=== 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: <spiv> Ok, found it. <spiv> I wonder who changed that. <kiko> spiv, you merged that in?
=== Topic (#launchpad): set by kiko at Tue Sep 21 20:51:14 2004
spivdaf: The vegemite?  Please bring it :)02:13
=== SteveA [~steve@224-21-161-212.DSL.ONCOLT.COM] has joined #launchpad
=== SteveA [~steve@224-21-161-212.DSL.ONCOLT.COM] has joined #launchpad
dafspiv: will do :)04:13
=== 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
dilysNew bug 2038 for Launchpad/Database: Desirable sqlobject improvements07:52
=== 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
limioh joy, new arch changes ;)10:26
=== 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
limiBradB|London :)10:44
limihow's things?10:45
dafhow are you?10:45
dafI hear you've been invading mainland Europe10:45
limibeen down with the flu all weekend, so - recovering ;)10:46
limionly the nasty cough left10:46
limi:] 10:46
dafhow was the conference?10:46
limiand I sound like Tom Waits10:46
limiconference was excellent10:46
limiI saw there were some arch changes, anything I need to know, or update as normal?10:47
dafI'm not sure, actualyl10:48
dafI need to work out whether they affect me too10:48
spivdaf: When will we see you? :)10:48
dafspiv: sometime this afternoon, I expect10:48
spivdaf: Chances are you just need to merge the latest rocketfuel.10:49
dafspiv: I'm meeting Dwayne at 5:3010:49
dafspiv: right, that's what I thought10:49
=== limi goes through the rest of the mails
stublimi: If you havn't merged for a few days I don't think you have to worry - just merge as normal10:51
limiok, great10:51
limiwhere in the world is stub these days?10:52
stubWhich is quite possibly straight down ;)10:52
dafspiv: how did your travels go?10:54
spivOk, a bit rough :)10:54
spivThe bus from Bridgend was full.10:54
spivAnd then the tube from Paddington had half the lines blocked due to a faulty train.11:02
stubBradB|London: Hows the launchpad setup on OSX going?11:10
spivAnd 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
dafheh :)11:13
dafI got a call from Steve at about 6 wondering where you were11:13
dilysNew bug 2039 for Launchpad/Rosetta: write a script for creating PO templates in the database11:17
dilysNew bug 2040 for Launchpad/Rosetta: allow searching for messages and translations within templates11:19
dilysNew bug 2041 for Launchpad/Rosetta: paginate Rosetta projects page11:26
=== carlos [~carlos@DWB-100-96.go.retevision.es] has joined #launchpad
carloslifeless: ping11:41
=== limi [~limi@] has joined #launchpad
lifelesscarlos: pong11:42
carloslifeless: my last patch is 41511:42
carloslifeless: could I do  a normal star-merge?11:42
carlosthank you11:43
dafspiv: is Steve around?11:46
cprovdaf: hi daf, yes he is here, do you want to talk with him now ?11:50
dafcprov: actually, never mind11:52
cprovdaf: sorry I was slow !!! we are having a meeting with mark11:53
dafcprov: no worries11:54
cprovdaf: ok11:54
dilysNew bug 2042 for Launchpad/Rosetta: PO import should update translation statistics11:58
dilysBug 1902 resolved: Can't add an assignee to a bug12:01
=== SteveA [~steve@host217-37-231-28.in-addr.btopenworld.com] has joined #launchpad
limiSteveA :)12:04
daflimi: you've seen all the bugs we've assigned to you, right? :)12:05
limisoon done with the mail catch-up :)12:08
dafhi Steve12:14
dilysNew bug 2043 for Launchpad/Launchpad: make Launchpad development servers run under the auspices of the launchpad user12:17
=== lalo [~lalo@200-102-142-102.paemt7013.dsl.brasiltelecom.net.br] has joined #launchpad
daflimi: whoops01:03
daflimi: we changed our minds -- it's going to be rosetta-users, not rosetta-testers01:04
limidaf: whoops?01:04
daflalo: hi01:04
=== lalo [~lalo@200-102-142-102.paemt7013.dsl.brasiltelecom.net.br] has joined #launchpad
laloyay composite.01:14
=== lalo goes to the bakery, brb
laloactually I don't01:20
=== daf goes to London
laloyay London01:24
=== limi goes to lunch
=== limi is now known as limi|llunch
sabdfllimi|llunch: when you get back, can you fix #1800 please?01:39
BradB|Londonlimi|llunch: Were you running launchpad on 10.2.8?01:43
lifelesslalo: around ?01:48
laloI am01:48
lifelessdo you know what you need to do to fix your branch ?01:48
lifelessok. First I need to know exactly what you did to cause the problem.01:49
lifelessit looks like you used an =all file - an advanced feature of tla, rarely needed.01:49
laloyeah, I would like to know that too01:49
lalono, I didn't even know these files exist01:50
lifelessdid you run tla explict-default at all ?01:50
lifelessdid you run any third party tla scripts ?01:51
lifelessthat makes it quite a mystery01:51
laloI'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
lifelesswe won't waste a morning.01:51
lifelessabout 10 minutes.01:51
lifelessok, do this:01:52
lifeless$ cd your-launchpad-dir01:52
laloit is *possible* that my revlib went corrupted.01:52
lifelessno, its not a misbehaviour on tla01:52
lifelessthis is a specific behaviour that has been enabled by a commit in your tree, for some unknown reason.01:53
lalolet me fill you in on the history of the problem01:53
laloFriday I submitted a merge, and it failed01:53
lalowhen I tried the same merge in my local checkout of rocketfuel, it didn't fail01:53
lifelessprobably one of the random-segfault-on-chinstrap problems.01:54
lalothen I found that my patch-135 was bogus; if I disabled my revlib, the merge *would* fail locally01:54
lifelessoh, thats interesting.01:55
lifelessthere are string checks to prevent that, I'd love to know how that occured. are you using NFS ?01:55
laloI 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/removes01:55
lalono nfs01:55
lifelesswhat do you mean when you say 'reverted' ?01:56
lalo(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)01:56
lifelesswhat do you mean when you say 'reverted' ?01:58
laloI removed the revision from my archive and the mirror, then replaced my lp tree with a fresh checkout to get rid of bogus patch-logs01:58
lifelessok, future reference: thats called altering history, and you should never ever ever do that.01:59
lifelessit breaks referential integrity on a global basis.01:59
lalosorry, but I don't agree. What's the alternative?01:59
lifelessits not a matter of agreement. Its a fact about how distributed systems interact. 02:00
lifelessas for alternative, there are many.02:00
laloif I know the revision is bogus, either I revert it or I abandon the version02:00
lifelessa) tag from the good part of that branch into that branch again,02:00
lifelessb) tag from the good part of that branch into a different branch.02:00
lifelessc) cacherev a good version of the same patch.02:00
laloI tried c too, before reverting02:01
lifelessdid you ask the arch team about this ?02:01
lifelessor on #arch ?02:01
laloand b is what I did now02:01
lifelessok, for b, have you tagged before or after those bogus deletes where added ?02:01
laloneither. I tagged from rocketfuel and completely abandoned my old version.02:02
=== BradB|London is now known as BradB|lunch
lifelessok. that will do. what was the bad merge that occured - was it from the now abandoned branch ?02:02
lifelessok, thats good.02:03
lifelessso, 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
lifelessbecause referential integrity problems raise havoc - for you more than for me.02:04
laloI 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
laloI wouldn't have done that, for example, if pqm used a revlib.02:06
lifelesspqm does have a revlib.02:06
lifelessif you had not mirror patch-135 to chinstrap, it wouldn't have been nearly as concerning.02:07
lalohmm. the reject message I got seems to imply pretty clearly it wasn't getting my tree from a revlib02:07
lifelessthat doesn't mean that it doesn't have one.02:07
laloah well02:07
laloI mean, if it was using a greedy revlib to get my tree - is that betrer?02:08
laloor even better - if it had my tree on its revlib02:08
lifelessright, it doesn't auto-add, because we have some tests that do bad things to rev libs.02:08
lifelessmy point is that once that there patch hit chinstrap, you no longer had control.02:08
laloin this case I did due to company policy02:09
lifelesscompany policy says no such thing.02:09
lalothe only legitimate users that *should* have my patch are [me, pqm] 02:09
laloam I wrong?02:09
lalosince I thoroughly checked that pqm didn't have it, then I had control02:10
lifelessand 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
lifelessyes, you are wrong.02:10
lifelessI 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
lifelessat 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.02:11
lalo*if* a "bogus rev" situation ever happens again, I'll use "tag from within the same version" and then cacherev just to be sure02:12
lifelesscool. thats a *safe* solution.02:12
laloat least for anything canonical.02:12
lifelessI'd be happy (and interested) in figuring out the root cause should this ever happen again: its a serious bug.02:12
laloI don't really have much clue, but *maybe* I ran out of space in the revlib partition02:13
lalo(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
lifelessah, that would do it.02:14
lifelessactually, no it wouldn't02:14
lifelessnew revs are built in a ,,foo dir, and renamed to the final name.02:14
lalomeanwhile, could you tell pqm about my new version? it's lalo@canonical.com--canonical-work-2004/launchpad--devel--0.002:15
lalohmm true.02:15
lifelessthat will work straight off.02:15
laloreally? funny then02:15
laloI submitted the request twice already yesterday and got neither success or failure02:15
lifelessI disabled pqm temporarily.02:16
laloah, ok02:16
lifelessuntil we'd had this chat.02:16
lalomakes sense.02:16
laloit 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 commit02:18
lifelesscould be02:20
lalobtw, 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
lifelessheh, I'm glad you wanted to understand.02:21
lalowell, 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
lalonow I'll get back to work02:23
=== limi|llunch is now known as limi
limisabdfl: fixed that (#1800) earlier this morning02:34
=== BradB|lunch is now known as BradB
SteveAthat's so confusing.  I looked at https://bugzilla.warthogs.hbd.com/bugzilla/show_bug.cgi?id=180002:36
SteveAand was confused :-)02:36
=== limi is also confused
limiof course, the accounts are not in sync02:40
SteveAwell, the counts are not in sync02:43
BradBAnyone know how to fix the pg_type.h error when trying to build psycopg 1.1.15 on OS X?03:08
BradBI did this: http://lists.initd.org/pipermail/psycopg/2002-December/001630.html, but it's still telling me it can't find it.03:08
stubHow did you install postgres? You might not have the headers installed?03:09
stubThe includes and libs are also often in /whatever/include/pgsql instead of /whatever/pgsql/include03:10
stubBradB: 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.03:11
BradBstub: The fact that it doesn't touch the existing Python install is the problem though, I think.03:12
limilulu: Amazon portlet added03:13
lululimi: thanks I'll have a look03:13
stubBradB: Try 'env INCLUDES="-I/usr/local/pgsql/include" ./configure --whatever-you-had-before' or 'env CFLAGS="-I/usr/local/pgsql/include" ./configure ...'03:16
BradBSame error.03:17
stubWhat was wrong with trying 'port install psycopg' and letting darwin ports install a fresh postgresql and python in /opt/local for launchpad work?03:18
BradBWe'll find out in a few minutes... :)03:22
stubIf 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.03:27
BradBThis'll take a long time yet...presumably I can go ahead with the RocketFuelSetup instructions, right?03:31
stubYup. All you need for that is arch and gpg.03:32
stubIt 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
BradBthat 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.)03:34
=== BradB goes to bug james
dilysBug 2031 resolved: PQM success or failure messages being bounced?03:36
stubBradB: a literal 'tla my-id'03:40
spivBut the firstname.lastname@canonical.com bit should be different ;)03:41
stubBradB, spiv: Do you now if it is possible to do FooTable.select(FooTable.q.foo = 'bar') in a case insensitive manner?03:44
BradBnot that i'm aware of. i did a manual ILIKE hack in a previous project using SQLObject.03:45
BradBSo far things are seeming to go okay with port install.03:47
stubMmm... I need to do it with CONTAINSSTRING actually. So I could just subclass CONTAINSSTRING and make it add a lower() in the right spot.03:49
lalolifeless: all yay, the merge was successful and doesn't seem to contain any bogosity.03:50
lalo*now* I'll go to the bakery; the oatmeal seemingly wasn't enough breakfast.03:54
spivstub: Yeah, there's no builtin way to do it with sqlobject that I know of.04:05
stubShould be possible to make Foo.q.bar.lower() work I think04:07
lalostub: +104:08
spivstub: Yep.04:09
laloin case you *are* looking for opinions, I find that a rather reasonable api for the task.04:09
=== lalo gets back to work
spiv(reminds me a little of http://svn.twistedmatrix.com/cvs/trunk/sandbox/cake.py?view=auto&rev=6369&root=Twisted)04:12
=== cprov [~cprov@host217-37-231-28.in-addr.btopenworld.com] has joined #launchpad
lululimi: are you working on the books portlet at themoment?04:55
liminot at this very second, no04:55
limihow come?04:56
luluok - the display seems to be broken - only showing the heading, not books.04:56
lulualso: in Documentation - could you move it to after the wiki link? cheers!04:56
liminot possible, it lists non-folderish items first04:57
limisorry, folderish04:57
limianyway, it's probably only showing the heading since it was moved04:57
limiI will try a re-index and see if that helps04:58
lululimi: thanks. let me know when you're done fixing it.05:00
limilulu: suddenly the books folder is at the root again - is anybody else moving it around that you know?05:02
limiI moved it into the docs area an hour ago or so05:02
lulunope - only you. It's appearing in the docs area, but not at the bottom05:02
luludid it have something to do with the tarball?05:03
limithat is the software itself, not the content05:03
limihm, just silly caching, it seems05:04
lululimi: 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
limiyes, it isn't working at the moment05:07
sabdfllimi: thanks, great05:11
sabdflstub: what's table x?05:11
limilulu: 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 works05:16
limiwill add the "More books" link next05:16
lululimi: thanks05:16
lululimi: 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?05:17
limiprobably something wrong in the indenting - will try to create an example document for you afterwards05:18
dilysNew bug 2044 for Launchpad/Rosetta: Rosetta should at the LEAST know about browser languages05:26
limilulu: http://www.ubuntulinux.org/examples/stx/05:29
lululimi: tx for that05:32
lululimi: let's keep that there for the moment.05:32
limiI won't touch it ;)05:32
limisent you the URL05:32
lululimi: gotcha.05:44
lululimi: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
lululimi: what's happening on this issue?05:45
limiit's doing what we are asking it to - showing both folderish and non-folderish items in the nav tree05:46
limiwhich 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 up05:47
limiit can't magically guess what types you think should show up ;)05:47
lululimi: 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
lululimi: in the news section05:50
limibut I can't see your mailbox ;)05:50
limiany other types than News Items?05:51
luluemails between us05:51
luluyes - all the help cenre (Documentation)05:51
limihaven't seen anybody mention news items specifically05:51
luluin news on canonical 05:51
luluwhen u add items05:51
luluwill they appear as a list (if more than one) on the index page?05:52
=== npmccallum [~npmccallu@69-162-252-7.ironoh.adelphia.net] has joined #launchpad
=== limi [~limi@] has left #launchpad []
lalospiv: was it you who suggested that I try commit/abort via fetching the connection object built by InitZopeless?07:16
lalohmm, doesn't seem to be around. I'll use the mailing list.07:18
=== 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
spivlalo: Yeah.07:55
spiv 07:55
spivWell, by keeping the connection you pass to initZopeless.07:55
laloyes :-) that doesn't work07:55
spivHow are you calling commit/abort?  What's the error (or just silently fails to do so?)?07:56
lalogive me 2 secs07:57
spivSure.  I'm very curious about this :)07:57
lalook, well07:58
lalothe default setup of sqlo is "autoCommit" - meaning, it commits after *EACH* query07:58
laloso you can, yes, commit or abort at the end of the script - but it will be meaningless :-)07:59
spivOh, right.07:59
laloyou wrap your connection in a Transaction object, and give *that*m to SQLBase08:00
lalothen it works. Kinda. There are two or three very simple bugs in the Transaction class.08:00
spivWhat are the bugs?08:00
spivThe iterSelect one?08:01
spivThat's fixed in SQLObject 0.6, which lifeless will apparently sync into our snapshot sometime this week.08:01
lalothe 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
spivRight :)08:02
spivAny others?  I already know about those two ;)08:02
lalo(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())08:02
spivYep, those are the two I know.08:03
spivNeed iter(...) in Transaction.iterSelect, and __iter__ in Iteration.08:03
spiv0.6 has both of those fixed, and I've attached a patch to a rosetta bug somewhere to do the same thing.08:03
lalogive me two more secs to check if that's all08:04
laloI was trying to commit those to my own sqlobject branch, and got lost in the conflicts :-)08:04
lalonay, sir, that's all.08:06
spivThat's good :)08:06
spivThanks for trying that out.08:06
laloyes. If it's known, and it's on the way to fixing, then it's good08:06
lalowell, I *had* to :-) our import script is kinda broken if we don't have transactions08:07
spivWell, thanks for letting me know how it went, then :)08:07
dilysBug 2011 resolved: Package browser08:38
=== mdz [~mdz@69-167-148-207.vnnyca.adelphia.net] has joined #launchpad
=== justdave [~dave@] has joined #launchpad
=== justdave [~dave@] has joined #launchpad
=== justdave [~dave@] has left #launchpad ["Poof"]
=== justdave [~dave@] has joined #launchpad
=== justdave_ [~dave@] has joined #launchpad
=== stub [~stub@dsl-] has joined #launchpad
=== justdave_ [~dave@] has joined #launchpad
=== stub [~stub@dsl-] has joined #launchpad
SteveAstub: hello09:16

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!