=== Guest30432 is now known as JackyAlcine [02:15] * gilbert is trying unity on 11.10, and isn't going completely mad [02:16] progress, ooh [02:17] but why oh why oh why can't i modify the panel??? come on... [02:20] wrong chan :( oops [02:20] disregard [08:01] good morning === almaisan-away is now known as al-maisan [10:01] anyone have the URL handy that searches UDD for uploads by an individual? [10:21] ooh, I like this nick! === Guest51058 is now known as Laney [10:21] boring :( [10:22] Laney: hi, do you have that link handy where I can look up my uploads in UDD? [10:22] ooh, err, ooh [10:23] the sponsoring thing? tumbleweed did that [10:23] yeah, ok, I wasn't sure who did that, let me see if I can find it now then [10:24] http://ubuntu-dev.alioth.debian.org/cgi-bin/ubuntu-sponsorships.cgi [10:24] micahg: that? [10:24] yes :), thanks [10:24] thank god for browser history [10:25] yeah, was lacking in mine and I thought I bookmarked it [10:25] * micahg needs to find who to hassle for endorsement [10:25] *endorsements [10:25] s/hassle/ask nicely/ :D [10:26] heh [10:54] tumbleweed: http://ubuntu-dev.alioth.debian.org/cgi-bin/ubuntu-sponsorships.cgi seems to not show sponsors anymore [11:01] micahg: it looks like the sponsors import is broken [11:01] the ubuntu_upload_history table seems to just have "N/A" for signed_by_name [11:03] http://pastebin.ubuntu.com/752368/ [11:06] BLAST! [11:10] erm [11:10] https://launchpadlibrarian.net/86055975/facter_1.6.2-1ubuntu1_source.changes [11:10] no signed-by? [11:11] oh, right, sanity checking? who does that? :) [11:11] epic fail [11:11] https://lists.ubuntu.com/archives/precise-changes/2011-November/003573.html [11:11] it is in the first part but not the changesfile [11:12] well it's in the gpg signature in the changesfile [11:12] which you don't have in changesFileUrl() [11:13] woe is us [11:13] oh, that's problematic [11:13] must be in the spph, though [11:14] the worst part is [11:14] http://anonscm.debian.org/gitweb/?p=users/laney/ubuntu-upload-history.git;a=blob;f=munge_ddc.py;h=11b3d28818110d84f23859de0b3e64fae80fcfed;hb=HEAD#l130 [11:15] the "some reason" is quite logical... [11:15] :P [11:16] i meant that i knew about it [11:16] and then promptly forgot said knowledge [11:16] right, so did I :/ [11:17] reimport time! [11:17] you think I can do it in one go? [11:18] launchpad downtime might get you [11:18] yeah :/ [11:19] I'll do per-release again [11:19] we should sanity check the dates [11:20] there was only ~10 uploads with borked ones so it's not too bad to ignore them [11:20] can we not trust date_created? [11:20] i guess fall back to that [11:20] or date_published even [11:21] hard to sanity check, though [11:23] parsedate_tz doesn't :( [11:24] if you want proper TZ handling in python you want dateutil and/or pytz [11:29] also doesn't validate it [11:29] i think the errors come from postgres itself [11:31] oh, I didn't see the errors you were getting [11:32] they were when importing [11:33] yes, what sort of problems? [11:33] #ubuntu-motu.log-20111124:23/11 00:26:02 psycopg2.DataError: time zone displacement out of range: "Thu, 12 Oct 2006 22:43:00 +2000" [11:33] like that [11:35] oh, /me remembers === yofel_ is now known as yofel [14:53] Laney: are you doing anything about ubuntu-upload-history (I'm about to start looking at it) [14:53] not atm, feel free === al-maisan is now known as almaisan-away [15:42] Laney: how about this: If abs(Changes Date - Date Created) > 24 hours, use Date Created ? [15:42] are there legitimate situations when that might be true? [15:43] stuck in a queue? [15:43] the +2000 case might not even be caught by that [15:43] embargoed uploads? [15:43] that's catchable by hand [15:43] you could just try to parse it [15:44] some of them were e.g. typoed days and months [15:44] which should hopefully fail to parse [15:44] but there were also things that were clearly wrong [15:44] and then check the tz is within a valid range [15:44] i.e. out by months [15:44] then making your check much longer (say 2 weeks) should be a good compromise [15:45] ok, I like that [15:45] I bet if the parser is lazy that it would accept invalid days though [15:45] as those are technically redundant [15:45] render it back to a string and compare? [15:46] sounds good [15:46] hrm, we don't have dateutil on syklone [15:46] * tumbleweed tries to work out if we need it [15:53] Laney: hrm, dateutil does sanity-check the TZ, but allows a greater range [15:53] -1439 to 1439 [15:54] (minutes) [15:58] another alternative is to let the udd importer skip invalid records [15:59] it's fine, I can easily enforce a tighter restriction here === nenolod_ is now known as nenolod [16:23] Laney: ok, running a smallish import of karmic from the beginning of 2009-10 (that'll also cover some odd e-mail addresses) http://people.ubuntuwire.org/~stefanor/out/ [16:24] cool [16:24] what did you do about the emails? [16:25] http://anonscm.debian.org/gitweb/?p=users/stefanor/ubuntu-upload-history.git;a=commitdiff;h=ab8f776c22ac367c28caa6fd2405d22d83557146 [16:25] WARNING:lp-udd:Date does not render back to itself: Mon, 12 Oct 2009 15:47:58 +0100 != Mon, 12 Oct 2009 15:47:58 +0100. Going with date_created [16:25] can we handle that? [16:25] I'm putting Signed-By on every record, seem sensible? [16:26] yes, we could special-case that [16:26] I guess signed_by <> 'N/A' was never good enough on its own, so I guess so [16:31] re-importing with double spaces special-cased [17:25] cjwatson, hey... are you working on flphoto? [17:28] lenios_: yes [17:28] er [17:28] l3on: yes [17:28] ah ok :) I see it fails :P [17:28] ok, don't worry.... let's see something else... :) [17:28] yes, I know, I'm working through the failures [17:29] flphoto for example... [17:29] ops sorry -.- [17:29] the failures in flphoto I mean [17:29] just a bit tedious [17:29] can I take git-annex ? [17:31] it's possible git-annex might succeed if retried [17:31] run it through pbuilder or something to chekc? [17:31] *check [17:31] it might have got caught up in the perl and imagemagick transitions [17:32] ok... let's see :) [17:32] still fails [17:34] sid version does too [17:37] it looks like a bug in whatever library that is it's using though? [17:38] i would say so [17:40] flphoto fix uploaded [17:48] Yep.. I'm going to test "perl -e 'use Image::Magick'" in precise [17:48] because in oneiric builds fine [17:50] Given all the transitions, problems in this area aren't especially surprising [17:59] that file hasn't been linked properly [18:09] hey men... I found the problem about myodbc ... [18:09] in precise we have libmysqlclient-dev > 5.1 [18:10] and myodbc seems to build only with libmysqlclient-dev <= 5.1 [18:10] in other words, in precise control should set "libmysqlclient16-dev" instead of "libmysqlclient-dev" [18:10] what the best way to do that? [18:11] and... debian sees libmysqlclient-dev > 5.1 in experimental... we have to file bug in debian or just merge in ubuntu for now ? [18:12] l3on: no, libmysqlclient16-dev is going away, we need it to build with libmysqlclient-dev (or 5.5) [18:12] micahg, ok... I'm going to figure out ... [18:13] libmysqliclient-dev should be libmysqlclient18-dev ATM [18:13] l3on: thanks :) [18:13] yes, you're right [18:36] hey micahg ... why don't we use myodbc 5.1.9 ? [18:37] I'm looking at code and bug has fixed there... [18:38] l3on: Debian hasn't updated yet? you can certainly propose a merge/debdiff with that update, might want to check with Spamaps in #ubuntu-devel first so as not to duplicate work [18:38] upstream changelog says http://paste.ubuntu.com/752778/ [18:38] ok.. I'm going.. [18:47] what diff between this channel and ubuntu-devel ?! [18:48] should I join both to get involved in ubuntu development cycle ?! [18:50] l3on: well, this channel is more geared towards universe/multiverse packages and mentoring, whereas ubuntu-devel is focused on main/restricted, patch piloting, and other development related discussions (such as transitions) [19:26] mmm... in precise: [19:27] perl -e 'use Image::Magick' [19:27] perl: symbol lookup error: /usr/lib/perl5/auto/Image/Magick/Magick.so: undefined symbol: MagickCoreGenesis [20:18] oh hey - the backports pocket does enforce some form of archive upload permissions! "tomcat6_6.0.18-0ubuntu3.3~hardy1_source.changes rejected [...] Signer is not permitted to upload to the component 'main'." [20:24] :) [20:25] Laney: finished running [20:41] hey guys... could you explain me something ? :) [20:41] I would understand... how to fix -as-needed .. [20:42] l3on: https://wiki.ubuntu.com/NattyNarwhal/ToolchainTransition [20:42] what I get is: put libs "-laaabbb" alway after "-o output" ... [20:42] tumbleweed, I'm reading that :) [20:42] no. put -laaabbb after foo.o [20:42] but.. what is not clear... what to do with '-L' ? [20:42] yeah yeah tumbleweed, I wrote it wrong... :) [20:43] -L probably doesn't matter here [20:43] ok, let's try :) [20:44] how exactly you go about fixing the build failure will depend on the package [20:44] most common cause is people misusing LDFLAGS [20:44] tumbleweed, problem here is not a fail in building, but something wrong at runtime [20:44] bug 897380 [20:44] Launchpad bug 897380 in imagemagick (Ubuntu) "PerlMagick fails at runtime in precise " [Undecided,New] https://launchpad.net/bugs/897380 [20:45] right, that's common for plugins, where we expect some symbols to not be resolved in the shared library [20:46] if only more packages had test suites... [20:47] Laney: just as well I ran that test, the weird e-mail addreses were in the changes files not lp. Duh! === Guest913 is now known as medberry [23:57] l3on: I informed mterry about that, don't know if he was looking at it but you might want to check [23:57] tumbleweed: yeah, that's where all of the dubious data is [23:58] * tumbleweed ran another test, but it got killed by ubuntuwire admins trying to add more disk space :)