=== cprov [~cprov@224-21-161-212.DSL.ONCOLT.COM] has joined #launchpad === cprov [~cprov@224-21-161-212.DSL.ONCOLT.COM] has joined #launchpad === mdz [~mdz@69-167-148-207.vnnyca.adelphia.net] has joined #launchpad === wartylog [~warthylog@port1845.ds1-khk.adsl.cybercity.dk] has joined #launchpad === Topic for #launchpad: Ok, found it. I wonder who changed that. spiv, you merged that in? === Topic (#launchpad): set by kiko at Tue Sep 21 20:51:14 2004 [02:13] daf: The vegemite? Please bring it :) === SteveA [~steve@224-21-161-212.DSL.ONCOLT.COM] has joined #launchpad === SteveA [~steve@224-21-161-212.DSL.ONCOLT.COM] has joined #launchpad [04:13] spiv: will do :) === npmccallum [~npmccallu@69-162-252-7.ironoh.adelphia.net] has joined #launchpad === SteveA [~steve@224-21-161-212.DSL.ONCOLT.COM] has joined #launchpad === SteveA [~steve@224-21-161-212.DSL.ONCOLT.COM] has joined #launchpad === SteveA [~steve@224-21-161-212.DSL.ONCOLT.COM] has joined #launchpad === SteveA [~steve@224-21-161-212.DSL.ONCOLT.COM] has joined #launchpad === SteveA [~steve@224-21-161-212.DSL.ONCOLT.COM] has joined #launchpad === SteveA [~steve@224-21-161-212.DSL.ONCOLT.COM] has joined #launchpad [07:52] New bug 2038 for Launchpad/Database: Desirable sqlobject improvements === SteveA [~steve@224-21-161-212.DSL.ONCOLT.COM] has joined #launchpad === limi [~limi@sparkit.easynet.no] has joined #launchpad === lulu [~lu@host217-37-231-28.in-addr.btopenworld.com] has joined #launchpad [10:26] oh joy, new arch changes ;) === cprov [~cprov@host217-37-231-28.in-addr.btopenworld.com] has joined #launchpad === BradB|London [~bradb@host217-37-231-28.in-addr.btopenworld.com] has joined #launchpad [10:44] BradB|London :) [10:45] limi! [10:45] daf! [10:45] how's things? [10:45] fine! [10:45] how are you? [10:45] I hear you've been invading mainland Europe [10:46] been down with the flu all weekend, so - recovering ;) [10:46] only the nasty cough left [10:46] ouch [10:46] :] [10:46] how was the conference? [10:46] and I sound like Tom Waits [10:46] conference was excellent [10:47] I saw there were some arch changes, anything I need to know, or update as normal? [10:48] I'm not sure, actualyl [10:48] I need to work out whether they affect me too [10:48] daf: When will we see you? :) [10:48] spiv: sometime this afternoon, I expect [10:49] daf: Chances are you just need to merge the latest rocketfuel. [10:49] spiv: I'm meeting Dwayne at 5:30 [10:49] spiv: right, that's what I thought === limi goes through the rest of the mails [10:51] limi: If you havn't merged for a few days I don't think you have to worry - just merge as normal [10:51] ok, great [10:52] where in the world is stub these days? [10:52] Melbourne [10:52] Which is quite possibly straight down ;) [10:54] spiv: how did your travels go? [10:54] Ok, a bit rough :) [10:54] The bus from Bridgend was full. [11:02] And then the tube from Paddington had half the lines blocked due to a faulty train. [11:04] urg [11:10] BradB|London: Hows the launchpad setup on OSX going? [11:13] And I realised I didn't know the address of the hotel, just that it was near Earl's Court station... thankfully a quick map purchase solved that :) [11:13] heh :) [11:13] I got a call from Steve at about 6 wondering where you were [11:17] New bug 2039 for Launchpad/Rosetta: write a script for creating PO templates in the database [11:19] New bug 2040 for Launchpad/Rosetta: allow searching for messages and translations within templates [11:26] New bug 2041 for Launchpad/Rosetta: paginate Rosetta projects page === carlos [~carlos@DWB-100-96.go.retevision.es] has joined #launchpad [11:41] hi [11:41] lifeless: ping === limi [~limi@193.71.38.142] has joined #launchpad [11:42] carlos: pong [11:42] lifeless: my last patch is 415 [11:42] lifeless: could I do a normal star-merge? [11:42] yes [11:43] ok [11:43] thank you [11:46] spiv: is Steve around? [11:50] daf: hi daf, yes he is here, do you want to talk with him now ? [11:52] cprov: actually, never mind [11:53] daf: sorry I was slow !!! we are having a meeting with mark [11:54] cprov: no worries [11:54] daf: ok [11:58] New bug 2042 for Launchpad/Rosetta: PO import should update translation statistics [12:01] Bug 1902 resolved: Can't add an assignee to a bug === SteveA [~steve@host217-37-231-28.in-addr.btopenworld.com] has joined #launchpad [12:04] SteveA :) [12:05] limi: you've seen all the bugs we've assigned to you, right? :) [12:08] yup [12:08] soon done with the mail catch-up :) [12:10] :) [12:14] hi [12:14] hi Steve [12:17] New bug 2043 for Launchpad/Launchpad: make Launchpad development servers run under the auspices of the launchpad user === lalo [~lalo@200-102-142-102.paemt7013.dsl.brasiltelecom.net.br] has joined #launchpad [01:03] limi: whoops [01:04] limi: we changed our minds -- it's going to be rosetta-users, not rosetta-testers [01:04] daf: whoops? [01:04] aha [01:04] lalo: hi === lalo [~lalo@200-102-142-102.paemt7013.dsl.brasiltelecom.net.br] has joined #launchpad [01:14] yay composite. === lalo goes to the bakery, brb [01:20] actually I don't === daf goes to London [01:24] yay London === limi goes to lunch === limi is now known as limi|llunch [01:36] ;) [01:39] limi|llunch: when you get back, can you fix #1800 please? [01:43] limi|llunch: Were you running launchpad on 10.2.8? [01:48] lalo: around ? [01:48] I am [01:48] do you know what you need to do to fix your branch ? [01:49] no [01:49] ok. First I need to know exactly what you did to cause the problem. [01:49] it looks like you used an =all file - an advanced feature of tla, rarely needed. [01:49] yeah, I would like to know that too [01:50] no, I didn't even know these files exist [01:50] did you run tla explict-default at all ? [01:50] nope [01:51] did you run any third party tla scripts ? [01:51] nope [01:51] that makes it quite a mystery [01:51] I'd much rather start a new version and work from there. It is quite possible that I won't work on Launchpad anymore after next week, so I don't think it's efficient to waste one morning trying to figure out what went wrong in my tree. [01:51] we won't waste a morning. [01:51] about 10 minutes. [01:52] ok, do this: [01:52] sigh. [01:52] $ cd your-launchpad-dir [01:52] it is *possible* that my revlib went corrupted. [01:52] no, its not a misbehaviour on tla [01:53] this is a specific behaviour that has been enabled by a commit in your tree, for some unknown reason. [01:53] let me fill you in on the history of the problem [01:53] Friday I submitted a merge, and it failed [01:53] when I tried the same merge in my local checkout of rocketfuel, it didn't fail [01:54] probably one of the random-segfault-on-chinstrap problems. [01:54] then I found that my patch-135 was bogus; if I disabled my revlib, the merge *would* fail locally [01:55] oh, thats interesting. [01:55] there are string checks to prevent that, I'd love to know how that occured. are you using NFS ? [01:55] I reverted 135, removed it from the revlib, and committed again; this time the merge succeded both locally and in pqm, *but* it had those bogus add/removes [01:55] no [01:55] no nfs [01:56] what do you mean when you say 'reverted' ? [01:56] (in fact I didn't remove 135 from the revlib - I removed lalo@canonical.com--2004/launchpad--devel--0 entirely from the revlib and let it rebuild) [01:58] what do you mean when you say 'reverted' ? [01:58] I removed the revision from my archive and the mirror, then replaced my lp tree with a fresh checkout to get rid of bogus patch-logs [01:59] ok, future reference: thats called altering history, and you should never ever ever do that. [01:59] it breaks referential integrity on a global basis. [01:59] sorry, but I don't agree. What's the alternative? [02:00] its not a matter of agreement. Its a fact about how distributed systems interact. [02:00] as for alternative, there are many. [02:00] if I know the revision is bogus, either I revert it or I abandon the version [02:00] a) tag from the good part of that branch into that branch again, [02:00] b) tag from the good part of that branch into a different branch. [02:00] c) cacherev a good version of the same patch. [02:01] I tried c too, before reverting [02:01] did you ask the arch team about this ? [02:01] or on #arch ? [02:01] and b is what I did now [02:01] ok, for b, have you tagged before or after those bogus deletes where added ? [02:02] neither. I tagged from rocketfuel and completely abandoned my old version. === BradB|London is now known as BradB|lunch [02:02] ok. that will do. what was the bad merge that occured - was it from the now abandoned branch ? [02:03] yes [02:03] ok, thats good. [02:04] so, for future reference, I repeat: do not alter revisions in your archive without discussing alternatives (we can even repair 'bogus patches' in all likelyhood) with someone on the arch team! [02:04] because referential integrity problems raise havoc - for you more than for me. [02:06] I think you're not being pragmatic with this referential integrity thing. If I have control over all places the revision may have ever been referred to, "altering history" may well be the best thing to do. [02:06] I wouldn't have done that, for example, if pqm used a revlib. [02:06] pqm does have a revlib. [02:07] if you had not mirror patch-135 to chinstrap, it wouldn't have been nearly as concerning. [02:07] hmm. the reject message I got seems to imply pretty clearly it wasn't getting my tree from a revlib [02:07] that doesn't mean that it doesn't have one. [02:07] ah well [02:08] I mean, if it was using a greedy revlib to get my tree - is that betrer? [02:08] better [02:08] or even better - if it had my tree on its revlib [02:08] right, it doesn't auto-add, because we have some tests that do bad things to rev libs. [02:08] my point is that once that there patch hit chinstrap, you no longer had control. [02:09] in this case I did due to company policy [02:09] ?!!??!! [02:09] company policy says no such thing. [02:09] the only legitimate users that *should* have my patch are [me, pqm] [02:09] am I wrong? [02:10] since I thoroughly checked that pqm didn't have it, then I had control [02:10] and anyone in the company who hacks on launchpad, who might want to review current status, or have a local mirror. there is nothing in the arch policy against either of those uses. [02:10] yes, you are wrong. [02:11] I agree, that in specific circumstances, replacing the revision is the right thing to do, but you have taken the referential integrity concerns far to lightly IMO. [02:11] ok [02:11] at a bare minimum you should have emailed everyone with potential access to your mirror stating that you've replace the patch so that we can take whatever actions are needed. [02:12] *if* a "bogus rev" situation ever happens again, I'll use "tag from within the same version" and then cacherev just to be sure [02:12] cool. thats a *safe* solution. [02:12] at least for anything canonical. [02:12] I'd be happy (and interested) in figuring out the root cause should this ever happen again: its a serious bug. [02:13] ok [02:13] I don't really have much clue, but *maybe* I ran out of space in the revlib partition [02:14] (I did run out of space, but I don't remember if it was in the same day as this happened, or even if I made any LP commit while I was out of space) [02:14] ah, that would do it. [02:14] actually, no it wouldn't [02:14] new revs are built in a ,,foo dir, and renamed to the final name. [02:15] meanwhile, could you tell pqm about my new version? it's lalo@canonical.com--canonical-work-2004/launchpad--devel--0.0 [02:15] hmm true. [02:15] that will work straight off. [02:15] really? funny then [02:15] I submitted the request twice already yesterday and got neither success or failure [02:16] I disabled pqm temporarily. [02:16] ah, ok [02:16] until we'd had this chat. [02:16] makes sense. [02:18] it may also be that the revlib has been bogus for quite long, maybe since I had bad ram, and the problem only surfaced last week due to some random coincidence in the commit [02:20] could be [02:21] btw, when I argue it's because I want to understand your point of view. In this case, you convinced me. If I simply believed I knew better than you I'd shut up and not care. [02:21] heh, I'm glad you wanted to understand. [02:23] well, I know more about arch than about i18n, and if you s/know/care/ the sentence remains true. I wish I had started *working* on it earlier, maybe I would be in the arch team instead of lp. [02:23] now I'll get back to work === limi|llunch is now known as limi [02:34] sabdfl: fixed that (#1800) earlier this morning === BradB|lunch is now known as BradB [02:36] that's so confusing. I looked at https://bugzilla.warthogs.hbd.com/bugzilla/show_bug.cgi?id=1800 [02:36] and was confused :-) === limi is also confused [02:40] of course, the accounts are not in sync [02:43] well, the counts are not in sync [03:08] Anyone know how to fix the pg_type.h error when trying to build psycopg 1.1.15 on OS X? [03:08] I did this: http://lists.initd.org/pipermail/psycopg/2002-December/001630.html, but it's still telling me it can't find it. [03:09] How did you install postgres? You might not have the headers installed? [03:10] The includes and libs are also often in /whatever/include/pgsql instead of /whatever/pgsql/include [03:11] BradB: btw. darwin ports won't touch your existing Python install so you can give that a go (assuming your python isn't installed in /opt/local), so it still might be worth giving 'port install psycopg' a go to see what happens. [03:12] stub: The fact that it doesn't touch the existing Python install is the problem though, I think. [03:13] http://paste.husk.org/1697 [03:13] lulu: Amazon portlet added [03:13] limi: thanks I'll have a look [03:16] BradB: Try 'env INCLUDES="-I/usr/local/pgsql/include" ./configure --whatever-you-had-before' or 'env CFLAGS="-I/usr/local/pgsql/include" ./configure ...' [03:17] Same error. [03:18] What was wrong with trying 'port install psycopg' and letting darwin ports install a fresh postgresql and python in /opt/local for launchpad work? [03:22] We'll find out in a few minutes... :) [03:27] If you go back to manually trying psycopg, there should be a confsomethingorother.log which will list the commands it was trying to run to test if pg_type.h exists. It might provide insight to why it failed. [03:31] This'll take a long time yet...presumably I can go ahead with the RocketFuelSetup instructions, right? [03:32] Yup. All you need for that is arch and gpg. [03:32] ok [03:33] It takes a while the first time if you havn't used arch before, so it might be worth getting someone there to shoulder surf. [03:33] ok [03:34] that tla my-id line: is that literally "tla my-id" or does some value belong in place of my-id? (i.e. the one that james [i think?] is going to be creating for me.) === BradB goes to bug james [03:36] Bug 2031 resolved: PQM success or failure messages being bounced? [03:40] BradB: a literal 'tla my-id' [03:41] But the firstname.lastname@canonical.com bit should be different ;) [03:42] heh [03:44] BradB, spiv: Do you now if it is possible to do FooTable.select(FooTable.q.foo = 'bar') in a case insensitive manner? [03:45] not that i'm aware of. i did a manual ILIKE hack in a previous project using SQLObject. [03:47] So far things are seeming to go okay with port install. [03:49] Mmm... I need to do it with CONTAINSSTRING actually. So I could just subclass CONTAINSSTRING and make it add a lower() in the right spot. [03:50] lifeless: all yay, the merge was successful and doesn't seem to contain any bogosity. [03:54] *now* I'll go to the bakery; the oatmeal seemingly wasn't enough breakfast. [04:05] stub: Yeah, there's no builtin way to do it with sqlobject that I know of. [04:07] Should be possible to make Foo.q.bar.lower() work I think [04:08] stub: +1 [04:09] stub: Yep. [04:09] in case you *are* looking for opinions, I find that a rather reasonable api for the task. === lalo gets back to work [04:12] (reminds me a little of http://svn.twistedmatrix.com/cvs/trunk/sandbox/cake.py?view=auto&rev=6369&root=Twisted) === cprov [~cprov@host217-37-231-28.in-addr.btopenworld.com] has joined #launchpad [04:54] limi:ping [04:55] pong [04:55] limi: are you working on the books portlet at themoment? [04:55] not at this very second, no [04:56] how come? [04:56] ok - the display seems to be broken - only showing the heading, not books. [04:56] also: in Documentation - could you move it to after the wiki link? cheers! [04:57] not possible, it lists non-folderish items first [04:57] sorry, folderish [04:57] anyway, it's probably only showing the heading since it was moved [04:58] I will try a re-index and see if that helps [05:00] limi: thanks. let me know when you're done fixing it. [05:02] lulu: suddenly the books folder is at the root again - is anybody else moving it around that you know? [05:02] I moved it into the docs area an hour ago or so [05:02] nope - only you. It's appearing in the docs area, but not at the bottom [05:03] did it have something to do with the tarball? [05:03] no [05:03] that is the software itself, not the content [05:04] hm, just silly caching, it seems [05:04] :) [05:07] limi: let me know when done - I've tried it in IE and Netscape and cleared my cache on Safari and Firefox. Still getting the problem on the books portlet. [05:07] yes, it isn't working at the moment [05:11] limi: thanks, great [05:11] stub: what's table x? [05:16] lulu: seems to be a bug in the ATAmazon product, I will report it to the author. moved the books to the top level for now, so the portlet works [05:16] will add the "More books" link next [05:16] limi: thanks [05:17] limi: headings and subheadings in Structured Text. I get a main heading with 2 - 4 spaces - can't seem to get a subheading unless I use html. what do you think the problem is? [05:18] probably something wrong in the indenting - will try to create an example document for you afterwards [05:26] New bug 2044 for Launchpad/Rosetta: Rosetta should at the LEAST know about browser languages [05:29] lulu: http://www.ubuntulinux.org/examples/stx/ [05:32] limi: tx for that [05:32] limi: let's keep that there for the moment. [05:32] I won't touch it ;) [05:32] sent you the URL [05:44] limi: gotcha. [05:45] limi:news item on canonical - can you set it so that they don't appear in the nav portlet....i assume this is the same prob as help centre items appearing in the nav. [05:45] limi: what's happening on this issue? [05:46] it's doing what we are asking it to - showing both folderish and non-folderish items in the nav tree [05:47] which will cause items like that to show up - which means we will have to explicitly tell it to hide certain types that you don't want to show up [05:47] it can't magically guess what types you think should show up ;) [05:50] limi: not asking you to magically guess - we have had emails on it - please refer to FAQ's appearing in the nav portlet - scalability issue. It's the same in news items... [05:50] limi: in the news section [05:50] but I can't see your mailbox ;) [05:51] any other types than News Items? [05:51] emails between us [05:51] yes - all the help cenre (Documentation) [05:51] centre [05:51] haven't seen anybody mention news items specifically [05:51] in news on canonical [05:51] when u add items [05:52] will they appear as a list (if more than one) on the index page? === npmccallum [~npmccallu@69-162-252-7.ironoh.adelphia.net] has joined #launchpad === limi [~limi@193.71.38.142] has left #launchpad [] [07:16] spiv: was it you who suggested that I try commit/abort via fetching the connection object built by InitZopeless? [07:18] hmm, doesn't seem to be around. I'll use the mailing list. === lulu [~lu@host217-37-231-28.in-addr.btopenworld.com] has left #launchpad [] === npmccallum [~npmccallu@69-162-252-7.ironoh.adelphia.net] has joined #launchpad [07:55] lalo: Yeah. [07:55] [07:55] Well, by keeping the connection you pass to initZopeless. [07:55] hello [07:55] yes :-) that doesn't work [07:55] Hmm! [07:56] How are you calling commit/abort? What's the error (or just silently fails to do so?)? [07:57] give me 2 secs [07:57] Sure. I'm very curious about this :) [07:58] ok, well [07:58] the default setup of sqlo is "autoCommit" - meaning, it commits after *EACH* query [07:59] so you can, yes, commit or abort at the end of the script - but it will be meaningless :-) [07:59] Oh, right. [07:59] so [08:00] you wrap your connection in a Transaction object, and give *that*m to SQLBase [08:00] then it works. Kinda. There are two or three very simple bugs in the Transaction class. [08:00] What are the bugs? [08:01] The iterSelect one? [08:01] yes [08:01] That's fixed in SQLObject 0.6, which lifeless will apparently sync into our snapshot sometime this week. [08:02] the iterSelect two, actually, if you also consider the fact that a list is not a valid thing to return from an __iter__ method :-) [08:02] Right :) [08:02] Any others? I already know about those two ;) [08:02] (or if this is the one you know, then the other one is that Iteration has to have a __iter__ method to be passed into list()) [08:03] Yep, those are the two I know. [08:03] Need iter(...) in Transaction.iterSelect, and __iter__ in Iteration. [08:03] 0.6 has both of those fixed, and I've attached a patch to a rosetta bug somewhere to do the same thing. [08:04] give me two more secs to check if that's all [08:04] I was trying to commit those to my own sqlobject branch, and got lost in the conflicts :-) [08:04] Heh. [08:06] nay, sir, that's all. [08:06] That's good :) [08:06] Thanks for trying that out. [08:06] yes. If it's known, and it's on the way to fixing, then it's good [08:07] well, I *had* to :-) our import script is kinda broken if we don't have transactions [08:07] Well, thanks for letting me know how it went, then :) [08:08] np [08:38] Bug 2011 resolved: Package browser === mdz [~mdz@69-167-148-207.vnnyca.adelphia.net] has joined #launchpad === justdave [~dave@24.247.63.44.gha.mi.chartermi.net] has joined #launchpad === justdave [~dave@24.247.63.44.gha.mi.chartermi.net] has joined #launchpad === justdave [~dave@24.247.63.44.gha.mi.chartermi.net] has left #launchpad ["Poof"] === justdave [~dave@24.247.63.44.gha.mi.chartermi.net] has joined #launchpad === justdave_ [~dave@24.247.63.44.gha.mi.chartermi.net] has joined #launchpad === stub [~stub@dsl-246.248.240.220.dsl.comindico.com.au] has joined #launchpad === justdave_ [~dave@24.247.63.44.gha.mi.chartermi.net] has joined #launchpad === stub [~stub@dsl-246.248.240.220.dsl.comindico.com.au] has joined #launchpad [09:16] stub: hello