:54 <orc_fedo> licquia: with the 'new' welcoming LSB, even if it is in the prdb, wouldn't we need to move to external visibility and resolution of this kind of thing
14:21:14 <licquia> yes, that's what i'm checking
14:21:30 <mwichmann> It might come from the test itself too
14:21:34 <licquia> it's not a question of whether it needs fixing or not; it's a question of where the fix needs to go
14:21:48 <mwichmann> remember we're using upstream's test for this, lightly wrapped
14:22:40 <licquia> https://www.linuxbase.org/prdb/?id=356
14:23:02 <mwichmann> okay, okay, prdb
14:23:17 <licquia> we probably need to update that link to point to the wiki page
14:23:48 <orc_fedo> in the short term add an out reference to that link as well
14:23:53 * licquia gets the new page from the other irc channel
14:24:02 <mwichmann> so the bug then is "prdb entry is stale"
14:24:03 <orc_fedo> it isn the URL field
14:24:08 <orc_fedo> is in
14:26:02 * mwichmann makes no promises the wiki page referenced is not also stale :(
14:26:51 <licquia> ok; think we should move it, too
14:27:16 <licquia> well, maybe; maybe we just need the "problemreport" keyword
14:27:35 <licquia> make sense?
14:27:54 * licquia doesn't think the prdb has its own bug section
14:29:05 <licquia> ok, going with that
14:29:05 <orc_fedo> mwichmann: wiki's rot, and that entry is toned in the first person, so it is clearly simply documentary and informational
14:29:23 <mwichmann> yeah
14:29:37 <licquia> #agreed 3969 is a problem_db issue, attach problemreport keyword and mark assigned
14:29:44 <mwichmann> a lot of stuff got documented there, most recently from stew's exploits setting things up right
14:29:54 <licquia> #action fix link for 3969 in problem db and check wiki page
14:29:55 <mwichmann> and a lot of that knowledge then went into dist-checker
14:30:23 <licquia> that, or we could embed the list into the problem db entry itself
14:30:54 <licquia> that way, we don't have to worry about bitrotting/spamming wikis, stale external links, etc.
14:31:37 <orc_fedo> ... or update the code to spit out a list of desired locales, when it feels a deficit ...
14:32:05 <licquia> that might be a good idea too
14:32:17 <mwichmann> dist-checker?
14:32:19 <licquia> the advantage with prdb: it's quick, and immediate
14:32:27 <orc_fedo> mwichmann: yes
14:32:31 <orc_fedo> thus my filing
14:32:34 <licquia> no software release needed
14:35:21 <mwichmann> and there we are in re: my comment yesterday, two "small" bugs and it's been 35 minutes... sigh...
14:36:08 <mwichmann> I'll get to withdraw my idea we'll have time for a bunch of extras
14:36:27 <licquia> ok, added comment to that effect, and setting target to vip updates
14:36:46 <licquia> #agreed 3969 should also be fixed in dist-checker; adding to vip rollup
14:37:43 <licquia> well, the rest should be fairly quick
14:38:02 <licquia> !lsbbug 3195
14:38:04 <lsbbot> licquia: 04Bug http://lsbbugs.linuxfoundation.org/show_bug.cgi?id=3195 normal, P2, ---, mats, REOPENED , LSB packages with unowned directories uninstallable on some rpm implementations.
14:38:17 <orc_fedo> oh yeah - that one
14:38:20 <licquia> no progress there; propose we leave it as it is and do the release already
14:38:28 <orc_fedo> unpushed updates as I recall?
14:38:37 <licquia> yup
14:38:51 <licquia> !lsbbug 3946
14:38:53 <lsbbot> licquia: 04Bug http://lsbbugs.linuxfoundation.org/show_bug.cgi?id=3946 normal, P2, 4.1-updates, licquia, PLEASETEST , Add scandir and alphasort interface with tests to olver-core
14:38:56 <licquia> !lsbbug 3947
14:38:58 <lsbbot> licquia: 04Bug http://lsbbugs.linuxfoundation.org/show_bug.cgi?id=3947 normal, P2, 4.1-updates, licquia, PLEASETEST , Extend mutex interface in OLVER to support prioceiling and protocol
14:39:06 <licquia> ran those tests yesterday, they can be closed
14:39:36 <mwichmann> yay!
14:39:44 <orc_fedo> ;)
14:40:22 <licquia> !lsbbug 3965
14:40:24 <lsbbot> licquia: 04Bug http://lsbbugs.linuxfoundation.org/show_bug.cgi?id=3965 normal, P2, ---, licquia, NEEDINFO , old lsb-runner packages in repositories
14:41:41 <licquia> so there's an outstanding issue, in that we could see this happening again w/o something going through and deleting old snapshot/staging build results every so often
14:42:01 * licquia proposes renaming, moving to infrastructure
14:42:11 <orc_fedo> licquia: makes sense +1
14:42:15 <mwichmann> +1
14:42:17 <licquia> ah, it's already on infrastructure
14:42:25 <orc_fedo> wiser than you knew
14:42:44 <licquia> #agreed do the extra task on 3965, mark assigned
14:43:20 <mwichmann> there's no bug on current s390 state, but looks like some more tweaking needed
14:43:36 <mwichmann> possibly more for puppet on setup?
14:44:30 <licquia> possibly; something odd is going on
14:44:50 <licquia> all the bits for 31-bit builds are there, but we're just not picking them up like we should be
14:44:59 <licquia> so something looks a little broken
14:45:04 <licquia> still trying to figure out what
14:45:17 <mwichmann> on 390x, tar (child): ../../saved/devel/tet-harness-s390x.tar.gz: Cannot open: No such file or directory
14:45:24 <mwichmann> dep-tet-harness step
14:45:48 <licquia> that, i think, might be a weird ordering thing; want to try that build again
14:46:05 <licquia> after making sure the tet-harness tarball is there
14:46:13 <mwichmann> sure
14:46:50 <mwichmann> libbat fail on s390 - it doesn't find the tarballs, not sure if it's set up to fetch them or not
14:47:13 <mwichmann> Checked entities from entities/package
14:47:13 <mwichmann> Checked packages from /opt/buildbot/lsb-slave/libbat-s390/appbat/packages
14:47:14 <mwichmann> Checked patches from patches
14:47:14 <mwichmann> Package entities found: 0
14:47:14 <mwichmann> Missing packages: 47
14:47:16 <mwichmann> and quits
14:47:20 <licquia> yeah, that's likely a holdover from the hercules build slave
14:47:51 <licquia> once builds start working there, i want to see them fail on real hw before trying to figure out what's going on
14:48:09 <mwichmann> ok
14:49:03 <licquia> and then we have that weird libstdc++ build fail, which i thought we had licked on the other archs
14:49:37 <mwichmann> fyi: just installed the 4.1 lsb-scribus on f19
14:49:44 <licquia> works?
14:49:53 <orc_fedo> licquia: but that looked like simply a missing BR dependency, not called out, and a possible artifact of build order?
14:49:55 <mwichmann> no complaints, gui is up
14:50:05 <licquia> hmm
14:50:19 <licquia> so it sounds very much like a distro issue
14:50:29 <mwichmann> but I'm not sure libpng1.5 is incompatible, while 1.6 is, so it may not prove enough
14:50:40 * licquia checks f20
14:50:45 <mwichmann> yes, sounds like distro issue
14:51:16 <licquia> f20 has libpng 1.6
14:51:23 <orc_fedo> licquia: as to tie IA64 lsb-scribus bug, it seemed the reporter might have been locally tampering w his environment as well 'to make stuff work'  he was present earlier and had linked in stray stuff in that regard
14:51:24 <licquia> so that should be the acid test
14:51:42 <licquia> that's what i'm wanting to try to eliminate
14:51:57 * mwichmann wonders idly if it really was ia64
14:52:04 <licquia> if we can reproduce it in our environment, we know something
14:52:10 <mwichmann> (not important to bug)
14:52:22 <licquia> if we can't, we can go back to him and say "try reinstalling from scratch and try again"
14:52:46 <mwichmann> clean environment definitely recommended
14:53:04 <licquia> anyway, want to clear away some fhs bugs?
14:53:19 <mwichmann> in eight minutes, why not
14:53:26 <licquia> !fhsbug 60
14:53:28 <lsbbot> licquia: 04Bug http://bugs.linuxfoundation.org/show_bug.cgi?id=60 normal, P2, ---, licquia, NEW , comments regarding /etc/[xml|sgml] & /usr/share/[xml|sgml] sections
14:54:17 <orc_fedo> only a decade old
14:54:22 <licquia> "new"
14:54:47 <licquia> wondering if the verbiage re: sgml and xml shouldn't just be dropped
14:55:32 <orc_fedo> it is noted as:  (optional)
14:55:35 <orc_fedo> at teh outlink
14:55:59 * licquia considers "optional" in a std to be a bit useless
14:56:24 <licquia> to the extent it's not, it means "you don't have to have /etc/xml, but if you do, it needs to work like this"
14:56:38 <licquia> and i'm not sure that claim is exactly true, given the other info in the bug
14:56:43 <orc_fedo> we could eradicate all such in lhs 3.0, and add a 'best practice annex to hold such cruft
14:57:05 <licquia> maybe a new bug for that?
14:57:08 <mwichmann> I think there's something like that already, haven't looked at the darn spec for a long time
14:57:13 <licquia> and then close this one as wontfix?
14:57:42 <orc_fedo> licquia: I am fine w a new rollup bug for planning fhs 3 and closing this as wontfix atm
14:58:05 <orc_fedo> (b/c of its optional natire
14:58:35 <mwichmann> appendix has contributors, maybe should update that too
14:58:51 <mwichmann> what *is* the timeline to wrap and relese this, is it just "together with LSB 5.0"?
14:59:17 <licquia> pretty much
15:00:14 * mwichmann looks at source, wonders what draft.sgml is?
15:00:55 <mwichmann> we do have    6. Operating System Specific Annex
15:00:55 <mwichmann> 6.1. Linux
15:01:39 <licquia> yeah, that may also get merged at some point
15:01:43 <mwichmann> which isn't quite "best practices"
15:01:50 <licquia> the idea was that fhs would apply to more than just linux
15:02:25 <mwichmann> yes... was just thinking of places to put commentary like discussed above
15:02:37 <orc_fedo> and intervening *nix politics have pretty well show that down -- the BSDs and OS/X just dont care to
15:02:40 <licquia> quick one more:
15:02:46 <licquia> !fhsbug 99
15:02:47 <orc_fedo> s/show/shot/
15:02:48 <lsbbot> licquia: 04Bug http://bugs.linuxfoundation.org/show_bug.cgi?id=99 major, P2, ---, licquia, NEW , Add /var/agentx to comply with RFC 2741
15:02:57 * licquia proposes wontfix here
15:03:09 <licquia> #agreed fhs bug 60 should be closed as wontfix
15:03:32 <orc_fedo> +1 ..
15:04:06 <mwichmann> 60?
15:04:18 <mwichmann> oh, earlier one
15:04:20 <mwichmann> ok
15:04:31 <orc_fedo> http://www.ietf.org/rfc/rfc2741.txt is ... a niche part
15:04:42 <licquia> yeah, not sure it's relevant anymore
15:04:43 <orc_fedo> seems out of scope
15:04:46 <mwichmann> seems like it was argued in bug
15:04:57 <mwichmann> the complaint is fhs as written disallows
15:05:07 <licquia> plus, there's nothing preventing a pkg from creating /var/agentx if they want it
15:05:26 <mwichmann> if they want to ignore the spec, yes
15:06:00 <mwichmann> essentially though: to me, this is not enough to create special-case wording out of, agree wontfix
15:06:05 <orc_fedo> or get a namespace in /opt and go wild ...
15:07:45 <licquia> #agreed close fhs bug 99 as wontfix
15:08:36 <licquia> alright, made a dent in those
15:08:41 <licquia> #endmeeting