[01:11] <ufo-> what happend to the export button in rosetta?
[01:11] <ufo-> can't seem to find it anywhere
[03:31] <salgado> morning stub
[03:35] <stub> Morning
[03:40] <salgado> stub, I created a new celebrity (shipit-admins) in my shipit branch. should I create the team myself or can you do that when doing the rollout?
[03:41] <stub> If you are a Launchpad admin, you can create it on production now.
[03:42] <salgado> no, I'm not an admin
[03:42] <stub> Can normal users create teams?
[03:42] <salgado> yes, they can
[03:42] <stub> Ok - then you can create it too ;)
[03:47] <salgado> done. :)
[06:30] <stub> spiv: Do you remember if list(foo) called len(foo) to preallocate the list size?
[06:33] <spiv> stub: Yeah, it does.
[06:33] <spiv> (And then swallows all exceptions from len(foo)!)
[06:35] <stub> The swallowing is fixed in CVS.
[06:36] <spiv> I remember reading that... that's good news :)
[08:34] <carlos> morning
[08:55] <SteveA> morning
[08:56] <SteveA> spiv: i have some great zcml evil to perform this morning.
[08:59] <spiv> SteveA: What sort of evil?
[08:59] <SteveA> i need to add a 'facet' attribute to page, subdirective-page, addform, editform directives
[08:59] <SteveA> i have worked out how to do so without duplicating code
[09:00] <SteveA> but the hack is necesarily evil
[09:05] <SteveA> i can wave my hands and imagine using events for zcml parsing in a way that would make this kind of customization easier
[09:05] <SteveA> but i don't want to change the zcml parsing upstream just for this.
[09:06] <carlos> SteveA, how is possible that the export page is failing now because a missing title? did you change anything mandatory with latest production update?
[09:07] <SteveA> i didn't do anything, and seeing as i can't merge into pqm at the moment, i don't think i can have.
[09:07] <carlos> SteveA, I mean with last week update
[09:07] <SteveA> sd-tux mentioned the problem here.  i looked at the error log page on the server, and saw an AssertionError about that pagetitle being missing.
[09:07] <carlos> anyway, it's an easy fix
[09:08] <SteveA> carlos: i haven't merged anything into pqm at all recently.
[09:08] <SteveA> i have no idea what changed to remove that pagetitle, or change the name of the template in that way.
[09:08] <carlos> me neither
[09:08] <SteveA> maybe someone else's refactoring?  and with no test to catch it, it didn't get properly refactored.
[09:09] <carlos> but one thing is true... a pagetest is missing :-(
[09:09] <carlos> yeah
[09:32] <lifeless> allo allo
[09:35] <SteveA> hello
[09:37] <lifeless> I spent the plane flight hacking up another proof-of-concept unittest plugin.
[09:37] <lifeless> as a -not-work-projet-
[09:37] <lifeless> http://www.robertcollins.net/unittest/testresources/
[09:50] <carlos> SteveA, I'm having problems with a tal macro, could you help me or are you too busy?
[09:50] <SteveA> carlos: i can help
[09:50] <carlos> SteveA, thank you
[09:50] <SteveA> lifeless: can you sort out my pqm access?
[09:51] <carlos> SteveA, I'm following the instructions we have at https://wiki.launchpad.canonical.com/PageTemplateHacking
[09:51] <carlos> but every time I try to refer to the macro, I get a KeyError with the name of the macro
[09:52] <SteveA> can i see your template code?  in Kinnison's paste bin?
[09:53] <lifeless> SteveA: on it
[09:53] <SteveA> thanks
[09:53] <carlos> SteveA, https://chinstrap.ubuntu.com/~dsilvers/paste/file5PQ1uO.html <- This is the error
[09:53] <SteveA> and the template source?
[09:53] <carlos> SteveA, https://chinstrap.ubuntu.com/~dsilvers/paste/filefQmE2H.html
[10:02] <Kinnison> Anyone object to me updating dogfood right now?
[10:02] <carlos> Kinnison, none at all
[10:03] <carlos> I mean no objections at all
[10:03] <Kinnison> cool
[10:03] <carlos> Kinnison, did you discover why your checkout was getting wrong code?
[10:03] <Kinnison> carlos: No, it's very very confusing
[10:05] <carlos> stub, hi, I was not able to test staging and this weekend my export failed because mawson ran out of disk space, so I'm doing the export now (again)
[10:05] <SteveA> carlos: i cannot see anything that would cause a problem.  we need to go into the debugger
[10:06] <Kinnison> carlos: how much more space does mawson need?
[10:06] <carlos> stub, I will tell you as soon as possible if it's ok to run that script on production
[10:06] <stub> carlos: I'll need to run it again - the database was reset yesterday
[10:06] <carlos> Kinnison, it has 55GB now, I suppose someone removed something
[10:06] <carlos> SteveA, ok
[10:06] <SteveA> carlos: can we pair-program the debugging?
[10:07] <carlos> stub, oohh...
[10:07] <SteveA> using screen or vnc?
[10:07] <carlos> stub, please, do it, I will stop the export then
[10:07] <carlos> SteveA, sure
[10:07] <carlos> I just need to setup it
[10:07] <carlos> SteveA, what do you prefer?
[10:07] <carlos> screen or vnc?
[10:07] <SteveA> let's try vnc
[10:08] <carlos> ok
[10:08] <BjornT> carlos, SteveA: just in case you didn't know, that error often happens when the template containing the macro contains some error, like a mal-formed or missing tag or something.
[10:09] <SteveA> BjornT: good point
[10:09] <carlos> hmm
[10:09] <SteveA> carlos: you can try removing surrounding parts of the page template
[10:09] <SteveA> one part at a time
[10:09] <lifeless> SteveA: fixed
[10:09] <SteveA> until the error goes away
[10:10] <BjornT> using the post-mortem debugger and print obj probably tells you something useful as well
[10:11] <lifeless> stub: ping
[10:11] <lifeless> stub: I'd like to copy my reply to you to Jim.
[10:12] <stub> lifeless: Can you paste it somewhere? My mail provider is down
[10:12] <SteveA> lifeless: anything i should know about?
[10:13] <carlos> grrr, I killed my X session
[10:14] <SteveA> carlos: you might also try removing all metal: stuff (change the namespace to xmetal) to ensure the page renders okay without it
[10:14] <SteveA> fix any problems in the page, then re-enable the metal
[10:15] <carlos> SteveA, the page has no problems at all
[10:16] <lifeless> SteveA: nope, PBCAK at this end
[10:16] <lifeless> stub: I've pasted it to you, are you happy for me to send that to jim and you ?
[10:16] <carlos> and the error is still there if I leave the macro empty
[10:16] <stub> beep beep beep.... I meant the web paste service ;)
[10:17] <lifeless> stub: ?
[10:17] <lifeless> I'm not really in the habit of pasting draft-public-mails to a web paste service.
[10:17] <stub> https://chinstrap.ubuntu.com/~dsilvers/paste/ (but I have it now - stopped beeping ;)
[10:17] <lifeless> draft-private I mean.
[10:18] <carlos> SteveA, fixed
[10:18] <SteveA> okay
[10:18] <stub> lifeless: Sure - Send a copy to Jim too
[10:18] <SteveA> what was it?
[10:19] <carlos> SteveA, just moving it inside the <div metal:fill-slot="main"> tag was enough
[10:19] <lifeless> SteveA: I've copied you
[10:19] <carlos> I suppose it's a scope problem
[10:19] <SteveA> if you can produce a minimal example, then i can probably explain what is going on, and see if it is a bug in page templates.
[10:29] <carlos> SteveA, https://chinstrap.ubuntu.com/~dsilvers/paste/fileAhWTyu.html vs. https://chinstrap.ubuntu.com/~dsilvers/paste/fileZNg3lJ.html
[10:29] <carlos> That's the only change I did
[10:30] <SteveA> ah
[10:30] <SteveA> okay, so in the example that didn't work, you weren't actually defining the macro at all
[10:30] <SteveA> because you were using the define-macro in a part of the template that was being thrown away
[10:31] <carlos> yeah, I guess
[10:32] <zyga> carlos: hello
[10:32] <carlos> zyga, hi
[10:34] <carlos> SteveA, I need to send an argument to that macro, something like a fill-slot but as a <tal:block define=....>
[10:34] <carlos> SteveA, I don't see anything about that in our FAQ
[10:34] <carlos> zyga, I suppose you want to talk about GNOME imports, right?
[10:38] <Kinnison> dogfood back running, thanks for your patience
[10:40] <SteveA> carlos: you cannot put arguments into a macro.  there are other ways to do what you need to do.
[10:40] <SteveA> for example, you can put the define="..." around the use-macro
[10:44] <zyga> carlos: exactly
[10:44] <zyga> carlos: pitti told me yesterday that they are invoked manually
[10:45] <zyga> carlos: but since 2.12 is already frozen we could trigger that import to get an up-to-date view of the situation
[10:45] <zyga> (especially since polish translators were busy enough to get near 100%) :-)
[10:51] <carlos> SteveA, something like?:
[10:51] <carlos> <metal:use
[10:51] <carlos>                               metal:use-macro="template/macros/translation-suggestions"
[10:51] <carlos>                               tal:define="submissions python:messageSet.getSuggestedSubmissions(index)">
[10:52] <SteveA> carlos: maybe.  i don't know if that will work.  i'd say, to be clear, <tal:block tal:define="submissions ..."><metal:block metal:use-macro="..."> .... </tal:block>
[10:52] <carlos> zyga, we are not imoporting any GNOME module outside the Ubuntu's packages we have that is done automatically
[10:53] <carlos> SteveA, I have that already
[10:53] <carlos> SteveA, but when I try to use submissions inside the macro, I get a KeyError
[10:54] <carlos> I tried to move the tal:define inside the meta:use-macro with the same error
[10:54] <ddaa> stub: help!
[10:54] <stub> eh?
[10:55] <ddaa> I need to get "make check" in launchpad to apply BranchDataStorage.sql before running tests
[10:55] <ddaa> since I'm in a launchpad/branches sprint now
[10:55] <ddaa> I renamed pending/BranchDataStorage.sql to patch-25-22-0.sql
[10:56] <SteveA> carlos: show me the code
[10:56] <ddaa> and tried cd database/schema ; make
[10:56] <ddaa> but then I get
[10:56] <ddaa> 08:52:14 CRITICAL patch-25-22-0.sql failed to update LaunchpadDatabaseRevision correctly
[10:56] <cprov> jamesh: ping ?
[10:57] <stub> Add to the end of the  patch  'INSERT INTO LaunchpadDatabaseRevison VALUES (25, 22, 0)'
[10:57] <carlos> SteveA, https://chinstrap.ubuntu.com/~dsilvers/paste/filexc96R7.html
[10:57] <ddaa> stub: how should I diagnose/fix handwave-away the problem?
[10:57] <stub> ddaa: Actually - pick a higher number like 78 or something - makes it easier for me if it accidently lands in rocketfuel
[10:57] <carlos> SteveA, the <tal:block condition="submissions"> raises the KeyError
[10:57] <carlos> inside the macro definition
[10:58] <stub> ddaa: (ie. call itpatch-25-90-0.sql, with 'INSERT INTO LaunchpadDatabaseRevision VALUES (25, 90, 0)' added to the end)
[10:58] <ddaa> ack
[11:00] <ddaa> hu... ha, okay, I see, the Makefile checks whether the LaunchpadDatabaseRevision table looks consistent with the names of applied patches
[11:01] <SteveA> carlos: although you can say <tal:block condition="...">, I think it reads more clearly to always use tal:condition
[11:01] <SteveA> so, use <tal:block tal:condition="...">
[11:01] <SteveA> that way, people can visually scan the source for tal:condition
[11:02] <carlos> ok
[11:02] <SteveA> so, you have:
[11:02] <SteveA>     <div metal:fill-slot="main">
[11:02] <SteveA>       <metal:define metal:define-macro="translation-suggestions">
[11:02] <SteveA>         <tal:block condition="submissions">
[11:02] <SteveA> you're defining a macro
[11:02] <SteveA> however, the macro will still be output at this point in the template
[11:03] <SteveA> there is no variable 'submissions' defined at this point
[11:03] <carlos> right
[11:04] <SteveA> you have a choice.  you can choose that you want to display the macro here, where is is defined, so you need to make 'submissions' mean something.  or say tal:condition="submissions|nothing".  OR, you can decide you don't want to show the contents of the macro here, and surround it in tal:condition="nothing"
[11:04] <carlos> how could I do to fix that?
[11:04] <carlos> hmmm
[11:04] <carlos> last option looks better
[11:04] <ddaa> stub: now, it blows up in the "* Security setup"
[11:04] <ddaa>   File "security.py", line 237, in main
[11:04] <ddaa>     assert obj_name in schema.keys(), 'Bad object name %r'%(obj_name,)
[11:04] <ddaa> AssertionError: Bad object name 'public.changeset'
[11:05] <stub> Did we drop the changeset table?
[11:05] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: debbugssync, hct enabling, and ui fixes. r=jamesh (patch-2018: mark.shuttleworth@canonical.com, scott@canonical.com)
[11:05] <dilys> Merge to thelove@canonical.com/bazaar--devo--1.5: Fixed relative dest path for "baz export" (used to panic), mini API for management of user errors (patch-63: Matthieu.Moy@imag.fr)
[11:05] <stub> Or renamed it? Same effect
[11:06] <carlos> SteveA, thanks
[11:06] <SteveA> it should become 'revision'
[11:06] <ddaa> Yes, "ALTER TABLE Changeset RENAME TO Revision;"
[11:06] <ddaa> where are the security chicken bones, so that I rethrow them?
[11:07] <stub> ddaa: Edit security.cfg and change references to public.changeset to public.revision
[11:07] <ddaa> ack
[11:07] <stub> (in database/schema)
[11:09] <lifeless> jamesh: ping
[11:19] <ddaa> stub: now, I have a sampledata conflict. I reverted my tree and db, but I'm looking for the magic word to apply the db patch to the live sampledata so I can "make newsampledata". How do apply a sql patch on the sampledata?
[11:21] <lifeless> jamesh: ping
[11:21] <stub> First build a fresh database without your patch (rename it to something not patch-25-??-?.sql to do this)
[11:22] <stub> ddaa: Then run your patch manually (psql -d launchpad_dev -f my-renamed-patch.sql)
[11:22] <stub> ddaa: Then modify your sampledata, and run 'make newsampledata' as per the documentation
[11:23] <stub> ddaa: You can then put your dbpatch back in place
[11:23] <stub> (yes - this is a pita but I'm not sure if it can be improved easily without making other compromises)
[11:45] <niran> i'm trying to get a bazaar mirror of a project that's currently developed using cvs, so if i register the project with launchpad and tell it the cvs server, will the mirror automatically be created?
[11:45] <Kinnison> spiv: ping?
[11:46] <ddaa> stub: I have another problem...
[11:46] <ddaa> When trying to update the sampledata, I get this error:
[11:46] <ddaa> psql:pending/BranchDataStorage.sql:88: ERROR:  new row for relation "branch" violates check constraint "valid_name"
[11:47] <ddaa> oh...
[11:47] <ddaa> I think I see... the update script produces a number of empty names...
[11:47] <lifeless> niran: it will get tested for importabailty yes
[11:47] <ddaa> also, we are going to need to add "@_" as valid chars for branch names
[11:47] <lifeless> niran: and you can chat with jblack or ddaa here to get the mirror checked more carefully
[11:48] <ddaa> stub: is that okay to add those chars to the valid_name regex?
[11:48] <niran> lifeless, ok, thanks
[11:48] <stub> ddaa: There is already valid_branch_name - use that. We can't add those characters to valid_name
[11:48] <stub> I must have added it after creating that basic patch.
[11:48] <ddaa> Oh, right, I missed it in my confusion :)
[11:59] <spiv> Kinnison: pong
[12:00] <Kinnison> spiv: any chance you can go over the remaining stuff in cprov's buildd branches?
[12:00] <Kinnison> spiv: We need them mergable asap
[12:01] <Kinnison> jamesh has done a bunch of review already and I believe cprov has dealt with most, if not all, of his comments
[12:03] <spiv> Kinnison: Ok.  celso.providelo@canonical.com/launchpad--buildd--0 and celso.providelo@canonical.com/launchpad--buildd-queuebuilder-archaware--0?
[12:04] <Kinnison> cprov: ?
[12:04] <cprov> Kinnison: spiv: yes, they two
[12:04] <zyga> carlos: is there any way to star translating a template without any translated messages?
[12:05] <spiv> cprov: Ok.
[12:08] <carlos> zyga, sure
[12:08] <cprov> spiv: great, special attention to the buildd which contains killing process issues, would be nice if you can suggest something based on your librarian experience
[12:08] <WaterSevenUb> zyga, 'morning zyga... how went your installation tests of breezy?:)
[12:08] <carlos> zyga, why do you think you cannot do that?
[12:08] <WaterSevenUb> zyga, in the translation side of course.
[12:09] <spiv> cprov: I'll take a look, but the librarian never spawns subprocesses, so I don't think there's any particular insight to be gained there.
[12:09] <lifeless> pqm does
[12:09] <lifeless> so whatever you do, it can feed inot PqmRobustness :0
[12:10] <cprov> spiv: right, get inside, you'll probably have something better than that in mind
[12:11] <zyga> WaterSevenUb: morning :-)
[12:11] <zyga> WaterSevenUb: pretty good but I've found one annoying issue
[12:12] <zyga> WaterSevenUb: installation path: warty->upgrade is terrible due to clashes in xfree86-common and xorg.common (rgb.txt somewhere)
[12:12] <zyga> WaterSevenUb: as for colony CD it's not perfect yet but I'm already pulling the fresh one :)
[12:12] <WaterSevenUb> zyga,and what about those issues of "Time Zone Configuratoin" and "ubuntu configuration" (post-install)... did you find them?
[12:13] <zyga> WaterSevenUb: well I'd love to but those templates just don't appear on the list
[12:13] <zyga> WaterSevenUb: no, they were not in any of the .po files breezy has - probably unmarked
[12:14] <WaterSevenUb> zyga, well... the point is .... should we annoy Carlos again to point us someone we could talk to before we report the problem somewhere?;)
[12:15] <WaterSevenUb> carlos, :)
[12:15] <zyga> WaterSevenUb: I'd rather grep the source package
[12:16] <zyga> WaterSevenUb: write a patch and then annoy the hell out of people :)
[12:16] <zyga> WaterSevenUb: similar case: 'hibernate this computer'
[12:16] <zyga> WaterSevenUb: do you know where that is?
[12:17] <WaterSevenUb> zyga, that one is on my "todoWhereTo" list too :) 
[12:18] <zyga> WaterSevenUb: translators should be able to pull a list of .pot files with *every* string extracted 
[12:18] <zyga> and a list with only marked strings extracted
[12:18] <zyga> that would solve 99% of the problems
[12:18] <zyga> if we had a clone of pitti he could spare some time to do this
[12:22] <hannosch> jordi: ayt?
[12:23] <WaterSevenUb> zyga, yeah, I agree.
[12:23] <zyga> WaterSevenUb: I'll be hacking synaptic in the meantime, minor issues though
[12:27] <zyga> ddaa: ?
[12:27] <ddaa> zyga: when running it for the first time, after updating packages, it displays a dialog "everything went fine" with a checkbox to automatically close the dialog at the end of future updates.
[12:27] <ddaa> Clicking on the checkbox closes the dialog.
[12:28] <ddaa> That is helpful, in the sense that when you click the checkbox, you want the dialog to go away pretty soon, but that is suprising because, well... checking a checkbox should not close the window!
[12:28] <ddaa> just a minor WTF, not really a blocker
[12:33] <carlos> WaterSevenUb, zyga I cannot help you too much with that kind of issues as I hadn't time too look into those things. Usually, is better if you ask the maintainer of those packages or the Ubuntu team if you don't know the exact package.
[12:35] <zyga> carlos: that's okay - we know you are really busy :)
[12:38] <carlos> zyga, thank you
[12:38] <carlos> stub, what's the ETA finish for the whitespace script on staging?
[12:38] <lifeless> 2006
[12:38] <bob2> haha
[12:39] <stub> carlos: 10.2459 done (400000 of 3904014). eta 21:24:20.121362
[12:39] <carlos> lifeless, kiko would hate you :-)
[12:41] <carlos> stub, thanks
[12:41] <SteveA> lifeless: i haven't received a reply from pqm yet
[12:42] <bob2> did you check the pqm status page?
[12:42] <zyga> carlos: why is synaptic not in rosetta?
[12:42] <lifeless> http://pqm.ubuntu.com/
[12:42] <lifeless> suggests you should not have
[12:42] <Kinnison> cprov, spiv: It seems jamesh was reviewing one of the branches, Can you coordinate and check that everything looks sane on that?
[12:42] <SteveA> lifeless: thanks
[12:43] <zyga> carlos: (as far as /pl is concerned)
[12:45] <zyga> carlos:  synaptic + rosetta = b0rked https://launchpad.net/products/synaptic/+translations
[12:58] <sivang> is launchpad still donw? I still can't open launchpad integration pages..
[12:59] <carlos> zyga, automatic imports are off for a couple of days while pitti does some changes to the buildd <-> rosetta path, it should be restored soon and synaptic will appear 
[01:00] <ddaa> stub: I'm almost there, but I still need your help for a little thingy
[01:00] <ddaa> 10:58:16 WARNING No permissions specified for ['"public"."branchmessage"', '"public"."branchsubscription"', '"public"."revisionparent"'] 
[01:00] <ddaa> what should I do to make it happy?
[01:02] <stub> ddaa: You need to add some security declarations to security.cfg for those tables. Currently only superuser accounts have access to them, which is a problem.
[01:02] <ddaa> Well, I do not think that any code is actually using those, yet...
[01:03] <ddaa> so it's not a terribly big problem, but lifeless thinks I should fix the problem right now.
[01:03] <stub> ddaa: Then ignore the warnings. If you want to silence them, stick a section 'public.branchsubscription=' in the [public]  section should do the trick
[01:03] <ddaa> nothing after the = sign?
[01:03] <stub> ddaa: Nope. That explicity says 'give no permissions'
[01:04] <ddaa> thx
[01:04] <stub> So anything attempting to access them will raise an exception, and the relevant entries added to security.cfg then.
[01:04] <ddaa> Yup, way to go.
[01:05] <ddaa> I'm not going to give speculative permissions that I will forget about and which will come back to hit me later.
[01:17] <zyga> WaterSevenUb: do you know where can we translate lanuchpad integration details?
[01:19] <WaterSevenUb> zyga, no. Let me know if you find it. I will let you know if I find it. (I'm currently busy )
[01:27] <zyga> WaterSevenUb: well... that was easier than I though ... apt-get source launchpad-integration
[01:46] <niemeyer> Morning!
[01:49] <cprov> spiv: ok, nice idea to kill process, I'm implementing it .
[01:50] <spiv> cprov: No worries.
[01:50] <spiv> cprov: I'm about to go to bed, so if you want anything else from me, send me an email and I'll do it tomorrow morning.
[01:51] <spiv> (I see that jamesh has followed up to the reviews)
[02:02] <Kinnison> thanks spiv
[02:34] <Keybuk> right
[02:34] <Keybuk> so next thing to try
[02:34] <Keybuk> cd /tmp
[02:34] <Keybuk> hct source dbus
[02:35] <niemeyer>   File "/home/niemeyer/src/mango-sorbet/pybaz/backends/forkexec.py", line 384, i    raise ExecProblem(self)
[02:35] <niemeyer> ExecProblem: baz exited with code 2 (expected exit code 0)
[02:35] <niemeyer> argv: 'get', 'TO-BE-DELETED@REALLY-DO-NOT-USE.ubuntu.com/dbus--orig--0.36.2--pat* error report
[02:35] <niemeyer> get: could not connect to the archive for (TO-BE-DELETED@REALLY-DO-NOT-USE.ubunt2005-09-05 09:34:40,373 INFO     hct.source.SourceTree.Library Removing library
[02:35] <Keybuk> ah, d'oh
[02:35] <Keybuk> baz register-archive sftp://casey.ubuntu.com/srv/launchpad.ubuntu.com/archive
[02:35] <Keybuk> and try again :p
[02:36] <Keybuk> (you may have to delete a "dbus" directory)
[02:37] <niemeyer> Hummm..
[02:37] <niemeyer> It looks like I don't have an account on that machine
[02:38] <Keybuk> elmo: ? ^
[02:38] <niemeyer> Humm.. 
[02:38] <niemeyer> I do
[02:38] <Keybuk> you should have an sftponly account
[02:39] <niemeyer> I thought the mail to db.warthogs.hdb.com would install the key everywhere, but it looks like it wasn't the case.
[02:40] <niemeyer> Ok, I've registered the frightening TO-BE-DELETED@REALLY-DO-NOT-USE.ubuntu.com archive. :)
[02:41] <Keybuk> it's supposed to, bug elmo about that <g>
[02:41] <Keybuk> yeah, Mark picked that name
[02:41] <Keybuk> it's trying to hint that you shouldn't branch off it
[02:42] <Keybuk> ah ... elmo's actually usefully given you an account
[02:42] <Keybuk> (a real one, that is)
[02:43] <niemeyer> Great!
[02:43] <niemeyer> I've installed my key manually, so everything is fine now. :)
[02:43] <niemeyer> You've mentioned that I'd need Launchpad working. What parts of hct interact with launchpad right now?
[02:44] <Keybuk> right
[02:44] <Keybuk> so mango-sorbet is my development config, which I use for testing distro-side stuff
[02:44] <Keybuk> it's got two main bits in it ...
[02:45] <Keybuk> hct -- which is the generic module bit, the utility library (hct.util) and the command-line interface (hct.cli)
[02:45] <Keybuk> sourcerer -- which is the source package importer -- it uses hct, but isn't "part" of it
[02:45] <Keybuk> it doesn't have any of the launchpad stuff
[02:45] <Keybuk> now, if you look at your launchpad config, it also has a lib/hct and lib/sourcerer which match these two
[02:46] <Keybuk> launchpad also has the server-side stuff
[02:46] <Keybuk> lib/canonical/launchpad/hctapi.py <-- the direct-to-database backend code
[02:46] <Keybuk> lib/canonical/launchpad/daemons/trebuchet.py <-- the xml-rpc server
[02:47] <niemeyer> Hummm... I see
[02:47] <Keybuk> so, how the backends work:
[02:47] <Keybuk> hct doesn't have any real concept of a source package, or how to find them. or anything like that.  It just wants a Manifest object
[02:47] <Keybuk> so it uses a backend to convert an arbitrary string into a Manifest object
[02:48] <Keybuk> so when you typed "hct source dbus", the following happened:
[02:48] <Keybuk> - it parsed it as a URL, saw there was no scheme, so added the default ("hct") scheme to it
[02:48] <Keybuk> - it looked up the handler for the "hct" scheme (hct.backends.xmlrpc)
[02:49] <Keybuk> - it called the hct.backends.xmlrpc.get_manifest() function passing "hct:///dbus" to it
[02:49] <Keybuk> - the backend made an XML-RPC call to the trebuchet server on casey
[02:49] <Keybuk>   (but with the "hct:///" bit stripped off)
[02:49] <Keybuk> - trebuchet made the same call to hctapi.get_manifest()
[02:50] <Keybuk> - which parsed the string, looked up the record in the database, and made a Manifest object
[02:50] <Keybuk> - trebuchet pickled the Manifest object by turning it into XML
[02:50] <Keybuk> - the xmlrpc backend unpickled it and returned it
[02:50] <Keybuk> -- 
[02:51] <Keybuk> there's two backends in hct -- xmlrpc which handles the "hct" scheme and is the default
[02:51] <Keybuk> and xmlfiles which handles any other scheme
[02:51] <Keybuk> so, infact, you can do something like:
[02:51] <Keybuk> hct source file:///tmp/dbus/{hct}/++manifest
[02:51] <niemeyer> (I'm following.. keep going :)
[02:52] <Keybuk> that'll parse it as a URL, see the "file" scheme, use the hct.backends.xmlfiles.get_manifest() function which treats the rest of the string as a filename, and loads it directly
[02:52] <Keybuk> the logic for all of this is in hct.url
[02:53] <Keybuk> backend modules "plug" themselves into that
[02:53] <Keybuk> the top of hct.cli.command imports the xmlfiles and xmlrpc backends to make sure they're both available
[02:53] <Keybuk> the top of sourcerer.bubblewrap imports the hctapi backend from launchpad (which shares the API)
[02:54] <Keybuk> the hctapi backend is the "best" so the "lp:///" scheme it implements becomes the default
[02:54] <Keybuk> so using the hct code within sourcerer reads and writes directly to/from the database rather than going over xml-rpc
[02:55] <niemeyer> Who's the guy parsing the manifest
[02:56] <niemeyer> (and doing something useful with it)
[02:56] <niemeyer> ?
[02:56] <Keybuk> depends what you mean by parsing
[02:56] <Keybuk> if you're reading a database object, it's hctapi
[02:56] <niemeyer> Sorry.. rephrasing
[02:56] <Keybuk> if you're reading an xml file, it's xmlfiles
[02:56] <Keybuk> if you're reading an xml-rpc response, it's xmlrpc (which actually uses xmlfiles)
[02:56] <niemeyer> Who runs baz by looking at the manifest?
[02:56] <Keybuk> hct.manifest
[02:56] <Keybuk> and, partly, hct.sourcetree
[02:58] <niemeyer> Right
[03:00] <Keybuk> that stuff needs cleaning up a bit, I want to move a lot from sourcetree into manifest itself
[03:01] <niemeyer> You mentioned that sourcerer reads/writes directly to/from the database. It looks like it asks the backend to do that, and the backend is xmlrpc. What piece am I missing?
[03:01] <Keybuk> at the top of bubblewrap it imports "canonical.launchpad.hctapi"
[03:01] <Keybuk> which is in launchpad/lib/canonical/launchpad/hctapi.py
[03:02] <Keybuk> that registers itself as a backend with a higher priority than the xmlrpc one
[03:02] <Keybuk> (and with the "lp:///" scheme)
[03:02] <niemeyer> Ah, ok
[03:02] <niemeyer> So it doesn't use xmlrpc when being run from launchpad
[03:02] <Keybuk> right
[03:02] <Keybuk> make a copy of your launchpad tree (or another get)
[03:03] <Keybuk> then:
[03:03] <Keybuk> baz switch -d launchpad scott@canonical.com--2005/launchpad--sourcerer-production--0
[03:03] <Keybuk> baz switch -d launchpad/lib/hct scott@canonical.com--2005/hct--devel
[03:03] <Keybuk> baz switch -d launchpad/lib/sourcerer scott@canonical.com--2005/sourcerer--devel
[03:03] <Keybuk> -- 
[03:04] <Keybuk> you'll then get a scripts/sourcerer-import.py
[03:04] <Keybuk> as this deals with the database directly, it cheats and uses the hct.where_am_i() function to construct a url with an "lp:///" scheme -- forcing the database backend to be used for the import
[03:07] <niemeyer> Keybuk: I'd like to build a "package" from scratch (rather than asking sourcerer to import it). Is it possible?
[03:09] <Keybuk> nope
[03:10] <Keybuk> it will be, one day, but right now existing sources is much more important
[03:13] <niemeyer> Sure.. I'd just like to go through the process to understand it.
[03:13] <niemeyer> I'll try to follow it internally
[03:18] <niemeyer> Keybuk: Is there any dummy database backend which doesn't require launchpad/a xmlrpc server to be running?
[03:19] <Keybuk> no, unless you could the xmlfiles one
[03:19] <Keybuk> which only implements the get_manifest function
[03:22] <niemeyer> Keybuk: Is hct currently storing anything in the database, besides manifests? I mean, those entries mentioned in the manifest, they're all under a baz archive, right?
[03:24] <Keybuk> right
[03:24] <Keybuk> so there's a bunch of tables
[03:24] <Keybuk> Manifest -- which is the top-level table
[03:24] <Keybuk> ManifestEntry -- which is the record for each entry in the manifest, and linked to the top-level table
[03:24] <Keybuk> ManifestAncestry -- which records the descent of each manifest, which one it was based on and which were "merged in"
[03:25] <Keybuk> each ManifestEntry references Branch and Changeset, which in turn reference ArchNamespace and ArchArchive, etc.
[03:25] <Keybuk> those are the ones we write to
[03:25] <Keybuk> we read from quite a few, the Product, ProductSeries, ProductRelease descent
[03:25] <Keybuk> and the Distribution, DistroRelease, SourcePackagePublishingHistory, SourcePackageRelease descent
[03:26] <Keybuk> but we never create those, we rely on dyson and gina (respectively) to do that
[03:27] <niemeyer> Ok
[03:29] <Kinnison> What's PQM's situation currently?
[03:29] <Keybuk> there's a spec (DevelopmentManifests) for adding another table to link Person to Manifest (so you can have "my dbus") but that's still a spec
[03:29] <Keybuk> it's almost 10 minutes work, but depends on a bunch of stuff to be useful, so it's waiting in the wings
[03:30] <cprov> Kinnison: possibly it got lost with spiv's job at 8 AM
[03:30] <niemeyer> Understood
[03:30] <cprov> lifeless: or elmo: could you kick PQM ?
[03:33] <lifeless> done
[03:43] <cprov> lifeless: thank you
[04:05] <sabdfl> mpt: are kiko and salgado around?
[04:07] <sabdfl> carlos: we have a problem
[04:07] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [r=spiv]  something/enumvalue:VALUE for page templates (patch-2344: steve.alexander@canonical.com)
[04:08] <sabdfl> we should not have the review-foobar-1 templates building up like this
[04:10] <mpt> sabdfl: No, I'm not sure where they are
[04:11] <mpt> kiko came back from Rio last night, apparently
[04:11] <sabdfl> mpt: k thanks
[04:11] <mpt> and salgado left the apartment before I did, I think, but he's not here, so I'm not sure where he is
[04:11] <sabdfl> could you ask kiko to ping me when he gets in?
[04:11] <mpt> sure
[04:11] <sabdfl> thanks
[04:11] <sabdfl> how's it going over there?
[04:11] <mpt> the usual
[04:12] <mpt> Warm, windy, and not enough RAM for baz :-)
[04:12] <sabdfl> does anybody else have the ./makepagetest thing fail? SteveA?
[04:12] <sabdfl>     assert (request and response) or not (request or response)
[04:12] <sabdfl> AssertionError
[04:12] <sabdfl> is the end of the traceback
[04:13] <SteveA> sabdfl: interesting.  there are a couple of open bugs on it.  can you put the whole traceback in https://chinstrap.ubuntu.com/~dsilvers/paste ?
[04:14] <sabdfl> https://chinstrap.ubuntu.com/~dsilvers/paste/file7ljmEl.html
[04:14] <mpt> sabdfl: kiko just called me -- he's working on the monthly report for Jane, and he'll be in the office in about 1.5 hours
[04:15] <sabdfl> mpt: thanks
[04:15] <SteveA> sabdfl: do you still have /tmp/page-test.2gvIFd around?
[04:17] <sabdfl> slinky% ls -al /tmp/page-test.2gvIFd                ~/projects/ubuntu/launchpad
[04:17] <sabdfl> total 72
[04:17] <sabdfl> drwx------   2 mark mark  4096 2005-09-05 15:12 .
[04:17] <sabdfl> drwxrwxrwt  22 root root 12288 2005-09-05 15:16 ..
[04:17] <sabdfl> -rw-r--r--   1 mark mark  1729 2005-09-05 15:11 watch0001.request
[04:17] <sabdfl> -rw-r--r--   1 mark mark 24473 2005-09-05 15:11 watch0001.response
[04:17] <SteveA> basically, there is some request data that did not meet with a response.  the zope test creation stuff is bitching about that.
[04:18] <sabdfl> -rw-r--r--   1 mark mark  1731 2005-09-05 15:12 watch0002.request
[04:18] <sabdfl> -rw-r--r--   1 mark mark 24483 2005-09-05 15:12 watch0002.response
[04:18] <sabdfl> yes
[04:18] <sabdfl> interesting
[04:18] <SteveA> okay, can you tar that directory up, and mail it to me
[04:18] <SteveA> i'll be able to use that to see what's up
[04:18] <sabdfl> i've noticed it does some weird stuff, like reordering the pages sometimes
[04:18] <sabdfl> there are quite a few dirs like that
[04:18] <sabdfl> do you want the lot?
[04:18] <SteveA> just the one called /tmp/page-test.2gvIFd
[04:19] <SteveA> that's the one that caused it to fail
[04:19] <sabdfl> it was small
[04:19] <sabdfl> you got the lot
[04:19] <sabdfl> it's been failing quite a bit for me
[04:20] <sabdfl> right now it just will not handle the sequence of pages i am trying to record
[04:23] <SteveA> there's nothing obviously suspicious about the captured data.  i'll run it through the "create a page test from this" thing, and see why it is failing.
[04:33] <SteveA> okay, reproduced the error
[04:36] <sabdfl> phew
[04:36] <sabdfl> my hair loss is accelerating
[04:36] <ddaa> SteveA: heya
[04:36] <sabdfl> not one of my attempts has worked :-)
[04:36] <SteveA> i haven't fixed the error -- just reproduced the exact problem with the data you have sent to me
[04:36] <SteveA> but, gimme a few minutes...
[04:37] <SteveA> hi ddaa
[04:37] <ddaa> SteveA: got a problem with the new Branch goo.
[04:37] <SteveA> ddaa: can we talk in 15 mins?
[04:37] <lifeless> SteveA: who else knows NotFoundError on attributes with new interfaces?
[04:37] <lifeless> SteveA: we'll ask them
[04:38] <SteveA> Bjorn may be able to help
[04:38] <lifeless> BjornT: ping
[04:38] <BjornT> hi lifeless 
[04:38] <lifeless> BjornT: we may have solved it, una momento
[04:39] <lifeless> BjornT: we have, never mind
[04:39] <BjornT> cool
[04:39] <ddaa> did not save interfaces/pyarch.py after remove the IBranch from __all__ there...
[04:42] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: Added initial support to show suggestions for multiline entries r=kiko (patch-2345: carlos.perello@canonical.com)
[04:51] <SteveA> sabdfl: want an instant totally hackish fix?
[04:52] <SteveA> python -O makepagetest.py (usual args here)
[04:58] <Keybuk> hmm, I appear to have misunderstood the way Python's logging works
[04:59] <Keybuk> if I have Logger("foo") and Logger("foo.child")
[04:59] <Keybuk> and foo's level is 30, but foo.child's is 10
[04:59] <Keybuk> I thought foo.child.debug(...) shouldn't appear, because it's parent level is higher
[05:13] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [r=spiv]  Additions and some minor bug fixes to the email interface. (patch-2346: bjorn.tillenius@canonical.com)
[05:18] <mpt> SteveA: How's menus going?
[05:19] <SteveA> mpt: very nicely!  a little later today, you'll have some page<->menu assignment to do
[05:19] <SteveA> but now, i must go out and collect a bicycle.
[05:20] <mpt> heh, cool
[05:21] <SteveA> it's taking way too long to walk to the office
[05:21] <SteveA> so, bjorn inspired me to get a bike
[05:23] <mpt> I have a bike, and I was going to spend this lunchtime getting a helmet
[05:23] <mpt> but I'm too hungry
[05:25] <SteveA> added how to work around failing makepagetest.py in https://wiki.launchpad.canonical.com/LaunchpadHackingFAQ
[05:30] <Kinnison> it's so slow
[05:33] <mpt> hmm
[05:33] <mpt> Does a team have packages?
[05:33] <mpt> Or is it people only who do?
[05:33] <mpt> (er, "only people")
[05:35] <lifeless> both
[05:36] <mpt> thanks lifeless
[05:43] <cprov> interesting -> http://www.bestpractical.com/rt/ , why not use Malone for it ? smells hypocrisy from our part :( 
[05:43] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [r=kiko]  More karma for the Malone users. (patch-2347: bjorn.tillenius@canonical.com)
[05:44] <elmo> cprov: eh?
[05:46] <Kinnison> malone isn't a trouble-ticket tracker
[05:47] <cprov> elmo: just wondering what does it have better than we can do in malone ? not criticising at all, just wondering ... 
[05:47] <elmo> cprov: RT is a dedicated issue/ticket tracker, malone is a bug tracker?
[05:47] <cprov> Kinnison: but I can't see why it can't be, what differs from one to other 
[05:48] <cprov> elmo: right, TRAC is also a issue tracker and can be used for both purposes 
[05:48] <elmo> cprov: IME bugs and ticket tracking are entirely different things
[05:48] <elmo> TRAC is a software issue tracker
[05:48] <elmo> which is just really an extension of bugs
[05:49] <elmo> (AFAIK, I have to admit I've never used TRAC)
[05:50] <vinsci> hi carlos
[05:50] <vinsci> carlos, how is your email request going?
[05:50] <cprov> elmo: perhaps I'm just missing the small features differences, but IMHO they are quite same thing .. but ok, you have a deeper look on RT docs and see, probably my fault on this
[05:51] <cprov> s/you/I'll
[05:56] <cprov> Kinnison: actually, I'm using your chair, more discret but still painful
[05:57] <Kinnison> hehe
[05:57] <Kinnison> you can sit on some screws and nails if the chair is painful
[06:01] <kiko> G
[06:01] <kiko> M
[06:01] <kiko> V!
[06:01] <Kinnison> pardon?
[06:02] <elmo> ood
[06:02] <elmo> orning
[06:02] <elmo> ietnam
[06:02] <Kinnison> oh
[06:02] <ddaa> mpt: ping
[06:03] <cprov> Kinnison: the chair is extremely confortable, stay with the fingers under the chair is what I meant.
[06:03] <Kinnison> cprov: heh
[06:03] <kiko> how is everyone?
[06:05] <cprov> kiko: morning sounds a bit late for br time and early for vietnan though, how is it going ? 
[06:06] <kiko> well, morning is whenever I arrive at the office until lunch
[06:06] <kiko> not too bad
[06:06] <kiko> have a report to send off in a second
[06:09] <mpt> ddaa: pong
[06:12] <cprov> kiko: fair enough, should I say: "this life will kill you" ehe
[06:13] <kiko> sent
[06:16] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [r=jamesh,spiv]  Fix buildd job abort action, unicode handling on slave and buildd-monitor improves. (patch-2348: celso.providelo@canonical.com)
[06:18] <cprov> ouw ... finally
[06:29] <salgado> hey SteveA. have one minute to talk about ShipitStandalone?
[06:31] <SteveA> salgado: okay, let's doit
[06:33] <salgado> so, should I just copy what we have for launchpad and change the text and the main template used?
[06:34] <SteveA> i need to catch up on where you're at with it
[06:34] <SteveA> i think i have an email explaining the work so far...
[06:34] <SteveA> yep, got it
[06:34] <SteveA> salgado: let me spend a few minutes reading this
[06:35] <salgado> sure
[06:38] <elmo> who's responsible for the CoC stuff?
[06:38] <SteveA> cprov and salgado 
[06:39] <elmo> cprov/salgado: https://launchpad.net/people/mail-slomosnail/+codesofconduct
[06:39] <cprov> elmo: yup, did anything go wrong ?
[06:39] <elmo> cprov/salgado: why is it listed in both active and inactive?
[06:39] <elmo> and is there any way for me to get the actual signed CoC?
[06:39] <cprov> elmo: no perm for me, only the own user can see it or admin
[06:40] <cprov> elmo: btw, why i'm not admin, can someone promote me ?
[06:40] <elmo> cprov: I've no idea
[06:40] <SteveA> salgado: okay, i've read it
[06:40] <Arnia> salgado: Hi. Could you tell me what the reasons were for not using DOAP in launchpad's RDF please. lifeless pointed me to you.
[06:40] <elmo> you're the launchpad team ;-P
[06:41] <cprov> elmo: how do you mean ? "get" is download the the original CoC or see if the CoC signature of this guy ? 
[06:41] <elmo> cprov: http://people.ubuntu.com/~james/x/screenshot.jpg
[06:42] <elmo> cprov: download the CoC as he signed it
[06:42] <elmo> so I can verify it for myself etc.
[06:42] <cprov> elmo: yes, unfortunatelly a poor member ..
[06:43] <cprov> elmo: lp.net/codesofconduct/console/
[06:43] <elmo> 404?
[06:43] <elmo> aha, codeofconduct
[06:44] <salgado> Arnia, the right person to answer this would be morgs, but he's on vacation now. 
[06:44] <salgado> a
[06:44] <cprov> elmo: oh, pardon me
[06:44] <salgado> Arnia, I guess we're not using DOAP because we had to extend it a little to suit our needs
[06:44] <elmo> cprov: what do I put in the From: box?
[06:45] <cprov> elmo: displayname ..
[06:45] <Arnia> salgado: Hum... ok, not sure why you'd need to create a whole new vocab for that. Any idea when morgs will be back?
[06:45] <lifeless> salgado: oops,thought it was you. my bad.
[06:45] <elmo> cprov: aha, ok, thanks
[06:45] <cprov> elmo: dsilvers has tried and found 
[06:45] <elmo> cprov: how does the code work, does it check the md5sum?
[06:46] <cprov> elmo: yes, checks the GPG signature then the content md5
[06:47] <salgado> Arnia, no, I don't know when he'll be back. sorry
[06:48] <salgado> SteveA, so, do we have a plan?
[06:49] <cprov> elmo: did I solve your problem ? in this console you can also acknowledge paper signatures and deactive signatures.
[06:49] <SteveA> salgado: 2 mins
[06:53] <elmo> cprov: yes, thanks, that's great
[06:54] <cprov> elmo: nice, thank you for testing it 
[06:54] <SteveA> salgado: okay, let's go
[06:55] <cprov> SteveA: could you include my lp user in the LaunchpadAdmins team ?
[06:55] <SteveA> cprov: no
[06:55] <SteveA> cprov: i am not an administrator of that team
[06:55] <SteveA> https://launchpad.net/people/admins
[06:56] <salgado> SteveA, for that you only need to be a member of that team
[06:56] <salgado> because that makes you a launchpad admin, which can do almost everything
[06:57] <SteveA> okay
[06:57] <SteveA> done
[06:57] <cprov> SteveA: thank you 
[06:58] <cprov> SteveA:  btw, why do we have Daniel Silverstone listed twice in the team member list /
[06:58] <SteveA> maybe he didn't merge his accounts
[06:58] <cprov> ok two lp users 
[06:58] <SteveA> Kinnison: any idea?
[06:59] <SteveA> salgado: what do we need to talk about?
[06:59] <Kinnison> SteveA: I *did* merge my accounts
[06:59] <Kinnison> SteveA: It's a holdover from a broken peoplemerge
[06:59] <Kinnison> the dsilvers account is the one I use
[07:00] <Kinnison> the kinnison one is defunct
[07:00] <salgado> SteveA, what I need to do to have a standalone /shipit, where people can login, place new orders and change existing ones
[07:00] <SteveA> do you have an apache you can configure, and a locally running launchpad running your shipit code?
[07:00] <SteveA> that would be the best, to test this out
[07:01] <salgado> I can arrange that
[07:01] <SteveA> ok
[07:02] <elmo> you guys know fix-whitespaces-in-translations.py is using > 1Gb and rising of memory on staging, right?
[07:02] <lifeless> yup
[07:05] <salgado> SteveA, so, what I need is to create something like a ShipItApplication, rooted at /shipit and then start registering the pages under it?
[07:06] <SteveA> okay, so you haven't made that yet?
[07:06] <salgado> no, I haven't
[07:06] <SteveA> yeah, that's the way to do it
[07:06] <SteveA> that wiki page described the right approach
[07:06] <SteveA> i'm not exactly sure about the details of logging in and out though
[07:07] <SteveA> but, let's get the basic thing up there
[07:07] <SteveA> in the suburl directive, you can say what the new layer is beneath that suburl
[07:07] <SteveA> so, you'll also need to define a Shipit layer
[07:07] <SteveA> and register a new main template on the shipit layer
[07:07] <salgado> right, I think I already have that
[07:08] <SteveA> you can get that far, and test it out, without using apache
[07:08] <SteveA> but it will be good to make sure it works properly with all the proxypass stuff for virtual hosting
[07:08] <salgado> right. how do I register the main template for the new layer? I couldn't find that out
[07:08] <SteveA> you know how the main template is registered generally?
[07:09] <salgado> no, I don't
[07:10] <SteveA> actually, it is registered twice!  but the registration in registry.zcml should go away
[07:10] <SteveA> so, the main template is registered like this:
[07:10] <SteveA>   <browser:page
[07:10] <SteveA>      for="*"
[07:11] <SteveA>     permission="zope.Public"
[07:11] <SteveA>      name="main_template"
[07:11] <SteveA>       template="...../main-template.pt"
[07:11] <SteveA>      />
[07:12] <SteveA> and you need to register one on the shipit layer, so in the shipit.zcml, repeat the above, but add in layer="canonical.launchpad.layers.ShipitLayer"
[07:12] <SteveA> and that says, when you have the shipit layer, use this other template
[07:12] <SteveA> and of course, it shoudl be called shipit-main-template.pt
[07:12] <salgado> ok. this is in lp/configure.zcml?
[07:12] <SteveA> the original is
[07:12] <SteveA> but you should put it all in shipit.zcml
[07:13] <SteveA> i need to remove the rest of the stuff from canonical.lp sometime
[07:13] <Kinnison> SteveA: I have a branch which is effectively an agglomoration of a bunch of otherwise trivial changes, which by the time I want to merge will probably total around a 500 line diff. Now a 500 line diff isn't trivial in-and-of itself, yet each change is trivial. Once I have it ready, would you be okay to cast your eye over it and give me an rs= ?
[07:14] <SteveA> Kinnison: when will it be?
[07:14] <Kinnison> SteveA: depending on when apt-ftparchive finishes, this evening some time
[07:14] <sabdfl> SteveA: any progress on the makepagetest front?
[07:15] <SteveA> sabdfl: yes, i replied on channel ages ago, and it is in the LaunchpadHackingFAQ
[07:15] <Kinnison> SteveA: I can drop the patch as it currently stands somewhere for you to look at before it's ready if that'll help
[07:15] <SteveA> Kinnison: okay, depending when i finish tonight, either tonight, or tomorrow morning
[07:16] <Kinnison> thanks. This is a blocker for a couple of cprov's branches but we can work around the delay
[07:16] <SteveA> salgado: let's talk through the ShipitStandalone open issues
[07:17] <elmo> person.id is stable right?
[07:17] <elmo> in the sense that y'all aren't going to run around changing it under me
[07:17] <Kinnison> I certainly hope we won't
[07:18] <SteveA> elmo: it is stable.  but it is nice if external systems don't use it.
[07:18] <SteveA> salgado: ping
[07:18] <salgado> SteveA, so, #1. can we do self.context = getUtility(ILaunchBag).user ?
[07:20] <SteveA> so, you want a person as the context
[07:20] <SteveA> i guess you can do that.  there are a couple of ways to arrange this
[07:20] <SteveA> you can do it on traversal from your ShipitApplication, so that traversal from there is by looking up views on a Person
[07:21] <sabdfl> SteveA: worked a treat
[07:21] <SteveA> but for this, it is probably better to have views be registered to IShipitApplication
[07:21] <SteveA> and have the __init__ of the view store a person that can be used.
[07:21] <salgado> yes, that's what I thought
[07:21] <SteveA> sabdfl: cool.  so, there is some weird problem with a GET request for some image, and the response to that being lost or mangled.
[07:22] <SteveA> sabdfl: using -O should still give you the important stuff.  i'll check sometime to see if the code has been improved upstream.
[07:23] <SteveA> salgado: using IShipitApplication as the context also means you won't get so many inappropriate pages being available
[07:23] <SteveA> on "How hard is it to do +register and +forgottenpassword? Should we do it?"
[07:24] <SteveA> the hard part is not the URLs, but the need for emails going to a different domain than usual
[07:24] <sd-tux> hallo carlos , I'm waiting for a fix of the bug  https://launchpad.net/malone/bugs/1978 .. is there any workaround solution to download translation..?
[07:24] <Kinnison> SteveA: If spiv gets up before I'm done, I'll ask him to review it
[07:24] <SteveA> also, as regards shipit in general, note that canonical_urls may possibly not work as you want.
[07:25] <salgado> SteveA, you mean, they'll point to launchpad.net/shipit/* instead of shipit.ubuntu.com/* ?
[07:26] <SteveA> salgado: i'm actually not sure, without checking the code.  they might well work properly.
[07:26] <SteveA> salgado: but, we need to test it using apache / virtual hosting stuff
[07:26] <SteveA> so, when you've done the new layer, new main template, and a couple of pages, let's try it out with apache in the mix
[07:26] <salgado> sure
[07:27] <SteveA> salgado: okay.  i'm going to get back into some infrastructure hacking.
[07:35] <mpt> ddaa: pongity-pong
[07:36] <ddaa> mpt: nevermind
[07:41] <kiko> BjornT, a question for you. for a zope3-widget (used on a regular view), how would I go about adding an extra attribute to the HTML? I wanted to add an onchange to it
[07:42] <SteveA> isn't there an "extras" for that, or some such keyword argument
[07:42] <kiko> yeah, but that's only for add/editforms
[07:42] <SteveA> i think there is an equivalent thing for individual widgets
[07:43] <kiko> I couldn't figure it out from reading the z3 code,but I only spend 45s looking at it
[07:44] <SteveA> 'extra' as a kw arg, dude
[07:45] <vinsci> SteveA, the LaunchpadHackingFAQ doesn't seem to mention how to get the released code.
[07:45] <SteveA> zope/src/zope/app/form/browser/widget.py, line 462
[07:46] <BjornT> kiko: yeah, SteveA is right. you can specify extra in the browser:widget directive as well.
[07:46] <kiko> browser:widget, hmm
[07:46] <kiko> vinsci, the code isn't publically available yet -- that's why :)
[07:47] <SteveA> vinsci: that's not what the LaunchpadHackingFAQ is for.  that FAQ is to help people who are coding on launchpad, and may be helpful for people coding on zope3 in general.
[07:47] <vinsci> kiko, well, SteveA said some of it was, and promised to get me the link days ago ;)
[07:47] <kiko> ah, right
[07:47] <SteveA> carlos has the release of the poparsing stuff on his todo list.  i thought it had already been done and put out there.
[07:47] <SteveA> we need to find time to release the librarian too...
[07:48] <vinsci> SteveA, ok, I've not heard back from him yet
[07:48] <sabdfl> SteveA: i'm running breezy now, and it seems as though a bunch of IntegrityError's got turned into ProgrammingError's
[07:49] <SteveA> sabdfl: these are database-level things?
[07:49] <sabdfl> SteveA: yes
[07:49] <SteveA> i guess they're being emited by postgresql 8 on breezy
[07:49] <SteveA> is it causing a problem?
[07:52] <sabdfl> i'm using 7.4
[07:52] <sabdfl> i think it is in the python module that talks to psql
[07:53] <SteveA> apt-cache policy python2.4-psycopg
[07:53] <SteveA> what does that tell you?
[07:54] <SteveA> i'm on 1.1.18-1ubuntu5
[07:54] <SteveA> which is from hoary/main
[07:55] <SteveA> there is a 1.1.19 release
[07:55] <SteveA> nice!
[07:55] <SteveA> 	* ZPsycopgDA/db.py (DB.query): applied patch from Jonathan
[07:55] <SteveA> 	Stoneman to automatically try to reconnect *once* on
[07:55] <SteveA> 	OperationalError. This fix the problem with Zope loosing the
[07:55] <SteveA> 	connection to the database when PostgreSQL is restarted.
[07:56] <SteveA> 	* Applied patch from James Henstridge to avoid segfaults in
[07:56] <SteveA> 	_curs_doall() (see psycopg mailing list "PATCH: psycopg1
[07:56] <SteveA> 	thread-safety bug in commit() and rollback()" for details.
[07:56] <SteveA> 	* James Henstridge did it again: another patch to fix segfaults
[07:56] <SteveA> 	related to calling certain methods on a closed connection object.
[07:56] <kiko> fun
[07:57] <SteveA> and the thing mark's seeing is probably
[07:57] <SteveA> 	* cursor.c (_psyco_curs_execute): applied patch submitted to
[07:57] <SteveA> 	tracker (ticket #10) to fix the problem with wrong exception types
[07:57] <SteveA> 	on PostgreSQL 8.x.
[07:57] <Keybuk> kiko: I still can't change anything on launchpad wrt hrt
[07:57] <SteveA> http://initd.org/pub/software/psycopg/ChangeLog
[07:58] <kiko> Keybuk, really? walk me through this. what URL are you on?
[07:59] <kiko> SteveA, but sabdfl said it was 7.4?
[07:59] <Keybuk> https://launchpad.net/products/hct/+bugs/1039/+index
[07:59] <Keybuk> as me
[07:59] <Keybuk> it says "You cannot change this status page: you are not the assignee or a maintainer of upstream hct."
[08:01] <kiko> Keybuk, try again.
[08:01] <kiko> swears even
[08:01] <Keybuk> that's better
[08:01] <kiko> let me know if you have any other issues.
[08:02] <Keybuk> okies
[08:03] <sabdfl> 1.1.18-1ubuntu6
[08:03] <SteveA> kiko: that's the postgresql version
[08:04] <sabdfl> SteveA: yes, i've got postgres 7.4, which is the same as hoary
[08:04] <SteveA> sabdfl:  that's interesting for two reasons.  1. why has the exception changed?  2. why isn't the most recent release in breezy?
[08:05] <SteveA> it was released on 2005-07-16
[08:05] <SteveA> i'll ask doko
[08:07] <sabdfl> hmm
[08:07] <sabdfl> i'm on pg 7.4.8
[08:07] <sabdfl> what's in hoary?
[08:07] <kiko>   Installed: 7.4.7-2ubuntu2.1
[08:07] <SteveA> 7.4.7-2ubuntu2.1 from security
[08:16] <vinsci> rosetta is rather slow
[08:17] <vinsci> seems I can translate 10 strings faster than it saves them... :-/
[08:18] <vinsci> oh well, back to kbabel for the real stuff.... :)
[08:27] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial]  Fix bug 1968, fix bug 1663, and make portlet links underlined for Kinnison (patch-2349: mpt@canonical.com)
[08:27] <zyga> launchpad-dev-people: how often is launchpad upgraded on the production system
[08:28] <vinsci> weekly, according to the faq
[08:28] <Nafallo> zyga: every tuesday IIRC
[08:28] <SteveA> weekly, unless there is something that requires an urgent upgrade
[08:28] <SteveA> like a serious bug that there's a quick fix for
[08:28] <zyga> ah, good to know
[08:29] <SteveA> special things are usually decided during the weekly developers' meeting
[08:29] <SteveA> the date and time of it is always in the channel's title
[08:29] <SteveA> the next one is on thursday, 12:00 UTC
[08:31] <zyga> SteveA: could the karma page inform the user how often the stats are updated?
[08:31] <SteveA> salgado's the karma master
[08:32] <SteveA> let's see what salgado says
[08:33] <sabdfl> kiko: did you send me a spec-system review in the end?
[08:35] <kiko> sabdfl, no, I'm still intending on doing it this week -- last week had too many last-minute requirements, I'm afraid
[08:35] <kiko> sabdfl, I'm going to start using it this week, so I'll send you feedback together perhaps
[08:35] <sabdfl> ok. i'm asking spiv for a review of the ticket tracker today
[08:36] <kiko> okay cool
[08:37] <salgado> zyga, it's updated daily, but we don't store the time when it was last updated
[08:45] <sabdfl> mpt: help
[08:45] <sabdfl> with tickets, i'd like to put the "ticket description" in a nice box. what class should I use on the div?
[08:46] <zyga> salgado: I was rather talking about 'This page is updated daily"
[08:49] <mpt> sabdfl: boardComment, perhaps?
[08:49] <mpt> or boardCommentDetails
[08:51] <salgado> zyga, sorry, I misunderstood what you said. the +karma page (where you see your activity log) is dinamically generated, so when you file a bug the correspondent action will show up there instantaneously. it's your karma counter that's updated only once a day
[08:51] <zyga> salgado: yes, yes I was talking about that
[08:52] <zyga> salgado: I've noticed it's updated daily but a message saing so would remove some confusion probably :)
[08:57] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [r=kiko]  ignore emails with an empty Return-Path header. (patch-2350: bjorn.tillenius@canonical.com)
[08:59] <kiko> any clue where carlos is?
[09:16] <niemeyer> Keybuk: What's your idea regarding the pristine assembled tree vs. the real seembled tree?
[09:17] <salgado> SteveA, ping?
[09:17] <Keybuk> means that the assembled tree can be changed, and we can work out how
[09:18] <niemeyer> Keybuk: Is this being used currently?
[09:18] <Keybuk> no
[09:19] <SteveA> salgado: 
[09:20] <niemeyer> Keybuk: Understood.. was just curious if I'd be missing something.
[09:20] <Keybuk> there's lots of things like that
[09:20] <salgado> SteveA, is it possible to override the "+login" page that's registered for all interfaces? I need to use a custom one for shipit
[09:21] <SteveA> salgado: not easily
[09:21] <niemeyer> Keybuk: You have a very nice code style.
[09:22] <SteveA> salgado: i suggest that you just write a +login view for IShipitApplication
[09:23] <SteveA> called something different
[09:23] <salgado> yes, I did that
[09:23] <salgado> the problem is when you try to access a page that's protected while you're not logged in
[09:24] <salgado> SteveA, is there anything I could do in this case?
[09:25] <SteveA> salgado: yes
[09:26] <SteveA> salgado: you need to write an UnauthorizedView for the ShipitLayer that does what you want 
[09:26] <SteveA> salgado: see lib/canonical/launchpad/webapp/login.py for the standard UnauthorizedView
[09:27] <sabdfl> kiko: any news on the langpack front?
[09:27] <salgado> SteveA, cool. thank you
[09:28] <kiko> sabdfl, I'm trying to talk to carlos, but apparently, the script is still running. I'll give carlos a call if he doesn't show up in another 30mins
[09:28] <SteveA> it's getting late in spain
[09:29] <sabdfl> and in lithuania :-)
[09:29] <kiko> I know, but I didn't get any updates today and carlos knows I want to know how this is going
[09:32] <SteveA> sabdfl: i'm on a roll.  can't stop the momentum
[09:39] <niemeyer> Keybuk: I was wondering how do you plan to do the move from baz to bzr. Do you plan to reimplement WorkingTree and Branch, or perhaps create a "pybzr" layer which behaves like pybaz?
[09:41] <niemeyer> Keybuk: (I hope you don't mind if I bother you with those questions)
[09:44] <Keybuk> WorkingTree and Branch changes
[09:44] <Keybuk> they're actually designed a bit more like bzr-style than baz-style anyway
[09:45] <niemeyer> Cool
[09:46] <Keybuk> you could theoretically implement hct on any revision control system that supported merge history
[09:48] <niemeyer> Keybuk: Interesting indeed. I'm slowly being enlightened about what hct does. Initially I belived that hct was also a kind of "meta source", which was able to build different package formats from the same information.
[09:49] <niemeyer> Keybuk: But it looks like it's more like a "meta storage" specialized in package-styled version control.
[09:50] <Keybuk> yeah, it's definitely the latter
[09:50] <niemeyer> In other words, it's able to fetch information from different sources to rebuild a "state", and make it very comfortable to hack on that information, and rebuild the planned structure.
[09:51] <Keybuk> martin calls it "meta-version-control" :p
[09:52] <niemeyer> But the different pieces must be in their correct format already (e.g. a control file must be a control file wherever it is).
[09:52] <Keybuk> right, then there's some fun bits ... because we have each state in a useful structure and have the history of that structure, we can compare two different states of the same thing
[09:52] <Keybuk> in other words, we can compare the Debian and Ubuntu source packages
[09:52] <Keybuk> or the Ubuntu and RedHat ones ... because they share the same underlying branches (even if it's just the CVS import)
[09:52] <niemeyer> Right!
[09:53] <Keybuk> as you point out, it doesn't understand debian/changelog, debian/control, RPM spec files, etc.
[09:53] <niemeyer> Uhhhh.. that's *very* interesting. 8)
[09:53] <Keybuk> you have to (as the user) maintain those yourself
[09:53] <Keybuk> we always figured we'd change the source format instead
[09:53] <niemeyer> Ok.. so I'm on the right track.
[09:54] <Keybuk> yup
[09:58] <sabdfl> mpt: the thing is, there's comments on the page
[09:58] <sabdfl> and i don't want it to look exactly like a comment
[09:58] <sabdfl> it should be thpecial
[10:00] <carlos> Hi, people, sorry, I had to go out this afternoon
[10:01] <vinsci> hi carlos
[10:01] <carlos> hi
[10:01] <carlos> kiko, new languagepack ready
[10:03] <mpt> sabdfl: Well, the description of a bug has no special background or border, and it looks fine
[10:03] <mpt> Perhaps the same would work for tickets
[10:05] <sabdfl> mpt: 
[10:05] <sabdfl> sorry
[10:05] <sabdfl> the bug has lots of other stuff on the page though
[10:05] <sabdfl> a ticket doesn't
[10:05] <kiko> carlos, that's great
[10:05] <sabdfl> it's just a ticket
[10:05] <kiko> carlos, give me good news!
[10:06] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [r=BjornT]  added a facet attribute to zcml directives for page.  refactored and added tests for the previously overridden defaultView directive. (patch-2351: steve.alexander@canonical.com)
[10:06] <carlos> kiko, as I told you by phone, the whitespace problem is still there as the db was not fixed when I generate that tarball
[10:08] <salgado> SteveA, ping?
[10:08] <kiko> carlos, the phone line was really bad, so I didn't pick up all the details. it's still running now, right? any clue on the ETA for it finishing?
[10:09] <SteveA> salgado: yes
[10:09] <salgado> SteveA, I'm having problems testing the virtual hosting stuff. can you help me?
[10:10] <carlos> kiko, it was 21 hours this morning
[10:10] <carlos> kiko, I think it should be done atm
[10:10] <kiko> hmm
[10:10] <kiko> carlos, but it's still running?
[10:10] <carlos> kiko, I don't have access to asuka to know that
[10:11] <SteveA> salgado: how can i refuse?
[10:11] <kiko> oh, right, I always forget we're handcuffed.
[10:11] <kiko> sabdfl, so the script may have finished running on staging, but we need to wait till stuart wakes up to know for sure.
[10:12] <sabdfl> how do we tell?
[10:12] <salgado> SteveA, heh. :)
[10:12] <kiko> carlos, can you email stub and ask him to run the language-pack generation script once/if the script finishes running when he's up?
[10:12] <kiko> sabdfl, via email.
[10:12] <salgado> SteveA, this is what I'm using for the vh: https://chinstrap.ubuntu.com/~dsilvers/paste/filejacmfF.html
[10:12] <sabdfl> elmo says it is finished
[10:13] <kiko> great.
[10:13] <kiko> I am unsure if we can do anything else without stuart right now, though. carlos?
[10:13] <SteveA> salgado: you don't need the ProxyPassReverse
[10:14] <salgado> SteveA, it seems to work, but when I login it doesn't work as it does when I don't use the vh
[10:14] <carlos> kiko, I don't need stuart at all for this
[10:14] <carlos> kiko, I will run the script again now
[10:14] <kiko> carlos, really? how do you run the script if you don't have access to asuka?
[10:15] <salgado> SteveA, what I tried to say is that the redirect works, but the code works only when I don't use the redirect
[10:16] <carlos> kiko, I have db access
[10:16] <SteveA> salgado: what exactly goes wrong?
[10:16] <carlos> kiko, as readonly
[10:16] <carlos> kiko, I have a modified tree that works with a readonly db
[10:16] <kiko> carlos, then you can of course ps auxw | grep to see if the script is still running, no?
[10:17] <carlos> kiko, no, I have db access using psql, nothing more
[10:17] <carlos> kiko, no ssh access
[10:17] <kiko> oh
[10:17] <kiko> I see
[10:17] <kiko> you could query a known broken string :)
[10:17] <salgado> SteveA, right after successfully logging in, I call getUtility(ILaunchBag).user. but that's always None when I use the vh.
[10:17] <carlos> well, I have tcp access to the db
[10:18] <salgado> SteveA, there's also this error: https://chinstrap.ubuntu.com/~dsilvers/paste/fileKnNKS8.html
[10:18] <SteveA> salgado: so, when you
[10:19] <SteveA> salgado: so, when you are virtual hosting via apache, then you get that ForbiddenAttribute error, and users don't get logged in?
[10:20] <salgado> SteveA, the users are getting logged in, but ILaunchBag.user is None right after you login.
[10:21] <salgado> and the ForbiddenAttribute I get only when virtual hosting, yes
[10:31] <SteveA> salgado: i need to debug this using the same setup you're using. 
[10:31] <SteveA> can you commit what you have to a branch, so i can see it
[10:32] <SteveA> and prepare for me explaining what i need to do to see this in action
[10:34] <salgado> SteveA, can I send you a diff instead of commiting?
[10:34] <SteveA> salgado: i can't look at this tonight
[10:34] <SteveA> i'm about to get some more menus infrastructure landed for mpt to look at, and then i'm going home
[10:36] <SteveA> if you commit, it means i can branch from it, and work on it
[10:37] <salgado> SteveA, okay. I'll mail you the branch name and the setup details. I guess it's not a problem if I do it later tonight, right?
[10:37] <SteveA> no problem
[10:37] <SteveA> i'll look at it tomorrow morning
[10:37] <salgado> great. :)
[10:45] <salgado> SteveA, one last question. how do I specify that I want to use a specific UnauthorizedView for the ShipItApplication?
[10:47] <SteveA> register the UnauthorizedView for the ShipitLayer
[10:48] <salgado> ah, ok
[10:50] <kiko> cprov, shouldn't https://launchpad.net/malone/bugs/2070 be private?
[10:51] <cprov> kiko: not really, since it doesn't expose the code IMO. do you think it should ? I've just changed it after a chat with dsilvers
[10:52] <kiko> I don't see any advantage to it being open, jus that
[10:52] <kiko> it wouldn't make sense beyond confuse any end-user
[10:52] <kiko> but it's a nit
[10:56] <cprov> kiko: you're right, sorry for the noise, I'll change it back to private. What about the other ? since the component buildd is a bit restrict, should I keep also them private ?
[10:58] <cprov> question about the tester app
[10:58] <cprov> python test.py -f --test='builder.txt'
[10:58] <cprov> Ran 2 tests in 2.806s
[10:59] <cprov> $find -name 'builder.txt'
[10:59] <cprov> ./lib/canonical/launchpad/doc/builder.txt
[10:59] <ondrej> hi people, I still receive bugs from malone to gnome-desktop-environment (since I am DD for this package).  I tend to ignore these bug reports (since they are ubuntu specific), could we do something to reassign gnome-desktop to MOTU?
[10:59] <cprov> kiko: any clue ?  why 2 if I've specified only one, and there is, of course, only one file builder.txt
[11:01] <salgado> cprov, ./launchpad/doc/buildd-queuebuilder.txt
[11:01] <salgado> this is the second one
[11:01] <salgado> the test stuff uses re.search('testname') to find the tests it needs to run
[11:02] <salgado> python test.py -f --test='^builder.txt' should run only the builder.txt, I guess
[11:02] <cprov> salgado: erm, ok how the regexp to enforce I want just one
[11:02] <cprov> salgado: yup, thank you
[11:03] <cprov> salgado: works ;)
[11:03] <salgado> :)
[11:05] <kiko> heh
[11:06] <cprov> kiko: what was the decision about buildd bugs ?
[11:06] <kiko> no official recommendation, just that one consideration
[11:09] <SteveA> kiko: is there a reviewers' meeting tomorrow?
[11:11] <kiko> it's on wednesdays, SteveA 
[11:13] <SteveA> goddamnit, my clock says "tuesday"!
[11:13] <kiko> go to bed
[11:13] <salgado> SteveA, I added this (https://chinstrap.ubuntu.com/~dsilvers/paste/filerWfvNv.html) on browser-exceptions.zcml, but it's not using that view for ShipItLayer. am I missing something?
[11:13] <SteveA> when dylis says so...
[11:14] <SteveA> salgado: that looks correct
[11:14] <SteveA> so, there is something else wrong
[11:14] <SteveA> i'll look at your branch tomorrow
[11:16] <salgado> ok. I'll leave it as it is now
[11:16] <SteveA> it is strange.  we did this before, for the old ubuntu forgotten passwords pages
[11:32] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [r=BjornT, trivial]  added facet attribute for browser:addform and browser:editform, and for the outer browser:pages directives.  with tests. (patch-2352: steve.alexander@canonical.com)
[11:32] <SteveA> yay
[11:32] <kiko> does anyone know if it is possible to create a new account on staging?
[11:32] <kiko> SteveA, perhaps?
[11:32] <SteveA> i know next to nothing about staging
[11:32] <SteveA> stub will be around in a while
[11:32] <kiko> :-(
[11:32] <kiko> ok
[11:32] <carlos> kiko, you can create anything you want there
[11:33] <carlos> kiko, but it will be lost after the sync with production happens
[11:34] <kiko> carlos, new users need email to be able to validate accounts. are you sure you can create new users there?
[11:34] <carlos> kiko, hmm
[11:35] <carlos> good point