[12:06] <sabdfl> ddaa: phew, i was worried about our weenie buildbot
[12:06] <sabdfl> seemed almost... french
[12:06] <sabdfl> much happier with a manly pyarch
[12:08] <lifeless> a manly man
[12:12] <lifeless> spiv: so, its a larger job than you thought?
[12:13] <spiv> lifeless: A little..  I misunderstood lalo's work a little.
[12:14] <lifeless> lalos work will be useful for me, FWICT.
[12:14] <lifeless> auto-commits bad for taxi, v. bad.
[12:14] <spiv> Yeah, I'm basing this off his stuff.
[12:15] <spiv> It's a good start, it just doesn't go far enough :)
[12:15] <spiv> I'm close to having the threads-happy code written and factored somewhat nicely, but I need to test it a bit before I make you my guinea pig for it ;)
[12:16] <lifeless> sniff sniff, fart, sniff sniff fart
[12:17] <spiv> lifeless: commentary on my work, or diary of you morning?  Or would I rather not know? ;)
[12:17] <lifeless> being a guinea pig.
[12:17] <lifeless> you ever observed one of those beasties ?
[12:19] <spiv> Not really.
[12:20] <lifeless> well then. 
[12:21] <spiv> Oh, I see.
[12:35] <BradB> drinks all around!
[12:39] <lifeless> ddaa: where are you at now with stuff ?
[12:41] <ddaa> Still not run buildbot again with the fixed PYTHONPATH
[12:42] <ddaa> been fixing my arch-submit-merge, checking the test suite failure, got phoned twice by gfriend...
[12:46] <Kinnison> just gotta get distroarchreleases hacked in and I can do a "gold" import of a katie snapshot
[12:47] <cprov> Kinnison: is gina working on new DB  ?
[12:51] <ddaa> lifeless: anything in particular you want to tell me. I'm going to bed soon.
[12:53] <Kinnison> cprov: She will be
[12:54] <lifeless> ddaa: I need you to run with the fixed python path and tell me it works.
[12:54] <Kinnison> cprov: is this dump only i386?
[12:54] <lifeless> then, you are set to start checking off imports tomorrow, and I can relax a little.
[12:55] <ddaa> lifeless: sent you the message just after starting the import run...
[12:55] <ddaa> waiting for it to complete
[12:55] <Kinnison> cprov: aah no; I see; gina only does one arch at a time
[12:55] <lifeless> ohohoh, bad file descriptor again.
[12:55] <ddaa> lifeless: 
[12:55] <ddaa> not my descriptor! :-P
[12:56] <lifeless> dude:
[12:56] <lifeless> importdtest@chinstrap:~$ PYTHONPATH=~/buildbot/launchpad/lib:~/buildbot/launchpad/lib/canonical/sourcerer/util python
[12:56] <lifeless> Python 2.3.4 (#2, Jun 21 2004, 11:11:53) 
[12:56] <lifeless> [GCC 3.3.4 (Debian)]  on linux2
[12:56] <lifeless> Type "help", "copyright", "credits" or "license" for more information.
[12:56] <lifeless> >>> import twisted
[12:56] <lifeless> >>> import arch
[12:56] <lifeless> >>> print arch._tla.spawn
[12:56] <lifeless> that is so not the right spawning strategy.
[12:56] <lifeless> <arch._tla.PyArchSpawningStrategy object at 0x4043754c>
[12:56] <Kinnison> cprov: if I can get her importing overnight then I'll commit the changes tomorrow
[12:57] <lifeless> it acked
[12:57] <ddaa> check for my archive if in doubt
[12:57] <ddaa> pqm has not yet sent the reply...
[12:57] <lifeless> patch-49 in rocketfuel.
[12:57] <ddaa> Now it should be DelayedGuessedSpawningStrategy
[12:58] <lifeless> oh, silly me.
[12:58] <lifeless> blush.
[12:58] <lifeless> had missed a star-merge.
[01:01] <cprov> Kinnison: I can't create the new DB on zhongshan, so I can't run them. have you modified gina or is it working fine ?
[01:01] <lifeless> elmo: around ?
[01:04] <elmo> lifeless: yeah
[01:05] <lifeless> any luck on that basic->certificate conversion ?
[01:05] <cprov> elmo: hi dude, can you grant me perms to createlang on zhongshan ?
[01:06] <elmo> cprov: meh
[01:06] <Kinnison> cprov: I had a lot of modifications to gina
[01:06] <elmo> lifeless: I'm ready - I asked Steve when I could convert rosetta, didn't get a response..
[01:06] <Kinnison> cprov: I just commented out the plpython stuff from the makefile
[01:06] <lifeless> elmo: now.
[01:06] <elmo> lifeless: rosetta?
[01:06] <elmo> or macquarie?
[01:06] <lifeless> (its launchpad, not rosetta)
[01:06] <lifeless> yeah, macquarie.
[01:07] <elmo> nono, rosetta the machine
[01:07] <lifeless> I'm blocked - can't do any conversions - till this happens.
[01:07] <lifeless> elmo: I don't know about rosetta - I don't use it.
[01:07] <lifeless> rosetta the machine I mean
[01:08] <elmo> cprov: I hgave you createuser too - if that doesn't work, shout.. or do what kinnison did
[01:09] <Kinnison> elmo: Can you close that psql so I can re-build my db? postgres won't let me createdb while someone is using template1
[01:09] <cprov> elmo: now works :)
[01:10] <cprov> elmo: tks
[01:10] <Kinnison> cprov: if you want to play; take a copy of gina from my ~ on zhongshan
[01:10] <Kinnison> cprov: although my gina does expect a librarian too and I'd appreciate you not putting stuff into the librarian I have running just now :-)
[01:10] <elmo> Kinnison: done
[01:10] <elmo> lifeless: ok one sec
[01:11] <cprov> Kinnison: ok
[01:12] <Kinnison> Well; I'm leaving gina importing against the new DB and I'm off to bed. If all goes well I'll commit the changes tomorrow morning
[01:12] <Kinnison> And if it went *really well* then I'll be starting to extract archives from the db again tomorrow :-)
[01:12] <cprov> Kinnison: anyway I intend to run gina and nicole today to have a full DB tomorrow
[01:12] <Kinnison> g'night all
[01:12] <lifeless> night
[01:15] <ddaa> lifeless: test inconclusive... same weird exceptions in enforceTargetNewOrCleanAndUpdated
[01:15] <ddaa> then I have some issues setting up the sandboxing right
[01:15] <lifeless> ddaa: thats strange, I put that in as a 'it should never happen but lets be thorough' check.
[01:16] <ddaa> That's prolly a problem with my revlib.
[01:16] <ddaa> aalib get revlib entries
[01:16] <lifeless> ah.
[01:16] <ddaa> thus the attempts to sandboxing
[01:16] <ddaa> will continue to go down that road tomorrow
[01:16] <lifeless> well, if I may suggest - shortest path - disable your rev lib.
[01:17] <lifeless> if that fixes is, so that it then works, then start sandboxing.
[01:17] <ddaa> That's not shortest path. I need the revlib for my actual VCS needs.
[01:17] <lifeless> aalib is known bad, it has default branches.
[01:17] <lifeless> anything in arch.ubuntu.com is known good.
[01:18] <ddaa> Great. I'll test a few of them to become confident that my setup is right, and then I'll bodly go when no buildbot has gone before.
[01:18] <lifeless> ddaa: it is shortest path - it eliminates a variable, and doing an a52dec test run is fast - about 10 minutes.
[01:18] <ddaa> a52dec breaks, you remember?
[01:18] <lifeless> when that works, then reenable the revlib, and try sanboxed.
[01:19] <lifeless> how does it break ? (in what manner)
[01:19] <Kinnison> night all
[01:19] <Kinnison> (really this time)
[01:19] <ddaa> In any case I have to go bed now.
[01:19] <lifeless> ok, night ddaa
[01:19] <lifeless> thanks
[01:19] <lifeless> bah, chinstrap is still 70
[01:19] <lifeless> % idle.
[01:21] <ddaa> lifeless: it breaks on the cvs version detection.
[01:22] <ddaa> I showed you the backtrace yesterday.
[01:22] <lifeless> ddaa: oh thats right. 
[01:22] <lifeless> did you breakpoint the cvs version call and see what it was tring to do ?
[01:22] <ddaa> I did not.
[01:22] <lifeless> or equally, add a logger.* message.
[01:22] <lifeless> ok. go to bed, see you tomorrow
[01:23] <sabdfl> cprov: excellent, looking forward to seeing the results of the run
[01:23] <sabdfl> i am finishing up the product release and product series stuff now
[01:24] <sabdfl> then i'd like to update nicole so that for each project it finds, it create the ProductSeries for each branch, then populates it with ProductReleases
[01:24] <ddaa> G'night warthogs, guinea pigs, suidae of all sorts
[01:24] <sabdfl> night ddaa
[01:30] <cprov> sabdfl: sorry, I was afk for while, yes gina is running on zhongshan !
[01:32] <cprov> sabdfl: I will work on you request for ProductSeries til tomorrow
[01:34] <sabdfl> cprov: ok, tomorrow is fine
[01:34] <sabdfl> should be easy for you to create them now
[01:34] <sabdfl> and by tomorrow afternoon, productrelease should be done too
[01:36] <cprov> sabdfl: of course, add comments on bug #2088 if you need more something  
[01:53] <elmo> lifeless: sent you a client cert - please tell me if it works for you
[01:56] <lifeless> checking
[01:59] <lifeless> elmo: erm:
[01:59] <lifeless> Hi,
[01:59] <lifeless> <#part type="application/octet-stream" filename="~/ubuntu/ssl/ca/launchpad.p12" disposition=attachment description="launchpad client cert">

[01:59] <lifeless> that all I got
[01:59] <lifeless> --
[01:59] <lifeless> James
[01:59] <sabdfl> lifeless: you should save it locally, import it into your browser
[02:00] <lifeless> sabdfl: I know that, but I don't seem to /have/ it
[02:00] <sabdfl> ok
[02:02] <lifeless> elmo: yea, every way I try, I only end up with those 7 lines
[02:07] <elmo> ah, heh
[02:07] <elmo> I'm being stupid
[02:07] <elmo> lifeless: resent properly
[02:08] <lifeless> :] 
[02:09] <sabdfl> elmo: to me too please if it's a different one
[02:12] <lifeless> SteveA: around ?
[02:17] <lifeless> spiv: around ?
[02:17] <spiv> lifeless: Yeah.
[02:18] <lifeless> I need a launchpad account on the production launchpad
[02:18] <lifeless> my 'person' already exists.
[02:18] <spiv> lifeless: Ah, hmm.
[02:18] <spiv> gimme a sec, I'll see if you have a pw set.
[02:18] <lifeless> or should anyway, for the creation of the sourcesources.
[02:19] <spiv> You don't.
[02:19] <spiv> Hmm, I think there's a "reset my password" thing you can use?
[02:20] <lifeless> spiv: does it have an email address for me ?(I'm asking you as I haven't studied the guts of this..)
[02:21] <lifeless> if so, which email addy
[02:21] <spiv> It has robertc@robertcollins.net
[02:21] <lifeless> lets see if htis works
[02:24] <lifeless> yay
[02:24] <lifeless> woohoo!
[02:24] <lifeless> I'm IN BABY
[02:24] <lifeless> yayayayyayay
[02:24] <spiv> Excellent :)
[02:24] <lifeless> so, hows taxi-unbreakme-coming along ?
[02:24] <spiv> lifeless: So, how can I test my changes to taxi?
[02:25] <spiv> Heh :)
[02:25] <lifeless> well, you can setup a whole test environment...
[02:25] <lifeless> or just commit them and I'll cherry pick em into chinstrap and test
[02:25] <lifeless> commit but not merge I mean
[02:25] <spiv> Ok.
[02:25] <spiv> This will be entertaining...
[02:26] <lifeless> coool, computerbank are selling pc's with ubuntu on them.
[02:27] <spiv> Sweet.
[02:28] <lifeless> what projects do I need to update ?
[02:28] <spiv> You'll need sqlobject--test--0.6 if you're not already using it.
[02:28] <lifeless> done
[02:29] <spiv> I also have changes to launchpad, configs--buildbot-lifelesslap, and and buildbot
[02:30] <lifeless> ok.
[02:30] <lifeless> what changed in configs--buidlbot-lifelesslap ?
[02:30] <spiv> One line :)
[02:30] <lifeless>  ... ?
[02:32] <spiv> andrew.bennetts@canonical.com/configs--buildbot--0--patch-1 is the change you need.
[02:32] <spiv> Sorry, two lines, but one of them is an import.
[02:33] <lifeless> lets start with lunchbox
[02:34] <spiv> launchpad changes are in my launchpad--zopeless-transactions--0
[02:34] <spiv> which is still mirroring...
[02:34] <lifeless> ok
[02:34] <spiv> (and cau(and causing me lag)
[02:35] <spiv> Mirroring done.
[02:40] <lifeless> ok
[02:42] <spiv> My buildbot (i.e. taxi) changes are in my buildbot--devel--0
[02:47] <lifeless> are there sqlobject changes?
[02:49] <lifeless> ok, in my buildbot, I don't initialise the db in processDB
[02:49] <lifeless> :)
[02:50] <lifeless> lets try this sucker out
[02:51] <lifeless> I've started a test job, takes about 10 to import it
[03:00] <lifeless> ok, you missed an import in taxi
[03:00] <lifeless> at the top - can't import arch.database no more
[03:01] <lifeless> and database.connect is still called.
[03:01] <lifeless> as well as database.commit
[03:01] <lifeless> hmm
[10:09] <limi> morning, all
[10:22] <sabdfl> hi limi
[10:22] <sabdfl> rc today!
[10:23] <limi> cool
[10:23] <limi> :)
[10:38] <Kinnison> Morning
[10:39] <limi> how is Lithuania this morning, Steve? below zero yet? :)
[10:39] <limi> morning, Kinnison
[10:47] <SteveA> 2 degrees above freezing, with a clear sunny sky
[10:57] <limi> same templates as the ones supposedly fixed yesterday:
[10:57] <limi> lib/canonical/launchpad/templates/bugsystems-index.pt   x_Mark_Shuttleworth_<mark.shuttleworth@canonical.com>_Mon_Oct__4_21:56:28_2004_8303.0
[10:57] <limi> lib/canonical/launchpad/templates/bugtrackers-index.pt
[11:09] <Kinnison> sabdfl: Looks like the gina import last night completed without throwing any exceptions. I'm looking at the sample data now
[11:09] <sabdfl> Kinnison: excellent, thanks to cprov i think
[11:10] <Kinnison> sabdfl: oh? Was he prodding my copy of gina then? I admit I was very tired last night and possibly wasn't paying enough attention
[11:11] <sabdfl> Kinnison: no, just referring to the initial code
[11:11] <Kinnison> sabdfl: aah yes; cprov (and kiko?) did a fantastic job there
[11:12] <sabdfl> spiv: good work!
[11:13] <sabdfl> limi: i don't think it works well when you send me files that you have changes but mot perged through PQM
[11:13] <sabdfl> not merged
[11:13] <sabdfl> mot perged indeed
[11:14] <Kinnison> This morning I will be mostly merging my local changes from zhongshan into a working tree to commit to rf
[11:15] <limi> sabdfl: well, arch is not giving me much joy this morning (see mail to the list) but I am splitting things up and creating the files offline here - so you should be able to just drop them in your tree and commit
[11:15] <Kinnison> limi: Arch does seem to hate you :-(
[11:15] <limi> hard when I can't run launchpad to verify my changes, though :(
[11:16] <limi> Kinnison: it's reciprocal if that's any comfort :] 
[11:16] <Kinnison> limi: *g*
[11:17] <limi> archzoom doesn't have something like the "download tarball" feature of ViewCVS?
[11:29] <limi> stub: you awake?
[11:29] <stub> Yup
[11:30] <limi> can I send you a static HTML file and we can discuss the mapping I need for the shorthand form in the portlets?
[11:31] <stub> limi: sur
[11:31] <stub> e
[11:34] <limi> stub: should be there now
[11:35] <stub> Urk
[11:35] <limi> urk?
[11:37] <stub> Just my initial impression :-)
[11:37] <limi> well, wrong style sheet ;)
[11:38] <limi> it's the only way you can pack so much information into that little space
[11:38] <stub> Yer - doesn't make me like it any better ;)
[11:39] <limi> nope
[11:39] <limi> but sometimes functionality wins over elegance ;)
[11:39] <stub> So you want a TLA for distribution, and abbreviations for severities etc.
[11:39] <limi> yes
[11:39] <limi> it's a bit ugly to do it with tal:conditions, I think
[11:40] <stub> Your just pushing the ugly somewhere you can't see ;)
[11:40] <limi> 'tis my destiny!
[11:41] <limi> but a method shortPriority() or something would be nice
[11:42] <stub> But they are colour coded as well - or do you want tal:content="structure bug/priority" to output <span class="critical-priority" title="Critital priority with gratuitous unicode character to annoy mozilla">C</span>
[11:42] <limi> any functional comments, btw? does it work with the dotted underlines and hovering descriptions?
[11:42] <limi> Mozilla is broken ;)
[11:42] <limi> their engine can't handle multiline tooltips
[11:42] <limi> age-old bug
[11:43] <stub> Mmm... unfortunately it is are primary browser target, no? Or is this a way of 'encouraging' a fix? :-)
[11:44] <limi> yes and no - it's functional this way (just not optimal), and serves as a reminder for Mozilla, yes ;)
[11:44] <limi> the bug has 160 comments in bugzilla
[11:44] <limi> they are now quarreling about what to do with chinese unicode mappings or something
[11:44] <limi> :P
[11:45] <limi> and there is a patch attached that fixes the 99% issue, but nobody has applied it
[11:45] <limi> typical Mozilla community process :] 
[11:45] <limi> anyway
[11:45] <stub> get Thom to apply it to Ubuntu's firefox release :-)
[11:46] <limi> if you have the normal getPriority() return "Critical", I do something like tal:attributes="class string:priority${getPriority}"
[11:46] <stub> Functionally, I would think the colour coding would work better if the background colour was changed rather than the text colour. Might look spastic though since we have multiple colour coded columns.
[11:46] <limi> yes, I'm looking at that - this is not the final version in any sense, design-wise
[11:46] <limi> just need to get the functionality and classes in
[11:47] <limi> and I will experiment with colors afterwards
[11:47] <limi> (I agree on background being better)
[11:47] <SteveA> stub: we can put something in the /lp:whatever tal namesapce, if that will help
[11:47] <stub> I'm actually wondering if perhaps we should have a seperate page template for each priority or status
[11:47] <limi> sounds like overkill to me
[11:49] <stub> limi: It would mean the portlet template just does <td tal:content="bug/priority/condensed" />, which calls the 'condensed' view, which could do things like use an icon instead of a letter or change the descriptive text (currently hard coded in the .py file. You are right in that it might be overkill though.
[11:50] <stub> SteveA: yer. called 'abbreviation' or something.
[11:50] <SteveA> otherwise, don't have one template for each, but define one class, and have a method for each
[11:50] <limi> it makes it hard to track where to change things if it's too decomposed
[11:50] <SteveA> note that you'd probably need to use "structure" in the tal
[11:51] <limi> if I can have getPriorityShorthand() and getSeverityShorthand() and getDistroShorthand(), I would be happy :)
[11:51] <stub> limi: fairy 'nuff
[11:51] <limi> we can refactor if it makes sense later
[11:51] <limi> premature optimization etc ;)
[11:51] <stub> limi: It would be tal:content="whatever/distro/shorthand"
[11:51] <limi> perfect
[11:52] <stub> And I've forgotten the syntax for the priorities etc. ;)
[11:55] <limi> ok, ping me when you have settled on the syntax, and I will implement the display offline here (since my arch broke again)
[11:56] <stub> It could be worse - you could be a windows developer trying to use Arch ;)
[11:56] <limi> I was ;)
[11:57] <limi> the reason (or excuse, depending on how you look at it) why I got a PB was arch
[11:58] <SteveA> install ubuntu on it ;-)
[11:58] <stub> SteveA: The dbschema.py stuff will need to grow 'shorthand'. Are we going to squeeze that into the parse-it-all-from-the-docstring approach ('''Minor (m)\n\nBlah blah blah''')
[12:00] <limi> SteveA: yes, but I tend to like having an internet connection too ;)
[12:00] <SteveA> stub: like this?
[12:00] <SteveA> 
[12:00] <SteveA>     MINOR = Item(4, '''
[12:00] <SteveA>         Minor Severity (m)
[12:00] <SteveA>         This bug does not warrant an upload just to fix it, but 
[12:00] <SteveA>         should if possible be fixed when next the maintainer does an
[12:00] <SteveA>         upload. For example, it might be a typo in a document.
[12:00] <SteveA>         ''')
[12:00] <SteveA> 
[12:01] <stub> That is what I was thinking. I was just wondering if it is getting a little - erm - obtuse?
[12:01] <SteveA> It reads to me a bit like a dictionary or glossary definition
[12:01] <SteveA> which is kind of what it is
[12:02] <SteveA> doesn't look to bad to me
[12:02] <stub> limi: Are class attributes case sensitive?
[12:02] <SteveA> what other things than "abbreviation" need to be added?
[12:02] <limi> stub: yes, they are.
[12:02] <limi> but I can change the class names if we need that
[12:03] <stub> ok - sl <span tal:attributes="string:priority-${bug/priority/shorthand" tal:content="bug/priority/shorthand"/> is fine, given we have priority 'M' and priority 'm'
[12:04] <stub> SteveA: I think that is all, unless we want a 'tooltip' to use instead of the description
[12:04] <limi> the class name should use the real name, not the shortname
[12:04] <stub> limi: The real name has spaces in it
[12:04] <limi> why?
[12:04] <limi> :] 
[12:04] <limi> "Normal Priority"?
[12:05] <limi> that makes very little sense
[12:05] <stub> Critical Severity, Normal Severity, Wishlist
[12:05] <limi> bah, that sucks :] 
[12:05] <limi> no chance of changing that?
[12:05] <stub> Hey - I'm not the UI guy :-)
[12:06] <SteveA> what about the name in the class: MINOR, WISHLIST etc.
[12:06] <SteveA> those won't contain spaces or dashes
[12:06] <SteveA> so can be used as css classes
[12:06] <limi> ah, so you have that too
[12:06] <limi> that's fine
[12:06] <SteveA> if you don't, access to them can be arranged by a little programming
[12:06] <limi> I will try to repress my hatred for UPPERCASE ;)
[12:07] <SteveA> yeah, as item.name
[12:07] <sabdfl> limi: can i use the following syntax for a stylesheet that is locally accessible?
[12:07] <stub> It will survive l10n too, if we ever go non english
[12:07] <limi> I still don't see why you want the state name to be "Normal Severity" instead of "Normal"
[12:07] <sabdfl> <style ...>@import file(filname.css);</style>
[12:07] <sabdfl> ?
[12:08] <limi> sabdfl: yes
[12:08] <sabdfl> thanks
[12:08] <limi> that's what we do in the LP templates
[12:08] <limi> so I can work even if LP is not running ,)
[12:08] <stub> limi: It doesn't, and I don't care either way. I suspect that it is actually your call unless Mark wants to use his executive veto power
[12:09] <limi> sabdfl: any objections to changing the state name to "Normal" instead of "Normal Severity"?
[12:09] <sabdfl> limi: none, go ahead
[12:09] <limi> thanks
[12:09] <sabdfl> i think this is in dbschema.py
[12:10] <limi> stub: can you do that while you're in there anyway, seeing as I can't commit at the moment?
[12:10] <sabdfl> limi: @import file() is not working
[12:10] <limi> sabdfl:     <style type="text/css" media="screen">
[12:10] <limi>     @import url(plone.css);

[12:10] <limi> make sure it is in <head>
[12:10] <stub> SteveA: Did you want to to the dbschema.py stuff for limi or shall I?
[12:11] <limi> and that the CSS is in the same dir as the template
[12:11] <limi> sabdfl: sorry, my fault - didn't see that you said file()
[12:11] <limi> it's url()
[12:11] <SteveA> stub: what exactly needs doing?  Adding the abbreviation syntax, changing some other titles to remove "severity" from them?
[12:11] <sabdfl> even for a file in the same dir as the html?
[12:12] <limi> yes, CSS has no concept of file() afaik
[12:12] <limi> it's all URLs, baby
[12:12] <stub> SteveA: Yes, Yes, making sure we have access to the item's name, and giving access through tal
[12:13] <limi> SteveA: I need getSeverity() (returns: "Normal") and getSeverityShorthand() (returns: "N")
[12:13] <limi> and getSeverityDescription() or similar for the tooltips
[12:15] <SteveA> stub: if you file a bug on me, perhaps containing some tales that presents a bug that you want to see, I'll do it today. 
[12:15] <stub> ok.
[12:27] <dilys> New bug 2101 for Launchpad/Launchpad: abbreviations and tal: access to dbschema.py classes
[12:27] <dilys> https://bugzilla.warthogs.hbd.com/bugzilla/show_bug.cgi?id=2101
[12:34] <stub> Abbreviations for distributions - is there a web interface for creating or editing a distribution, or can I happily add a column to the database without breaking anything?
[12:35] <sabdfl> abbreviations?
[12:36] <stub> Limi's mock up needs a compressed distribution name (so 'Deb' instead of Debian, 'Ubu' instead of Ubuntu)
[12:37] <sabdfl> Hmm... no thanks
[12:37] <sabdfl> distro.name
[12:37] <limi> well, there's no way you can cram "Red Hat Linux" into that portlet
[12:39] <limi> might be useful in secondary views
[12:40] <spiv> "RHL" ;)
[12:40] <spiv> Or "RHAT", their stock ticker symbol ;)
[12:41] <limi> with accompanying stock value :P
[12:41] <limi> "adjusted for linux distro inflation"
[12:42] <stub> sabdfl: If you missed that, Limi cannot cram "Red Hat Linux" into the portlet he is working on
[12:45] <sabdfl> distro.name should be redhat
[12:45] <sabdfl> distro.name should follow the same rules as every other Table.name
[12:45] <sabdfl> it's a url component, mostly, sounds perfect for what limi is doing
[12:46] <sabdfl> limi: how do I add a portlet to the left and right of a page?
[12:46] <sabdfl> also, how do I get the WHOLE page to adopt the background colour, not just the part which has text on it?
[12:46] <sabdfl> hard to explain... but I'm seeing white at the bottom of my page
[12:46] <limi> you fill the portlet slots, look at how it is done on the existing bug templates
[12:47] <limi> sabdfl: body should be a solid color - do you have an example URL?
[12:48] <limi> sabdfl: btw, I need something shorter than distro.name if the portlet is going to work - "redhat"/"Ubuntu" is too big, unless you want to go with icons or similar instead
[12:50] <limi> I'm still not clear on the context in which these portlets should be shown - I assume the main bug page should have all this info in the body text area?
[12:50] <limi> (where we have more space)
[12:51] <limi> are we going to show these portlets in other parts of Malone, in Rosetta etc, is that the idea?
[12:57] <sabdfl> limi: what about using two lines per entry in the portlet?
[12:57] <sabdfl> limi: yes, if i understand you correctly
[12:58] <limi> sabdfl: I tried two lines, but it takes up quite a lot of space, so it's not ideal if you want to conserve space
[12:58] <limi> I'll give it another shot
[12:58] <sabdfl> limi: this is for an offline page
[12:59] <sabdfl> have done <body class="section-index_html"> but that did not help
[12:59] <sabdfl> otherwise, am using the ploneCOlumns, plone and ploneCustom.css links as per the web site pages
[12:59] <sabdfl> but the bottom of the page (below all the text) is still white
[12:59] <SteveA> limi: iirc, dashes are not allowed in css classes.  Is that right?
[12:59] <limi> wrong
[01:00] <limi> underscores are not allowed
[01:00] <limi> dashes are
[01:00] <limi> :)
[01:00] <SteveA> oh, right.  
[01:00] <limi> (but they work)
[01:00] <sabdfl> so why is it "section-index_html"?
[01:00] <SteveA> what an irritating spec.
[01:00] <limi> because Zope sucks
[01:00] <limi> ;)
[01:00] <sabdfl> ok...
[01:00] <SteveA> looks like a zope2-ism
[01:00] <limi> it is
[01:01] <sabdfl> is there a better body class to use?
[01:01] <limi> just use no body class
[01:01] <SteveA> in zope2, the index page is usually named "index_html" because using dots in the names of things used to be problematic
[01:02] <limi> sabdfl: hard for me to troubleshoot it without seeing the page - can you send me a copy?
[01:02] <sabdfl> limi: willdo
[01:02] <limi> great, thanks
[01:02] <sabdfl> limi: coming soon to info@plonesolutions?
[01:03] <limi> ok
[01:03] <sabdfl> limi: sent
[01:03] <sabdfl> this needs to get out there in 20 minutes :-)
[01:03] <limi> ok, I'll have a look right away
[01:06] <sabdfl> limi: any idea?
[01:06] <limi> no mail yet
[01:07] <limi> try limi@plone.org in case the company spam filter took it out? (did you have any body text, or just an attachment?)
[01:09] <sabdfl> just an attachment
[01:10] <limi> ok, chances are it didn't make it through, then - use my personal address limi@plone.org
[01:10] <limi> sorry about that
[01:10] <sabdfl> done
[01:10] <sabdfl> no problem, seems like a reasonable filter
[01:10] <limi> we get insane amounts of spam, and spamassassin is a bit overly eager at times ;)
[01:11] <limi> there we go, thanks
[01:11] <sabdfl> is there a good way to increase the size of all the text on the pag?
[01:11] <sabdfl> e?
[01:14] <limi> p { font-size: value }
[01:14] <limi> value being 12px or 1em or 110% or what you prefer
[01:14] <sabdfl> for the whole document?
[01:14] <limi> that will touch the paragraph text
[01:15] <limi> any reason why this is XML, btw?
[01:15] <sabdfl> no idea
[01:15] <limi> it's not valid XML, which may cause problems in some browsers
[01:15] <sabdfl> shit
[01:15] <limi> is Firefox where you are seeing it?
[01:16] <sabdfl> this came in from some community guys
[01:16] <sabdfl> yes
[01:16] <limi> ok, let me clean it up
[01:16] <sabdfl> i have some text changes, will send it now
[01:16] <limi> ok
[01:16] <sabdfl> sent
[01:17] <limi> where is this thing going to end up?
[01:17] <limi> on the CD?
[01:18] <sabdfl> yes
[01:18] <limi> ok
[01:18] <sabdfl> default home page after install
[01:18] <limi> I see
[01:18] <limi> give me a few minutes, and I will clean it
[01:18] <sabdfl> i want to add two portlets, left and right, i think that will make the page look neater
[01:19] <sabdfl> will send you the txt for the portlets
[01:19] <sabdfl> it's not essential, it's optional, based on what you think looks best
[01:19] <limi> ok
[01:20] <limi> this page has no portlet structures right now, so I think I would rather go with some extra margins to make it look nice
[01:20] <limi> if time is an issue :] 
[01:23] <SteveA> are usage messages supposed to go out on stderr or stdout ?
[01:26] <limi> sabdfl: should be in your inbox with some style changes (bigger font, margins, justified text)
[01:33] <limi> sabdfl: did you get my new version?
[01:33] <sabdfl> limi: looks great
[01:33] <limi> sabdfl: should be in your inbox with some style changes (bigger font, margins, justified text)
[01:33] <limi> ok
[01:33] <sabdfl> how can i indent the bullets a little more?
[01:33] <limi> need the portlets still?
[01:33] <sabdfl> no, i think it's fine
[01:33] <limi> just add style="margin-left: 1em;" on the ul tag
[01:33] <limi> let me test a value
[01:35] <limi>     <ul style="margin-left: 3em;">	
[01:35] <limi> change the two lists to that, sabdfl
[01:36] <sabdfl> done. let's SHIP IT!
[01:36] <limi> ;)
[01:36] <limi> need anything right now, or can I do lunch? :)
[01:37] <sabdfl> lunch, thanks limi
[01:37] <limi> see you in a bit
[02:47] <carlos> hi
[02:51] <ddaa> harg, there's a missing patchlog in launchpad again!
[02:51] <Kinnison> another one?
[02:52] <ddaa> mhhh... log for-merge does not barf...
[02:53] <Kinnison> launchpad has been missing patch 422 for ages
[02:53] <ddaa> okay...
[02:53] <Kinnison> but tla missing --skip-present doesn't list it
[02:53] <ddaa> just been doing a round of tree updating and sanity checking
[02:53] <ddaa> you remember the reason?
[02:53] <Kinnison> Someone committed something manic and bad so lifeless hard-reverted it or something
[02:54] <ddaa> I still 
[02:55] <ddaa> It's bad to remove patchlogs on permanent basis... gotta annoy lifeless about that.
[03:34] <ddaa> Who's the guy in charge of unbreaking taxi.py?
[03:38] <limi> is Malone running on chinstrap or another server at the moment? I would like to test some stuff, but arch has broken my local instance
[03:39] <limi> SteveA: do you know?
[03:39] <ddaa> "arch has broken my local instance"... If it's broken, just revert to a known-good revision.
[03:40] <limi> won't work, sorry
[03:40] <limi> all I did was to star-merge this morning
[03:40] <limi> and it broke lots of files I haven't touched
[03:40] <limi> see my mail to the launchpad list
[03:40] <limi> I'm tired of dicking around with arch
[03:42] <limi> daf: you here?
[03:43] <daf> limi: pong!
[03:43] <limi> daf: is launchpad running anywhere I can reach it?
[03:43] <daf> https://rosetta.warthogs.hbd.com/
[03:44] <limi> aha, so it runs the whole thing, not just Rosetta
[03:44] <limi> thanks
[03:44] <daf> although we're supposed to be referring to it by https://mawson.ubuntu.com/ these days
[03:44] <daf> no problem
[03:44] <daf> yeah, rosetta was the name of the machine, which was confusing, so we changed it
[03:45] <limi> daf: if I send you some files, could you commit them for me?
[03:45] <limi> fixed the CSS references to plone.org etcv
[03:46] <limi> daf: what was the temporary user again? foo.bar@canonical.com? passwd?
[03:47] <limi> bingo
[03:47] <ddaa> fyi: undo does not work because spiv has changed the tagging-methog marks conflicts file as unrecognized.
[03:47] <ddaa> tla inventory -u | xargs rm
[03:47] <ddaa> to get rid of them, effectively saying "I've taken care of, or do not care about, the conflict files".
[03:47] <limi> I see why arch is needed for the Grand Scheme, I can't see why it's needed for day-to-day development
[03:51] <carlos> limi: password test
[03:51] <limi> thanks
[03:52] <limi> "A system error occured."
[03:53] <limi> (everywhere)
[03:55] <limi> sabdfl: should we include bug status (New/Open/Closed) in the portlets, or is that implicit? The old code has it, but the new spec you gave me does not.
[03:56] <limi> (ie: does it only show up if it's in "Open" status?)
[04:10] <ddaa> limi: what's interesting is that when I merge your latest public (patch-38) revision with rf (patch-586) i have no problem.
[04:10] <ddaa> However you did not tell what revision you were using in your email, so I cannot 
[04:10] <limi> it's when I star-merge I get the problem
[04:10] <ddaa> say more than worksforme.
[04:11] <limi> it worked last night before I left, and the first thing I did today was star-merge - so it can't have changed locally
[04:11] <ddaa> I understand that. Star-merge with rf from a clean tree of your patch-38 revision works okay here.
[04:12] <ddaa> So please mirror your archive if you want me to look at it.
[04:12] <ddaa> You can also tell me to get lost.
[04:13] <SteveA> daf: we must stop using the machine name "rosetta.warthogs.hbd.com" and use the new name (which I can't remember, and is not yet listed in MachineOverview)
[04:14] <SteveA> oh, mawson
[04:17] <SteveA> daf: can you make dilys subscribe to all wiki pages, and announce changes here?
[04:17] <daf> elmo hasn't changed the actual machine name yet, only the DNS
[04:18] <daf> I think perhaps dilys would get very noisy if she watched the wiki too
[04:20] <SteveA> "machine name" has many meanings.
[04:20] <SteveA> the most useful one to everyone except the admins is "the name we should call the machine when discussing it or pointing people to it"
[04:20] <daf> of course
[04:20] <SteveA> Also, for the "people<-->potemplate" list, would you like to use the zodb for that rather than use the RDB?
[04:21] <daf> shrug
[04:22] <carlos> SteveA: those are personal preferences like the language a person can translate into
[04:22] <carlos> sorry, not personal preferences but "system" ones, not temporal information that could be removed
[04:22] <SteveA> ok
[04:23] <carlos> hmm, daf: I'm confused... shouldn't it be pofile instead of potemplate?
[04:23] <daf> hmm
[04:24] <daf> our translation process is template-oriented
[04:24] <daf> i.e. you translate a template
[04:24] <carlos> but if we are adding a unique key for person,potemplate....
[04:24] <carlos> what's the utility?
[04:25] <daf> they are unique together, not separately
[04:25] <carlos> oh, right
[04:26] <carlos> daf: I don't see the utility :-)
[04:26] <daf> utility?
[04:26] <carlos> I don't see it useful
[04:26] <daf> oh, right
[04:26] <daf> it's like a To Do list
[04:27] <daf> "My Templates"
[04:27] <daf> we can use it for notifications as well
[04:27] <carlos> ok, then it's not what I was thinking on, but yes, makes sense
[04:28] <daf> good :)
[04:28] <limi> ddaa: I appreciate your help, but right now I would prefer to just work, and not fix arch :)
[05:05] <ddaa> hey lifeless
[05:34] <kiko> BradB, about sourcepackageformat, the only issue is that your change precludes changing packaging formats mid-race. ;)
[05:36] <BradB> hm
[05:37] <BradB> does that happen?
[05:37] <kiko> I"m not sure that's a practical issue as long as sabdfl signs it off consciously.
[05:37] <kiko> I don't think it should, except if it's rpm3->rpm4 for instance.
[05:37] <sabdfl> no, that would be a new sourcepackage then i think
[05:38] <sabdfl> spformat is "deb" vs "rpm"
[05:38] <sabdfl> i don't think it will be version-specific
[05:38] <sabdfl> assumption its something the latest "deb" or "rpm" tools can process
[05:38] <sabdfl> if you switch from rpm to deb, well then that's a new sourcepackage :-)
[05:39] <kiko> agreed, sounds sane.
[05:39] <limi> sabdfl: sent you today's changes in a zip, if you would like to review and commit
[05:39] <sabdfl> limi: thanks
[06:00] <sabdfl> SteveA: i just edited a page, and it hasn't shown up
[06:01] <sabdfl> i guess the workflow is working :-)
[06:01] <sabdfl> how do i publish it?
[06:01] <sabdfl> it says it's state is published
[06:01] <SteveA> sabdfl: ubuntulinux site?  no clue.  I didn't work on the workflow parts, just the authentication/launchpad integration.  BradB?
[06:02] <sabdfl> BradB: how do i publish a changed page now?
[06:02] <limi> it should be published if the state is Published
[06:02] <BradB> sabdfl: Which page?
[06:02] <sabdfl> i'm using jane's account
[06:02] <sabdfl> downlaod
[06:02] <limi> unless somebody has changed that :] 
[06:02] <sabdfl> load even
[06:02] <limi> it might be cached too
[06:03] <BradB> I see the "Set up an Ubuntu mirror" page
[06:03] <BradB> And, of course, the folder called "download"
[06:03] <BradB> as an anon user
[06:03] <kiko> BradB, style guide rocks!
[06:04] <sabdfl> this is weird
[06:04] <BradB> kiko: :)
[06:04] <BradB> sabdfl: What if you clear your cache?
[06:04] <sabdfl> http://www.ubuntulinux.org/download/
[06:04] <sabdfl> now try this:
[06:04] <sabdfl> http://www.ubuntulinux.org/download/index_html/
[06:04] <sabdfl> i see different pages
[06:05] <limi> sabdfl: looks like Apache cache
[06:05] <limi> those pages are the exact same, technically
[06:05] <BradB> eeek, scary
[06:06] <limi> if elmo turned on full-page caching, this would be the effect
[06:06] <limi> I assume it has a sensible time-out 
[06:06] <limi> so either purge the cache manually or wait a bit and see?
[06:07] <BradB> it seems to have expired now
[06:08] <limi> they are still different here
[06:08] <spiv> Same for me.
[06:08] <sabdfl> still different
[06:08] <spiv> (i.e. the pages are the same for me :)
[06:08] <limi> in any case, it's not on the Plone level
[06:09] <limi> it's something in front of it
[06:10] <BradB> when i cleared my browser cache, they became the same page
[06:11] <limi> BradB: both have tree download sites?
[06:11] <limi> they are different here, in all the browsers I run
[06:12] <sabdfl> BradB: they should have only one download site
[06:13] <BradB> yeah, they do...now it's back to http://www.ubuntulinux.org/download/index_html/ making it look as though i'm logged in as jane
[06:14] <BradB> (i went first to https link, then typed that URL above again, and see the jane thing again)
[06:48] <limi|dinner> see y'all later
[07:38] <sabdfl> bradb: hmm... that must be because I was logged in as jane when i edited it, and so that must have been cached
[08:10] <BradB> If arch says this:
[08:10] <BradB> arch_commit: unable to acquire revision lock (could not rename file.)
[08:10] <BradB>     tree: /Users/bradb/launchpad/lp
[08:10] <BradB>     revision: brad.bollenbach@canonical.com/launchpad--devel--0--patch-36
[08:11] <BradB> isn't brad.bollenbach@canonical.com/launchpad--devel--0--patch-36 the one i'm supposed to unlock?
[08:11] <Kinnison> probably
[08:11] <BradB> when i try:
[08:11] <BradB> tla lock-revision --unlock brad.bollenbach@canonical.com/launchpad--devel--0--patch-36
[08:11] <BradB> i get:
[08:11] <BradB> lock-revision: error unlocking revision brad.bollenbach@canonical.com/launchpad--devel--0--patch-36 -- lock not held
[08:12] <daf> try -b instead of -u
[08:13] <daf> (--break?)
[08:13] <BradB> that didn't give any error, so maybe it worked
[08:19] <carlos> BradB: but the -b option should only be used if the --unlock one fails
[08:22] <BradB> yeah
[08:32] <dilys> Bug 2075 resolved: Rearrange all FIXME to new  suggested XXX layout
[08:32] <dilys> https://bugzilla.warthogs.hbd.com/bugzilla/show_bug.cgi?id=2075
[08:49] <Kinnison> ddaa: ping?
[08:49] <Kinnison> oh; never mind
[08:50] <Kinnison> tla was just takes ages to do something; but it was 'cos I had hidden the gpg verifications I thought it had gone to sleep or something
[09:04] <daf> SteveA: around?
[09:05] <Kinnison> ddaa: If you want; I won't snitch :-)
[09:17] <sabdfl> how's it looking at the lunchpad?
[09:21] <Kinnison> Well, now that I have my shiny new Ubuntu desktop; I'm going back to looking at generating an spanky-new pool/dists tree from the database I built last night :-)
[09:23] <SteveA> daf: am now
[09:23] <daf> SteveA: ah, good!
[09:23] <daf> SteveA: what news on the ftesting.zcml front?
[09:23] <SteveA> page tests are looking pretty sweet
[09:24] <SteveA> I've just returned from the gym
[09:24] <daf> hurrah!
[09:24] <SteveA> I still need to track down the inconsistency with ftesting.zcml that you noted, but I want to get a bunch of page tests running pre-checkin first
[09:24] <daf> sure
[09:24] <SteveA> and also get the changes to dbobject done for stub and limi
[09:24] <daf> did you make the changes you wanted?
[09:25] <SteveA> yes.  I rewrote your shell script in python, and made it do various things to enforce some standard policies
[09:25] <daf> sounds good
[09:25] <daf> do the changes require any changes to the documentation?
[09:25] <SteveA> yes
[09:26] <daf> will you do them or shall I?
[09:26] <SteveA> your docs read very well
[09:26] <daf> thank you
[09:26] <SteveA> perhaps I can check in the system, we can discuss the changes from your docs, and you can revise your docs and send them?
[09:26] <SteveA> that way, while you're working on the docs, I can work on the dbschema changes
[09:26] <daf> let's do that
[09:26] <SteveA> ok
[09:31] <sabdfl> SteveA: there's a steveb!
[09:32] <SteveA> where?
[09:32] <BradB> SteveA: Is there a more friendly way to run the functional doctests than: bradb@ozone:~/launchpad/lp$ PYTHONPATH=~bradb/launchpad/lp/lib python lib/canonical/ftests/test_pages.py?
[09:32] <SteveA> when I used to work at lancaster university, we had a good proportion of the alphabet of steves working in various IT roles
[09:33] <SteveA> BradB: yes, and I'll be checking it in RSN
[09:33] <BradB> ok
[09:33] <SteveA> there's a better way of creating them too
[09:33] <SteveA> keep going brad!  You can easily move the tests across to the new place when I've merged to RF
[09:33] <SteveA> I'm also going to turn them on pre-checkin
[09:36] <BradB> daf, SteveA: the docs should probably include a policy for how much "function" to test in one functional doctest :)
[09:39] <daf> it doesn't, but it probably should
[09:39] <daf> I'd advocate a minimum
[09:39] <daf> except in situations where it only makes sense to test pages in a sequence
[09:41] <SteveA> we can handle sequences
[09:41] <SteveA> there's a very simple dependency system that I'll be checking in
[09:41] <SteveA> saying "you must run these tests before this one"
[09:44] <BradB> SteveA: Did you guys stumble onto sorting out whether the ftests will use a DB called launchpad_test, or a DB called launchpad_unittest?
[09:45] <BradB> ftesting.zcml says the latter in my working dir, at the moment.
[09:47] <SteveA> One idea Mark and I discussed, but didn't decide either way: the page tests will use a launchpad_pagetest DB.  It will start off with minimal content.  The page tests themselves can be run to add to it.
[09:47] <SteveA> later page tests depend on earlier ones, so you can have the test to create a bunch of stuff run before the test to browse it, for example.
[09:47] <SteveA> What do you think?
[09:47] <BradB> The fewer the DB's we have to worry about, the better, IMHO. :)
[09:48] <SteveA> so, that's a +1 ?
[09:48] <BradB> -1
[09:48] <SteveA> oh, okay
[09:48] <SteveA> so, you want to use the launchpad_test db for page tests?
[09:48] <sabdfl> i don't think we want yet another db
[09:49] <SteveA> the thing is that the page tests will need to alter the data in the database.
[09:49] <SteveA> that's intrinsic to what they do
[09:49] <sabdfl> we also don't want launchpad_test buggered up every time the page tests run
[09:49] <sabdfl> i do think there should be a unittest db that is recreated when the tests are run
[09:49] <BradB> the DB should be initialized twice, I think
[09:50] <BradB> once for UT's, once for FT's (the latter being executed in sequence so that they get the data they expect; this is okay, since these tests are meant to test taking a walkthrough of the system anyway.)
[09:51] <SteveA> now I'm getting confused...  let's clarify the language:
[09:51] <SteveA> Page tests: PTs.  these are ones created by clicking in the browser, and writing a .txt file
[09:52] <SteveA> Functional tests: FTs.  these are code that uses the whole launchpad set-up in "functional testing" mode, right down to a database, and right up to being able to render pages, or to test code that accesses the database and does stuff.
[09:53] <SteveA> Unit tests: UTs.  code that tests particular classes or functions or groups of methods.  The only UT code that requires a database is tests of code that directly accesses the database.
[09:54] <SteveA> PTs will be executed in a well-defined sequence
[09:54] <SteveA> we should have a well-known state of a database before the first PT runs
[09:55] <BradB> PT's are just FT's. I'm not sure it's useful to fray that into another term.
[09:55] <SteveA> FTs run in a well-defined sequence provided you run all of them in one go.  This is often inconvenient, especially when you're using FTs while developing functionality.
[09:55] <SteveA> BradB: I agree that technically speaking PTs are FTs.  But, I think it is very useful to distinguish them.
[09:56] <SteveA> So, there's a problem with setting up the database just once before running any FTs -- when writing a FT you should be careful to undo whatever you did, if it will make a difference to other tests.
[09:57] <BradB> Okay, maybe one should take the same care then with PT's though.
[09:57] <SteveA> no need
[09:57] <SteveA> the PTs have a notion of dependency
[09:57] <SteveA> you're explicitly depending on other PTs
[09:58] <BradB> I'm just hoping to avoid introducing new terms, that's all. :)
[09:58] <SteveA> anyhow, for now, I'll make it so that the database is dropped and re-created before running PTs
[09:58] <SteveA> we can work out a more refined way later
[09:58] <BradB> sure
[10:00] <BradB> HTTP recorders rule; the fact that current the bug-index page is broken in Malone is only a further testament to the value that will be added to the system, well, right now. :)
[10:23] <BradB> daf, SteveA: do the docs give an example of how to actually document a FDT too?
[10:32] <sabdfl>     # return the bugtracker
[10:32] <sabdfl>     return bugtracker
[10:32] <sabdfl> um, that's probably overdoing it, self
[10:35] <BradB> SteveA: Have you seen errors like: IOError: [Errno 2]  No such file or directory: 'page-test-helper.XMY3jL/watch0001.response' after ^C'ing?
[10:35] <daf> yes
[10:36] <daf> that happens when no data was recorded for a response
[10:36] <BradB> Yah, but that's the bug. :)
[10:36] <daf> are you sure you're running Launchpad
[10:36] <daf> ?
[10:36] <BradB> yes
[10:36] <BradB> Hm, looking at the other window though i now see 2004-10-13T16:33:15 ERROR SiteError http://localhost:9000/favicon.ico.
[10:37] <daf> that might be it then
[10:37] <daf> I don't think TCPWatch deals with errors very well
[10:40] <BradB> daf: Is it also normal that you might have to very slightly tweak the captured output so that the test passes? E.g. I'm getting a failure like this, right after ^C'ing and running a test for the bug index of Malone (just part of the failure output coming up):
[10:40] <BradB> Differences (unified diff with -expected +actual):
[10:40] <BradB>     @@ -246,5 +246,5 @@
[10:40] <BradB>              </tr>
[10:40] <BradB>              <tr onclick="window.location='sourcepackageassignments/2/+edit'">
[10:40] <BradB>     -          <td>mozilla-firefox-dummy</td>
[10:40] <BradB>     +       <td>mozilla-firefox-dummy</td>
[10:43] <BradB> I'm going ahead and hand-editing it
[10:44] <BradB> caution, meet wind
[10:44] <daf> that's unexpected to me
[10:45] <BradB> it looks to have been recorded incorrectly
[10:45] <daf> curious
[10:45] <daf> by tcpwatch?
[10:45] <daf> or is it being mangled by the doctest converter?
[10:52] <BradB> I'm confused.
[10:52] <BradB> It was definitely converted incorrectly
[10:53] <BradB> but then, even when i fix it to look the way it was actually recorded, it still fails on that line, as the "actual" isn't the same
[10:53] <daf> strange!
[10:54] <daf> is it something to do with tabs?
[10:55] <BradB> ew, ouch, yeah, i had just seen that now
[10:56] <BradB> because it was telling me they STILL didn't match, even though they appeared identiccal
[10:57] <SteveA> what's the deal with tabs?
[10:58] <daf> I think they confuse the page testing machinery
[10:58] <daf> either when they're converted from TCPWatch logs to doctests or when the comparison is being made, or both
[10:58] <daf> (this is judging from what BradB said)
[11:04] <SteveA> There's an option NORMALIZE_WHITESPACE
[11:04] <SteveA> we can turn that on, as it would probably help
[11:05] <daf> I'm surprised it's happening
[11:05] <daf> seems like a bug to me
[11:05] <SteveA> well, I'll turn on NORMALIZE_WHITESPACE for now
[11:05] <SteveA> but, I'll make a note that the bug exists
[11:06] <BradB> the program seems pretty sketchy about when it feels like actually working too (as opposed to, say, dying with that IOError similar to the above)
[11:07] <BradB> even when there aren't exceptions i can see in the console where zope's running
[11:07] <SteveA> maybe it is because you're on a mac?
[11:07] <SteveA> I haven't seen such errors
[11:09] <BradB> not sure yet. anyway, i'm checking in these first three for malone, since they take the first step of making sure you can get to the point of actually seeing a bug report's page.
[11:09] <SteveA> I've just submitted my merge to pqm
[11:09] <SteveA> and I mirrored first ;-)
[11:09] <BradB> what do your changes do that would conflict with mine?
[11:10] <SteveA> probably nothing
[11:10] <spiv> SteveA: You can even mirror a bit after, it spends a bit of time doing a build-config of vanilla rocketfuel before it looks at your archive ;)
[11:10] <SteveA> cool... mirror racing
[11:10] <spiv> SteveA: I've taken advantage of that once or twice ;)
[11:10] <BradB> mirrorrracer
[11:13] <Kinnison> elmo: ping?
[11:16] <BradB> SteveA: On what date will our pqm merge requests fails if one or more tests fail?
[11:17] <BradB> s/fails if/fail if/
[11:18] <SteveA> if my merge goes through, then from about 5 minutes time, merge requests will fail if any pagetests fail.
[11:18] <SteveA> just pagetests for now
[11:19] <BradB> mm, good start
[11:20] <daf> SteveA: I think having (at least some of) the unit tests added to that would be a good goal
[11:21] <SteveA> we'll have all unit tests passing, mandatorily, but not right now
[11:21] <BradB> SteveA: will a failure message include the test failure output in the email?
[11:21] <SteveA> yes
[11:21] <SteveA> as I just found out ;-)
[11:21] <BradB> hehe
[11:22] <SteveA> aw crap
[11:22] <SteveA> looks like the pqm user can't do stuff to the launchpad_test database
[11:22] <SteveA> elmo: ping?
[11:26] <SteveA> ok, I'm going to disable the testing on merge stuff until I can get an admin to look at why it isn't working
[11:27] <BradB> quick, before my merge request goes through! :P
[11:27] <elmo> ?
[11:28] <SteveA> hi elmo
[11:28] <Kinnison> elmo: zhongshan has a couple of gzips and an apt-listtemplates stuck on it. are they reamnants of anything important?
[11:28] <SteveA> if the pqm user on chinstrap tried to admin the postgres database launchpad_test, would it give an error?
[11:28] <BradB> cool, mine went through
[11:28] <elmo> Kinnison: no
[11:28] <elmo> SteveA: probably yes
[11:28] <SteveA> BradB: arse, I'll have to merge again :-)
[11:29] <SteveA> elmo: can this be arranged?
[11:29] <elmo> requiring postgres super user privs is really rather crackful - do we have to do this?
[11:30] <SteveA> it needs to be able to createdb that database
[11:30] <SteveA> at least, with the way things are currently set up in launchpad for testing
[11:31] <SteveA> otherwise, we can't do full tests on merging
[11:31] <elmo> it can createdb
[11:31] <SteveA> ok
[11:31] <SteveA> I have a script that works locally
[11:31] <elmo> and always has been able to - the point is, you guys have changed it so it now needs createuser privs either
[11:32] <elmo> which is the postgres version of superuser
[11:32] <elmo> requiring that in a testsuite strikes me as fundamentally broken
[11:32] <SteveA> but, it fails with exit code 512 when I run it as using pqm on chinstrap
[11:33] <SteveA> where do we do that?
[11:34] <SteveA> the script drops the database
[11:34] <SteveA> and creates the database
[11:34] <SteveA> and shovels stuff into it
[11:34] <elmo> gar.  steve.  it has createdb privs.  it has since the pqm user existed.  the testsuite has changed recently so that more than createdb privs is required.
[11:35] <elmo> testsuite, schema, sample-data, whatever.  I don't know exactly what changed
[11:35] <SteveA> oh
[11:35] <BradB> the language thing
[11:35] <BradB> plpython
[11:36] <BradB> it needs to be a superuser because of that now
[11:36] <sabdfl> that makefile should really just test if it's installed
[11:36] <sabdfl> then it can be done once by someone, and never again
[11:36] <sabdfl> or does dropdb destroy that?
[11:38] <BradB> since createlang takes a db arg, it probably does destroy it
[11:42] <elmo> template1=# alter user pqm with createuser;
[11:42] <elmo> ALTER USER
[11:42] <BradB> yeewhooocrap
[11:43] <BradB> using plpython isn't worth it
[11:43] <Kinnison> Theoretically if you did the createlang in template1 then the createdb would inherit the language automatically. (Unless I'm misreading these docs horribly)
[11:43] <elmo> but, for the record, I still think it shouldn't need it.. ideally it shouldn't even need createdb - it's going to make it impossible to sanely have the testsuite interact with any unrelated or production database that happens to be on the same box
[11:43] <BradB> plpython shouldn't be used.
[11:44] <BradB> stub's made it a dependency so that he can write about 10 lines of validation code in python :)
[11:44] <cprov> sabdfl: I've finished a dump with gina+nicole working on main and restricted packages on zhongshan 
[11:45] <cprov> sabdfl: zhongshan:/home/cprov/lp_dump-gina+nicole-mr.sql.gz 
[11:46] <kiko> wow
[11:46] <kiko> it's *raining*
[11:46] <SteveA> really shouldn't use these cute names for software where there's "dumping" involved.
[11:46] <Kinnison> cprov: cor; excellent.
[11:46] <cprov> sabdfl: the only problem is the missed "shortdesc" field on productseries table
[11:47] <cprov> sabdfl: and broken productreleases ...
[11:47] <cprov> Kinnison: and you any news on lucille ?
[11:48] <Kinnison> cprov: Currently hacking on a quick-and-dirty way to dump the db as a pool
[11:49] <Kinnison> cprov: I should have it going tomorrow morning :-)
[11:52] <sabdfl> cprov: hmm... i think i have a pending sql update for the shortdesc issue
[11:52] <sabdfl> and am working on productreleases tonight
[11:52] <cprov> Kinnison: nice ... dump as pool should interest me too
[11:53] <cprov> sabdfl: should I wait for you ? otherwise I will run nicole again tonight, inserting productsreleases too
[11:54] <sabdfl> cprov: is your code all committed?
[11:54] <SteveA> elmo: I think I might have hung pqm -- some kind of database deadlock
[11:55] <elmo> bah
[11:58] <cprov> sabdfl: not yet 
[11:58] <sabdfl> ok, go ahead, we'll merge later