[01:56] <stub> kiko: morning
[01:57] <kiko> stub, very funny. ;)
[02:00] <sabdfl> stub: night!
[02:37] <lifeless> debonzi: what did you want yesterday?
[02:37] <debonzi> I had a little problem with my archive, but its already solved.. thanks
[02:37] <lifeless> ok
[03:34] <cprov> kiko: stub : here is the place :)
[03:35] <kiko> aha!
[03:35] <cprov> kiko: no, i think, why do you want displayname in lowercase ?
[03:37] <kiko> because of the URL section
[03:39] <cprov> kiko: in this case it loses its functionality
[03:39] <stub> cprov: Not displayname, name
[03:40] <kiko> I said *displayname*.
[03:41] <cprov> kiko: yep, ok names are suposed to be unique on distro and distrorelease, but in Person, we don't have nothing similar to use in as traverse
[03:41] <kiko> yes, in person it's freaking me out too -- there's a bug open though <wink>
[03:41] <stub> kiko: Hey - I can add it anytime :-)
[03:42] <kiko> stub, well, nobody seems to agree with need a login name, incredibly.
[03:43] <stub> I do
[03:43] <cprov> three lost on it 
[03:44] <kiko> they seem to think that an email address will work for everything
[03:44] <cprov> kiko: stub: having you seen the auth_component for launchpad written for spiv ?
[03:44] <stub> We can't use displayname in many places, because it is not unique. I'll have to display the integer id or something, or flag one of the email addresses as 'primary' and display that for all the spam robots :-)
[03:44] <cprov> how does it works ?
[03:44] <kiko> no, I haven't looked at it yet. does it actually work?
[03:45] <kiko> stub, one sec. we've shifted to discussing Person now, right?
[03:45] <stub> kiko: Yup
[03:45] <kiko> well
[03:46] <cprov> kiko: I haven't seen too, just the large commit on arch at morning 
[03:46] <kiko> stub, the issue there seems to be using the email address as a traversal string. do we want to do that, or do we want a login/nick?
[03:47] <stub> Even if auth is done using email address, I think we still need Person.name for traversal and display-name-when-we-need-it-to-be-unambiguous
[03:47] <stub> Traversal using email is evil, as email addresses fluctuate
[03:48] <stub> (well... so might login name, but that would be a conscious decision if we allow it)
[03:48] <stub> Also we will get complaints because the email addresses will be harvested by spambots
[03:49] <stub> Or people don't actually want them public, etc.
[03:49] <stub> And we have multiple email addresses per person atm, and no way to flag one as primary.
[03:49] <kiko> stub, well, we use distro release as a traversal string, and it can change too...
[03:50] <kiko> right.
[03:53] <stub> email address changes are forced on people. Changing distroname or login name is a choice, and people do it because they *want* those changes including the fallout to URL's.
[03:53] <kiko> btw, where I said "displayname should be lowercase" I mean "traversable names".
[03:54] <kiko> stub, well, not all email address changes are forced on people, though they may be (unemployed, server died, etc).
[03:54] <stub> kiko: I assumed that :-)
[03:54] <kiko> luckily it appears you are on a different sort of drug than I am
[03:55] <kiko> the kind I'm on makes me make mad statements in public
[03:55] <kiko> okay
[03:55] <kiko> well, if we want a unique name/nick.. if we want this to be painless we just add a field to the table and an editable field to edit personal info form and go on with life.
[03:56] <kiko> or we battle it out with steve and the gang about soyuz implying a globally unique nick/login space for OSS. :)
[03:57] <cprov> kiko: In that last way we'll experiment the live sensation to lost our canonical emails :)
[03:58] <kiko> yes, it will be thrilling to discover 
[03:58] <stub> Ok - I have to fly. Feel free to email this discussion to launchpad@, or I will do something similar this arvo.
[03:59] <kiko> arvo?
[03:59] <stub> afternoon
[03:59] <stub> occa speak
[03:59] <kiko> that's a valid abbreviation if you're strine?
[07:04] <lifeless> stub: ready?
[07:18] <stub> lifeless: ping
[07:18] <lifeless> hey
[07:18] <lifeless> so, ready ?
[07:18] <stub> Sure :-)
[07:18] <lifeless> I've tagged the launchpad & pyarch branches
[07:18] <lifeless> there is a last schema upate in your mail from me.
[07:19] <stub> Has emperor changed its domain recently?
[07:19] <lifeless> not that I know of
[07:19] <lifeless> butmaybe.
[07:20] <lifeless> can you the dbnazi thing on that schema update. (See how easily *I* let go ?)
[07:20] <stub> urgh... no... it is *not* .com.au :-P
[07:20] <lifeless> and then I can merge it into the production branch.
[07:21] <stub> debonzi patch == the one you emailed to launchpad@ ?
[07:21] <stub> oh... dbnazi... not debonzi
[07:24] <lifeless> both.
[07:29] <stub> Hopefully this will be the last time I have to watch arch crawl on this osx box - the new laptop arrived earlier :-)
[07:29] <lifeless> its the OS dude.
[07:29] <lifeless> :)
[07:31] <stub> Ok - its with PQM
[07:33] <lifeless> ok, the new production branch is launchpad--production--1.1
[07:34] <stub> I guess I need a new archive for emperor (hmm... or just use scp for now...)
[07:34] <lifeless> new archive ?
[07:34] <lifeless> why ?
[07:34] <stub> Oh... I can just check out rocketfuel. I just can't make changes.
[07:34] <lifeless> right.
[07:35] <lifeless> which is kinda the point of production-lockdown anyway :)
[07:35] <lifeless> I'm thinking we might want one branch per machine-type
[07:35] <lifeless> production-emperor
[07:35] <lifeless> production-appserver
[07:35] <lifeless> production-buildbot 
[07:35] <lifeless> etc
[07:36] <stub> I suspect that might be your call - I'm still in single thread CVS land where we just tag the trunk at certain points :-)
[07:39] <lifeless> I'm still dithering
[07:39] <lifeless> anyway, pqm is chundering away
[07:40] <stub> Hmm... 1.1? I thought we would be 2.0 now?
[07:40] <lifeless> I thought about going 1, 2, 3, decided that we have major and minor code drops
[07:40] <stub> (the database schema is already at path-1-11-0
[07:40] <stub> (the database schema is already at patch-1-11-0
[07:41] <stub> Suppose there is no real reason the numbers need to be in sync
[07:48] <lifeless> on, pqm luv done. did it work ?
[07:49] <stub> Hmm... no such version
[07:50] <stub> 'tla versions rocketfuel@canonical.com/launchpad--production' only gives  me 'launchpad--production--1'
[07:50] <lifeless> try now
[07:50] <lifeless> I think your merge failed
[07:51] <stub> got it
[07:51] <lifeless> bah, hang  on
[07:59] <lifeless> ok, production--1.1.--patch-1 is now available
[07:59] <lifeless> which has the patch merged down.
[08:02] <stub> No patch-1-11-0.sql
[08:04] <stub> tla revisions just gives me 'base-0'
[08:10] <stub> lifeless: If you have to go at 4:30, we are running out of time. I can copy the patches across and run them now, or we can do this later. I'm around all evening and tomorrow if we want to reschedule.
[08:15] <lifeless> stub. try now..
[08:15] <stub> ok - patch-1 now
[08:16] <stub> all there
[08:19] <stub> last backup is 18 minutes old
[08:20] <stub> Any last words?
[08:20] <lifeless> nope
[08:21] <stub> I'm going to have to kill connections, or would you rather shutdown stuff gracefully?
[08:21] <lifeless> doing it now
[08:22] <lifeless> done
[08:27] <stub> error on first patch - shortdesc contains nulls :-)
[08:28] <lifeless> hahaha
[08:28] <lifeless> mark is nuts thinking anyone will put sane data in two description fields
[08:29] <stub> Who bug #1 in there??? Its an issue dammit!
[08:32] <stub> ok - patches all run... I'll reenable remote access now
[08:34] <stub> lifeless: database all patched and live again
[08:34] <lifeless> ok, trying
[08:36] <lifeless> tla delta is a wonderful tool btw 
[08:37] <stub> I plan to actually read the tla docs one day ;)
[08:38] <lifeless> PYTHONPATH=/home/launchpad/dists/launchpad/sourcecode/zope/src:/home/launchpad/dists/launchpad/lib: python2.3 \
[08:38] <lifeless>             runlaunchpad.py -C launchpad.conf
[08:38] <lifeless> /home/launchpad/dists/launchpad/lib/zope/configuration/fields.py:383: UserWarning: You did not specify an i18n translation domain for the 'title' field in /home/launchpad/dists/launchpad/lib/canonical/lp/configure.zcml
[08:38] <lifeless>   warnings.warn(
[08:39] <lifeless> (oh, it makes two-commands to pick up the latest code, moving across branches, and preserving the dbhost changes we did for the connect strings.
[08:40] <stub> Those warnings should go in bugzilla. Do we have a zope version mismatch? 
[08:40] <stub> I'll do it now
[08:40] <lifeless> no, its running the latest
[08:40] <lifeless> I've got a system error accessing the project page
[08:42] <lifeless> whats the debugging url again ?
[08:42] <stub> do: /++skin++Debug/whatever/the/url/was
[08:43] <lifeless> #
[08:43] <lifeless> # Module zope.app.traversing.adapters, line 114, in traverse
[08:43] <lifeless> curr = traversePathElement(curr, name, path, request=request)
[08:43] <lifeless> # Module zope.app.traversing.adapters, line 167, in traversePathElement
[08:43] <lifeless> next_item = traversable.traverse(nm, further_path)
[08:43] <lifeless> # Module zope.app.traversing.adapters, line 52, in traverse
[08:43] <lifeless> raise NotFoundError(subject, name)
[08:43] <lifeless> __traceback_info__: (<canonical.soyuz.sql.SoyuzProduct object at 0x41a739ac>, 'name', [] )
[08:43] <lifeless>        'request': <zope.publisher.browser.BrowserRequest instance URL=https://macquarie.warthogs.hbd.com/launchpad/++skin++Debug/soyuz/projects/xfs/index.html>,
[08:43] <lifeless> I'm going to have to go.
[08:43] <lifeless> Its not likely to be a db problem AIUI, so, I'll just leave it down for now.
[08:44] <lifeless> as only mark and I have the password to macquarie, I don't have a problem doing this :).
[08:44] <stub> Cool. Have fun at the postgres sig. I'll be online later if you want to do this tonight, or tomorrow.
[08:45] <lifeless> great, thanks
[08:45] <lifeless> I'll try and track it down on my laptop during the SIG.
[12:23] <carlos> morning
[01:30] <stub> Morning
[02:04] <carlos> stub: please, could you add to rocketfuel the schema changes I sent last week?
[02:08] <SteveA> hello
[02:16] <stub> carlos: ok. I'll do that now.
[02:16] <spiv> stub: I'm setting up the authserver on macquarie, what DSN do I give to connect to the DB on emperor?
[02:17] <carlos> stub: thanks
[02:17] <spiv> And/or do you need to configure it to allow andrew@macquarie to connect? :)
[02:17] <stub> It shouldn't be connecting as andrew...
[02:18] <spiv> Hmm, I don't think there's been an 'authserver' user created for it... should there be?
[02:19] <stub> dbname=launchpad_prod, user=importd
[02:19] <stub> The users and permissions stuff needs to be audited and sorted. Connect as that user for now.
[02:20] <spiv> Ok.
[02:38] <stub> carlos: its with pqm now
[02:39] <carlos> stub: thanks
[02:40] <spiv> SteveA: production authserver is now running on macquarie.
[02:48] <kiko> morning
[02:49] <lifeless> spiv: how do I hook into the auth stuff to protect the sourcesource pages?
[02:51] <lifeless> kiko: just so you know, the SoyuzProjects & Products are /very much/ in use.
[02:51] <lifeless> (If it wasn't you that added the 'not in use' comment... sorry)
[02:52] <lifeless> I just had to track down a bit-rott problem because something got moved :[
[02:53] <kiko> lifeless, it wasn't me, but thanks for the heads up.
[02:53] <kiko> lifeless, maybe you could write to LP describing what you're using Projects/Products for, because it really confuses us
[02:53] <lifeless> they are the only bits of launchpad actually in production.
[02:53] <lifeless> the cscvs source code imports
[02:54] <kiko> yeah, I expected something like that, because to us they make no sense and wouldn't even fit into soyuz :)
[02:54] <lifeless> well they do fit into soyuz.
[02:54] <lifeless> and rosetta.
[02:54] <lifeless> so: soyuz source packages are in arch, in theory.
[02:54] <kiko> soyuz (the web app) doesn't know about projects/products (except via a thin line spiv pointed out to me the other day)
[02:55] <lifeless> so to add a source package you need the code imported, which means you need the project + product created and a sync created for that.
[02:55] <kiko> they are an integral part of rosetta though, which is why spiv brought it up in the first place (soyuz-rosetta integration)
[02:55] <kiko> lifeless, aham. interesting.
[02:55] <lifeless> and likewise, rosetta imports from tla, so to translate a project, you have to import it.
[02:55] <kiko> lifeless, so should we link from source packages to products and from there to projects?
[02:56] <kiko> don't shrug
[02:56] <lifeless> if the model allows a source package to exist /without/ a product... then I wouldn't worry.
[02:56] <kiko> but if it *does* have a product?
[02:56] <lifeless> but I was of the understanding that you could not create a source package without linking it to a product.
[02:56] <kiko> well
[02:56] <kiko> there are source package uploads that are distro/release-specific
[02:56] <kiko> but these don't exist IRL yet.
[02:58] <lifeless> I'm shrugging because I'm not the architect of the soyuz component - I can only relay what I recall from various discussions, the rdb model etc.
[02:59] <kiko> I know; still, your opinion is valid.
[02:59] <lifeless> yay, production launchpad happy again
[03:00] <kiko> sorry bout that. if you could get that mail to lp i'd appreciate it
[03:01] <lifeless> I'll drop a note there..
[03:01] <lifeless> but I've said all I know already :)
[03:01] <kiko> more than we know so far :)
[03:01] <kiko> lifeless, if you can, point out where the code that uses that lives
[03:02] <lifeless> sql.py database.py configure.zcml : where all of your other code is
[03:02] <kiko> I mean *callsites* that use that code
[03:03] <kiko> you don't use python -i, right? <wink>
[03:03] <lifeless> sorry, you've lost me.
[03:03] <kiko> well
[03:03] <kiko> code that uses the sql/database code?
[03:03] <lifeless> templates/project* product* sync*
[03:04] <kiko> lifeless, and how does your code interface with that? or is it web-driven?
[03:04] <lifeless> web driven
[03:04] <kiko> ah!
[03:04] <kiko> using http posting or do you actually use the web app?
[03:05] <lifeless> the arch module - lib.canonical.arch - has the machinery to model stuff and do heavy lifting
[03:05] <lifeless> the conversions etc all live in buildbot and cscvs which are different tla code trees.
[03:05] <lifeless> and you should be able to be completely ignorant off.
[03:06] <lifeless> But I had to write a think web UI to let the relevant data be entered in, so that as folk start doing rosetta & soyuz & malone things, it all hangs together.
[03:06] <lifeless> i.e. bugs in malone should track back to a project.
[03:06] <kiko> but you add projects *manually* or in an automated fashion?
[03:06] <lifeless> I should be able to get a list of all the source packages that exist for 'gnome', for instance. or bugs.. or translations.
[03:06] <lifeless> kiko: manually, Mark wants them reviewed.
[03:07] <lifeless> https://www.warthogs.hbd.com/ProjectProductSetup
[03:07] <kiko> aha.
[03:07] <kiko> and the buildbox/cscvs bits *use* these bits of the database directly, or is it more complex?
[03:08] <lifeless> cscvs uses the arch objects... and the arch objects use the database.
[03:08] <lifeless> buildbot is the scheduling and RPC components.
[03:08] <lifeless> + a readonly reporting UI.
[03:09] <kiko> it would really be nice if you could be there next week because this is all really fuzzy to us.
[03:09] <lifeless> one of Mark's goals with this whole system is the pervasive annotation and linking of all these things.
[03:09] <lifeless> were next week?
[03:11] <kiko> in london?
[03:12] <spiv> lifeless: There's a Soyuz sprint in London next week :)
[03:17] <lifeless> on, hahahha no,
[03:24] <lifeless> y just spent too long there (1 month sprint)
[03:30] <daf> that's not a sprint, it's a marathon
[03:30] <carlos> daf: :-P
[03:44] <SteveA> spiv: thanks
[03:45] <SteveA> spiv: Can you mail roche giving the hostname and port that the plone server will use to talk to the auth server on macquarie?
[03:50] <SteveA> spiv: I just talked to roche.  They started testing using the demo auth server on rosetta.  They made one xmlrpc query to create a user, but received an error.
[03:50] <SteveA> spiv: Did you test the server yourself with python's xml-rpc library?
[03:51] <spiv> SteveA: Yes.
[03:51] <spiv> (hello :)
[03:51] <SteveA> hello
[03:52] <SteveA> ok, roche is in a meeting with a client for about 1/2 an hour or so
[03:52] <SteveA> but he will be in touch soon to sort out why he is getting some error
[03:52] <SteveA> can you look at the logs on rosetta?
[03:52] <spiv> Yep.
[03:52] <SteveA> and see why he might have got that error
[03:53] <spiv> exceptions.TypeError: Non-hexadecimal digit found
[03:53] <spiv> (doing sshaDigestedPassword.decode('hex'))
[03:53] <SteveA> so, he is passing you the literal string 'hex' 
[03:53] <SteveA> ?
[03:54] <spiv> No, that's the way I'm invoking the hex decoding, i.e.
[03:54] <spiv> >>> 'abcdef'.decode('hex')
[03:54] <spiv> '\xab\xcd\xef'
[03:54] <SteveA> do you know what string he passed in?
[03:54] <SteveA> (I mean, are you logging xmlrpc queries?)
[03:54] <spiv> I'm not but I think I should be :)
[03:55] <spiv> I'll add that feature now.
[03:55] <SteveA> ok
[03:55] <SteveA> Maybe ask roche to join a chat so you can debug this together
[03:55] <spiv> (But it appears he's passing something other than a string of hex characters as the password)
[04:01] <SteveA> I'm off to catch a plane.  Back tomorrow.
[04:17] <lulu> spiv:ping
[04:17] <spiv> lulu: Pong.
[04:18] <lulu> spiv:have you got Roche's mail on the error?
[04:18] <spiv> I don't have working email :(
[04:18] <spiv> My email server is down :(
[04:18] <lulu> ok - i'll do an IRC message to you - standby
[04:18] <spiv> Thanks.
[04:21] <lulu> spiv:done
[04:21] <spiv> lulu: Got it.
[04:22] <spiv> The problem is he's sending the SSHA digest as a base64-encoded string, and I'm expecting a hexidecimal-encoded string.
[04:23] <lulu> ok - I'll pass that on.
[04:23] <lulu> snything more?
[04:26] <spiv> No, it looks good otherwise :)
[04:26] <spiv> I'll mail him too, so that he's got an email address for me that works.
[04:29] <kiko> spiv, does he get a reasonable error message back, btw? :)
[04:29] <lulu> didn't realise you had alternate access - good - I'll leave you to it then.
[04:30] <spiv> kiko: Fault 8002.  (i.e. no :)
[04:30] <kiko> spiv, very funny. :)
[04:33] <spiv> lulu: Sent.
[04:37] <lulu> spiv: thanks hon - we're aiming to have this done by close of play today so let's make sure it happens.
[05:18] <kiko> stub, you've got bugmail :)
[05:20] <stub> fun, fun, fun
[05:24] <sabdfl> db upgrade all taken care of, before it's too late .au time?
[05:26] <stub> sabdfl: db upgrade went well. Launchpad code  had issues Robert was going to look at tomorrow (I think)
[05:27] <kiko> stub, afaik they are solved, see comment to you in #canonical
[05:31] <elmo> I'm going to upgrade the postgres DB software now - this'll involve a restart - if this a problem for anyone shout RSN
[05:36] <elmo> (done)
[05:45] <lulu> spiv: ping
[05:52] <lulu> spiv: still problems with same error code
[05:57] <spiv> Yeah, all unexpected errors get that error code at the moment.
[05:57] <spiv> I'm just waiting for the PQM to accept my merge.
[05:57] <lulu> spiv: ok - thanks.
[06:17] <spiv> lulu: I just saw a createUser call succeed on the test server, and I'm pretty sure the other calls are fine.
[06:17] <spiv> (but I'm double-checking just in case...)
[06:18] <lulu> spiv: good news - well done - let me know after the double check!
[06:18] <spiv> lulu: Will do :)
[07:42] <kiko> ddaa!
[07:42] <ddaa> oi
[07:42] <ddaa> I'm doing bits of ProjectProductSetup to help lifeless
[07:43] <ddaa> I am a bit puzzled by the whole idea of project/products...
[07:43] <ddaa> I got an interesting sample right here...
[07:43] <kiko> ddaa, join the club. how can I help?
[07:44] <ddaa> You can try to fill me in with the philosophy of projects and products.
[07:44] <kiko> ddaa, are you familiar with DOAP?
[07:44] <ddaa> Hu... not especially...
[07:44] <kiko> http://usefulinc.com/doap
[07:44] <kiko> check that out.
[07:44] <ddaa> Okay, I guess i should start by reading some dog about that.
[07:56] <ddaa> Well.. not much more informative than what is already on the ProjectProductSetup page on the wiki...
[07:58] <ddaa> Consider zsh...
[07:58] <ddaa> Well, that's a product. And that's a product. And that's all as far as I can tell.
[07:59] <ddaa> Defining a zsh Project would mostly mean copy/pasting the displayname, title, shortdesc and description...
[08:01] <kiko> I think most projects have single products, but IMBW
[08:01] <ddaa> Or take zope. That's a project, and that's a product, but then I hardly see how how to come up with different descriptions for the two, I draw inspiration from the Debian package description and the plone.org website, and there does not seem to be a distinction.
[08:02] <ddaa> Or take gzip/zlib. That would make sense at two products of the same gzip project. But then there is not gzip project per se.
[08:03] <ddaa> The web pages for both products seem to prosent them as independent software which happen to overlap in purpose and it authorship..
[08:03] <ddaa> Or is that GNU project stuff?
[08:03] <kiko> the zope project has a *lot* of products.
[08:03] <ddaa> Name a few.
[08:04] <ddaa> I'm not particularly cognisant of the whereabouts of most free software projects.
[08:05] <ddaa> So for example, zope application server, zope database.
[08:05] <ddaa> That's the same project.
[08:06] <ddaa> But there is no info file for zodb...
[08:06] <ddaa> only one for zope
[08:07] <kiko> no no
[08:07] <kiko> or rather, yes
[08:07] <kiko> zodb and zope are zope products.
[08:07] <ddaa> Sure.
[08:07] <kiko> there are many others, just look at zope.org
[08:07] <ddaa> Yeah I saw a raft of software packages "which are hosted on zope.org"
[08:07] <kiko> right
[08:08] <kiko> many of them official Zope products
[08:08] <kiko> (medusa is another, for instance)
[08:09] <ddaa> okay... It's hard to tell from where I stand, but I guess one can tell given a bit of careful browsing...
[08:09] <ddaa> But what is the relation of all this to the info files.
[08:09] <kiko> right
[08:09] <kiko> that's beyond what I can tell you. lifeless otoh is the man with the plan :)
[08:11] <ddaa> Bah...  I'll stick to what I planned initially... Do some stuff, put it on the wiki, wait feedback...
[08:11] <ddaa> Successive approximations technique...
[08:11] <ddaa> ho, btw, how can I see the project/products which are already in the db?
[08:12] <ddaa> That might enlighten me and that will prevent duplication of effort.
[08:12] <kiko> ddaa, look at the sampledata file, for one.
[08:14] <debonzi> hi all, Ive made an star-merge from rocketfuel and my launchpad is not running anymore. It starts but I have only blanck pages and a traceback. Does anybody knows about it?
[08:14] <debonzi> s/blanck/blank
[08:14] <debonzi> ERROR root Exception during task
[08:14] <debonzi> AttributeError: 'NoneType' object has no attribute 'id'
[08:15] <debonzi> any ideia?
[08:15] <debonzi> s/ideia/idea
[08:15] <ddaa> kiko: bah... nevermind, I'll bug lifeless instead.
[08:16] <kiko> ddaa, as I said, he *will* know (and is possible the only person alive who does)
[08:16] <ddaa> maybe should fix that :->
[08:17] <kiko> possibly. yeah, we should steveA just the other day grilled us because we didn't know anything about it.
[08:17] <ddaa> ah.. I meant fixing the "alive" part
[08:17] <ddaa> not the "only person" part
[08:19] <kiko> LOL
[08:19] <kiko> lunchpad.
[08:19] <ddaa> gotto leave
[08:19] <ddaa> dinnerpad here
[08:19] <kiko> later then dude
[08:19] <kiko> bon apetit mon ami
[08:28] <debonzi> As far I can see, nobody knows :(
[08:28] <kiko> debonzi, it could be your database is foorked 
[08:29] <debonzi> I don't know.. Ive already rebuilded my db
[08:30] <kiko> are your tree and celso's identical?
[08:30] <debonzi> nop.. 
[08:31] <debonzi> I started to happens since Ive made an star-merge from rocketfuel
[08:31] <debonzi> s/I/It
[08:31] <kiko> have you looked through the diffs?
[08:31] <debonzi> Nop..
[08:31] <kiko> AFAIK lifeless hit the tree earlier today. could be that.
[08:32] <kiko> debonzi, something like:
[08:32] <kiko> https://rosetta.warthogs.hbd.com/soyuz
[08:32] <kiko>  -- ?
[08:33] <debonzi> yes... the same
[08:33] <kiko> that's fucked up.
[08:33] <kiko> daf?
[08:33] <debonzi> but here I can see a traceback in the terminal where launchpad is running
[08:33] <daf> kiko: hi
[08:34] <daf> hmm
[08:34] <kiko> daf, can you take a look at rosetta there. we're getting 100% blanks and silly tracebacks and we need help from above :)
[08:34] <daf> AttributeError: 'NoneType' object has no attribute 'id'                                                                                                                   
[08:34] <kiko> daf, it's not just in rosetta, of course, but you can probably evaluate easily from there.
[08:34] <debonzi> thats it
[08:34] <debonzi> the same traceback
[08:35] <daf> kiko: I'd be glad to help, but I'm not sure what I can do
[08:35] <daf> it seems to be something to do with publishing
[08:36] <kiko> 3 cheers for the spiver
[08:37] <spiv> kiko: No guarantees :)
[08:37] <spiv> But thanks for the vote of confidence :)
[08:38] <kiko> spiv, you had better fix it, we are going to be 'seeing' you this sunday!
[08:38] <kiko> spiv, daf: have I sent enough email and bugmail for today?
[08:39] <spiv> kiko: I only just got my email back about an hour ago, so I just got it all at once :)
[08:39] <kiko> :0
[08:39] <kiko> * From: kiko@async.com.br
[08:40] <kiko> /dev/null
[08:40] <daf> ?!
[08:40] <spiv> So, from reading the traceback, it appears that there's two problems...
[08:41] <kiko> daf, a procmail recipe? :)
[08:41] <daf> kiko: I know that! :)
[08:41] <spiv> There's an exception being raised in publish, and then it fails due to request.principal == None.
[08:42] <spiv> The original exception is a nameerror...
[08:42] <spiv> "global name 'transaction' is not defined"
[08:43] <kiko> weird.
[08:44] <spiv> Ok, found it.
[08:44] <spiv> I wonder how changed that.
[08:44] <kiko> spiv, lifeless did some hackage today, what file is it?
[08:45] <spiv> lib/canonical/publication.py
[08:45] <spiv> --- orig/lib/canonical/publication.py
[08:45] <spiv> +++ mod/lib/canonical/publication.py
[08:45] <spiv> @@ -169,10 +169,10 @@
[08:45] <spiv>          # the transaction in which they are instantiated - this is
[08:45] <spiv>          # no longer a problem as we are nuking the connection cache,
[08:45] <spiv>          # but it is still an issue in SQLOS that needs to be fixed.
[08:45] <spiv> -        name = getUtility(IConnectionName).name
[08:45] <spiv> -        con = sqlos.connection.getConnection(None, name)
[08:45] <spiv> -        t = transaction.get_transaction()
[08:45] <spiv> -        t.join(con._dm)
[08:45] <spiv> +        #name = getUtility(IConnectionName).name
[08:45] <spiv> +        #con = sqlos.connection.getConnection(None, name)
[08:45] <spiv> +        #t = transaction.get_transaction()
[08:45] <spiv> +        #t.join(con._dm)
[08:46] <spiv>          # Try to authenticate against our registry
[08:46] <spiv>          prin_reg = getUtility(IPlacelessAuthUtility)
[08:46] <spiv> (That patch works for me ;)
[08:48] <kiko> https://chinstrap.warthogs.hbd.com/archzoom/rocketfuel@canonical.com/launchpad--devel--0--patch-380/lib/canonical/publication.py.diff?diff
[08:48] <kiko> spiv, you merged that in?
[08:49] <spiv> Yeah.
[08:49] <spiv> kiko: Heh, I just found that too.
[08:50] <spiv> Looks like it was my fault :(
[08:50] <spiv> I thought I'd done "tla undo" :/
[08:51] <kiko> pity it can't be wittier but the old topic is stale.
[08:51] <spiv> :)
[08:51] <kiko> spiv, broken on the rock of tla again, huh? :)
[08:51] <spiv> Yeah, I guess so :)
[08:52] <spiv> Anyway, the merge request with the fix should be almost done by now.
[10:29] <kiko_> spiveroo!