[01:10] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: basic infrastructure to see the po/pot import queue (patch-867)
[07:14] <stub> elmo: ping
[09:33] <sabdfl> morning!
[09:33] <sabdfl> stub: couple of quick questions
[09:34] <stub> yo
[09:34] <sabdfl> we need to do a production code update today
[09:35] <sabdfl> we're nearly at the point where there's a nice web process for the whole arch import
[09:35] <sabdfl> i'll just need to make a few more code changes, test them and publish, do you have time?
[09:36] <stub> publish where?
[09:36] <sabdfl> rf
[09:36] <sabdfl> commit
[09:37] <stub> eh?
[09:39] <sabdfl> rocketfuel
[09:39] <sabdfl> erm... read-only filesystem here, got to sort that out first.
[09:47] <lifeless> stub: we've got db changes needed to enable code changes.
[09:47] <lifeless> so we want to drop them into a production code drop.
[09:47] <lifeless> may take several iterations to get right, but impt to do today.
[09:48] <stub> Several iterations of production database patching? Or will that be just once?
[09:48] <lifeless> hopefully once, but you never know
[09:48] <lifeless> :)
[09:49] <stub> Mmm... so 'production' is now 'dogfood' or 'staging'.
[09:49] <stub> How bought I create a launchpad_staging database on emperor we can screw with until we get it right?
[09:50] <lifeless> hmm, we should have it right, its just a little pressure cooker right now is all.
[09:52] <stub> ok - if your fairly confident we don't have to do multiple attempts at patching the database we can just update production as per normal. Otherwise it is trivial to make a second DB for some staging tests.
[09:55] <lifeless> it would more be adding more fields, not random patching
[09:56] <lifeless> i.e. might find we're missing X or Y later.
[10:48] <dilys> Merge to rocketfuel@canonical.com/cscvs--devel--1.0: sync cscvs to rf (patch-40)
[11:13] <dilys> Merge to rocketfuel@canonical.com/gnarly--devel--0: some fixes (patch-1)
[11:54] <dilys> Merge to rocketfuel@canonical.com/cscvs--devel--1.0: port to gnarly (patch-41)
[11:54] <stub> lifeless: Do you have a guestimate on when there will be a db patch for me to look at?
[11:59] <lifeless> stub: marks laptop fell down went boom
[11:59] <lifeless> :[
[12:00] <stub> So tomorrow morning my time earliest I gather?
[12:18] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: lib/gnarly symlink (patch-868)
[12:20] <dilys> Merge to rocketfuel@canonical.com/dists--devel--0: start of production-5 config (patch-34)
[12:32] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: _correct_ lib/gnarly symlink (patch-869)
[12:35] <lifeless> stub: mmm, I guess.
[12:47] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: Gina binary/source crosslinking (patch-870)
[01:00] <Kinnison> So how do I get tree-lint to shush about lib/gnarly?
[01:24] <Kinnison> never mind; I've sorted it
[01:42] <Kinnison> If I want baz to tell me what an undo dir contains can I find out without trying to apply it?
[01:45] <carlos> I'm not completely sure, but I think "baz changes --diffs ,,undo1" 
[01:45] <SteveA> daf, carlos: hello
[01:45] <carlos> could work
[01:45] <carlos> SteveA: hi
[01:46] <SteveA> how's the rosetta work coming along?
[01:46] <Kinnison> PANIC: exiting on botched invariant
[01:46] <Kinnison> is what I get if try that carlos 
[01:47] <carlos> Kinnison: no idea then. I saw doing something like that, ask him (or any member of arch team)
[01:48] <carlos> SteveA: I'm close to finish the queue moderation (I think it needs about 1 hour of work)
[01:48] <carlos> that's after I'm back from lunch
[01:48] <SteveA> what exactly needs doing?
[01:49] <carlos> the submit
[01:49] <carlos> and the process associated to the submit, of course
[01:50] <carlos> http://gollum.pemas.net:8085/rosetta/+uploadqueue
[01:50] <carlos> here you have the template I'm using
[01:50] <elmo> stub: can you [or find someone to]  ensure postgresql-contrib gets into hoary/main, if we're going to be using it in the long term
[01:50] <stub> Ok - I'll talk to martin
[01:50] <elmo> AIEE
[01:51] <elmo> and the stupid thing restarts postgres
[01:51] <carlos> you can use foo.bar@canonical.com with password: test to see it
[01:51] <elmo> so sorry, to anyone I just broke
[01:51] <elmo> and double AIEE, it install autovacuum by default
[01:51] <elmo> do we want that?
[01:51] <stub> Might as well.
[01:52] <carlos> SteveA: are we going to have a meeting now? I was thinking on leave to have lunch now
[01:52] <stub> Seems to be recommended procedure (provided you still do occasional full vacuums)
[01:52] <SteveA> carlos: no, we can talk later.  enjoy your lunch.
[01:52] <carlos> thanks
[01:52] <carlos> later
[01:53] <elmo> stub: how does it not cause random hangs?
[01:53] <elmo> i.e. launchpad has an idle spate, so autovacuum kicks in on some super-huowge table [not that we have (m)any of them ;)] , and locks updates out until it finishes?
[01:54] <stub> vacuum doesn't stop the server from responding - just slows it down a bit when it is running. And the more frequently you vacuum, the less work there is to do and the faster it completes. 
[01:54] <stub> The vacuum kicks in when it sees 'significant changes'. Not to sure about the details.
[01:55] <elmo> I thought it took a Big Global[per-table]  Write Lock(tm), but maybe I'm out of date
[01:55] <stub> 7.1 did that
[01:55] <stub> (or was it 7.0?)
[01:56] <Kinnison> I thought it just locked the rows it was moving
[01:56] <elmo> stub: presumably emperor will need this too?  and macquarie?
[01:57] <stub> eventually, yes. I was just running it up the flagpole yet - show people what can be done inside postgres to see if it is good enough.
[01:59] <elmo> okay, installed on mawson and chinstrap
[01:59] <stub> ta muchly :-)
[02:10] <lifeless> stub: ping
[02:10] <stub> lifeless: pong
[02:10] <lifeless> hey can we do a update of what we have ?
[02:11] <lifeless> just the current state. got a password fixing problem et al to correct.
[02:11] <stub> There are only three patches to roll out, only one of which may be of relevance to you (a constraint change in binarypackage). 
[02:12] <lifeless> elmo: ping
[02:12] <stub> I'd rather batch up more changes unless live stuff needs the changes
[02:12] <lifeless> stub: their are fixes to project and product status athat live stuff needs
[02:12] <lifeless> to increase the workflow stuff
[02:13] <SteveA> stub: there is a critical fuxkup in the forgottenpassword app in production
[02:13] <SteveA> can we fix it?
[02:15] <stub> lifeless: I don't think I have seen those fixes.
[02:15] <stub> SteveA: No idea. That is in plone somewhere, isn't it?
[02:15] <salgado> is there any problem in using something like dbschema.MembershipStatus.items[self.status] .title instead of iterating over dbschema.MembershipStatus.items looking for self.status?
[02:15] <SteveA> stub: no, it is in launchpad
[02:15] <SteveA> it just looks like the ul website
[02:16] <SteveA> I am currently checking that it works in rocketfuel
[02:16] <SteveA> it looks to be a 1 line fix
[02:16] <SteveA> salgado: that is fine
[02:16] <elmo> lifeless: ?
[02:16] <stub> If that is launchpad@macquarie, lifeless might be doing a rollout
[02:17] <lifeless> can we get python 2.4 on galapagos ? We've a nasty <foo> affecting cscvs that ddaa reckons doesn't happen in 2.4
[02:17] <lifeless> stub: stub mark is working again, he's looking like being done soonish
[02:17] <elmo> oh, jesus, you're kidding.  already?
[02:17] <lifeless> elmo: sorry, no I'm not.
[02:18] <SteveA> lifeless: you could compile python as cscvs user youself
[02:18] <SteveA> if we have no package availabe
[02:18] <lifeless> SteveA: do I then need to compile twisted too
[02:18] <lifeless> ?
[02:19] <lifeless> I'd really rather not go the whole do-it-myself route.
[02:19] <SteveA> does twisted use any C ?
[02:19] <lifeless> dunno. :|
[02:19] <lifeless> let me check
[02:20] <stub> debs should be available, at least for debian. It is pretty far into the beta cycle.
[02:20] <SteveA> it is at RC1
[02:20] <elmo> debs are available in hoary
[02:20] <elmo> that doesn't mean it's easily installable on our machines
[02:21] <lifeless> stub: mark says can you hang about for 30 minutes so he can finish one last db patch
[02:21] <lifeless> stub: should make check in launchpad work ?
[02:21] <stub> Yup, but he needs to buy me a beer ;)
[02:21] <stub> lifeless: yes
[02:22] <lifeless> he so owes you a 'best spanish beer'
[02:22] <lifeless> twisted has C code
[02:23] <SteveA> lifeless: get it from hoary
[02:23] <lifeless> SteveA: what from hoary ?
[02:24] <lifeless> SteveA: well, it fails all over the place.
[02:26] <elmo> guys, we REALLY don't want to start going down the road of requiring hoary or later packages for our infrastructure
[02:26] <elmo> it's just UTTER crack
[02:26] <lifeless> I can just imagine your accent as you say that
[02:30] <SteveA> daf: HELO
[02:30] <lifeless> elmo: so whats the resolution here ?
[02:38] <elmo> lifeless: sigh, I'll look at python2.4 packages for galapagos
[02:38] <elmo> but, be aware, I HATE YOU ALL
[02:38] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: steveA says the make test target should be nuked... nuke it (patch-871)
[02:40] <lifeless> elmo: so, we need pyscopg, twisted, in python 2.4 SteveA - thats the lot right ?
[02:43] <lifeless> SteveA stub : vvvvvv
[02:43] <lifeless> FAILED (failures=38, errors=20)
[02:43] <lifeless> Exception psycopg.InterfaceError: 'already closed' in <bound method Transaction.__del__ of <sqlobject.dbconnection.Transaction object at 0xb6359aac>> ignored
[02:43] <lifeless> ---- end test stderr ----
[02:43] <lifeless> make: *** [check]  Error 1
[02:43] <lifeless> I've dropped all my db's, and run make in the db/schema dir
[02:44] <lifeless> I'm fully up to date w/rocketfuel
[02:44] <stub> That is a spurious warning from SQLObject because the twonks decided that __del__ methods in python actually worked.
[02:44] <lifeless> FAILED (failures=38, errors=20)
[02:44] <lifeless> is what I'm concerned about
[02:44] <stub> Indeed.
[02:44] <lifeless>     ProgrammingError: ERROR:  relation "person" does not exist
[02:44] <lifeless>     SELECT name, displayname, givenname, familyname, password, teamowner, teamdescription, karma, karmatimestamp FROM Person WHERE id = 16
[02:45] <sabdfl> stub: sorry about this
[02:45] <sabdfl> can i msg you an sql fragment for approval?
[02:46] <stub> lifeless: Last time somebody complained about Person missing (bob2), rebuilding the database fixed it.
[02:47] <sabdfl> can i make that 5-04?
[02:47] <stub> ('make' in database/schema, with particular attention to the last few lines which copy launchpad_ftest_template to launchpad_dev)
[02:47] <stub> sabdfl: Yes - patch-5-04-0.sql
[02:48] <sabdfl> stub: ok, i will move it across and add the upda laucnhpaddatabaserevision;
[02:48] <stub> oh - hang. I didn't realize you ment I had the patch to look at now.
[02:49] <elmo> ask sabdfl for his sounds - they're hard to ignore
[02:50] <sabdfl> elmo: "hassole"
[02:56] <sabdfl> stub: * launchpad-5-00-0.sql is corrupt or has been modified
[02:56] <sabdfl> ?
[02:56] <salgado> lifeless, did you registered my gpg key in PQM?
[02:58] <stub> sabdfl: Run 'baz changes' to see if you accidently mucked with it.
[02:59] <sabdfl> stub: not currently, but maybe i committed a change accidentally? 
[02:59] <sabdfl> i certainly didn't edit it
[03:05] <sabdfl> this is weird
[03:06] <sabdfl> stub: i don't have any changes to that file in the baz logs
[03:06] <sabdfl> nuthin
[03:06] <sabdfl> make check was working just a second ago
[03:07] <stub> sabdfl: Can you run 'md5sum launchpad-5-00-0.sql'
[03:07] <sabdfl> 5f7922451b4619c7d73a4d0cd5bcefaa  launchpad-5-00-0.sql
[03:09] <stub> Well that is plain stupid. Same md5sum here, and it matches what is in the makefile. Yet it fails for you and works for me.
[03:09] <sabdfl> my Makefile has a different md5sum
[03:10] <sabdfl> # Confirm that launchpad-5-00-0.sql hasn't been messed with - this file
[03:10] <sabdfl> # is our baseline telling us what was installed into production
[03:10] <sabdfl> check:
[03:10] <sabdfl>         @if [ "`md5sum launchpad-5-00-0.sql`" != 'ec9ccdbcdcfc1f15c15368f2642c17ed  launchpad-5-00-0.sql' ] ; then echo "* launchpad-5-00-0.sql is corrupt or has been modified"; exit 1; else echo "* Using launchpad-5-00-0 as baseline"; fi
[03:10] <stub> Hmm... I must have done a bad commit or a partial one or something :-(
[03:17] <stub> Either change the md5 line to match what you have, or change line 14 of the Makefile to not do the check
[03:17] <sabdfl> hold on!
[03:17] <sabdfl> oh, no
[03:17] <sabdfl> wrong idea
[03:17] <sabdfl> i just did a refuel and got excited when the patch touched a Makefile
[03:17] <sabdfl> but it was a different Makefile
[03:17] <Kinnison> Heh
[03:27] <sabdfl> stub: pqm merge request sent, go ahead as soon as dilys says it's done
[03:28] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: Make taxi work again. (patch-872)
[03:28] <stub> That one?
[03:32] <sabdfl> nope
[03:40] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: finalise sourcesource import workflow (patch-873)
[03:40] <sabdfl> stub: that's it
[03:43] <dilys> Merge to rocketfuel@canonical.com/dists--devel--0: Use a working sqlobject. (patch-35)
[03:44] <ddaa> stub: can you please reset my pwd on the dogfood server?
[03:44] <ddaa> s/pwd/passwd/
[03:48] <stub> ddaa: You know your production password?
[03:48] <stub> If so, I'll mirror it to dogfood
[03:49] <ddaa> well, I did the "forgotten password" thing on prod, but I was never able to actually log in using that passwd. SteveA was unable to debug the problem because the "forgotten passwd" procedure failed on him...
[03:50] <ddaa> So, yes, I know I have set up something, but I have no evidence it's actually got where I intended it to go.
[03:50] <ddaa> in other words: okay, go ahead, and we'll see if something works.
[03:52] <stub> ddaa: There are two accounts on the production system - david@canonical.com and david.allouche@canonical.com. Which one of those are you most likely to know the password of?
[03:54] <ddaa> Mh... I have passwod for both of them in my notes... I'm totally confused about the scope of accounts...
[03:54] <ddaa> Just david.allouche@ for the moment...
[03:54] <ddaa> that's the one I updated recently
[03:55] <stub> ok. I'll merge the other while I'm at it.
[03:57] <stub> You are supposed to have one account on the production system, with multiple email addresses linked to it. The UI doesn't encourage this though, which needs to be fixed. The dogfood system talked to a different database, so you need another account there.
[03:59] <stub> ddaa: Done. Your dogfood account has the same password as your production one.
[04:16] <stub> lifeless, sabdfl: Production database updated to latest (5-04)
[04:17] <SteveA> stub: general baz discussion session in progress in mark's flat
[04:32] <carlos> OperationalError: no connection to the server
[04:32] <carlos> stub: I think the dogfood server lost the db connection
[04:37] <stub> ta. postgresql got bounced earlier due to a product installation - would have died then.
[04:37] <dilys> New Malone bug #89: "Launchpad does not survice database restarts", submitted by Stuart Bishop
[04:37] <dilys> https://dogfood.ubuntu.com/malone/bugs/89
[04:38] <SteveA> stub: any idea what we need to do to make launchpad survive?
[04:39] <stub> If you execute a query on a connection that is dead, it should attempt to reconnect and only raise an exception if that fails. I don't know if the culprit is in Z3 or SQLOS or SQLObject.
[04:39] <SteveA> sqlos or sqlobject I should think
[04:40] <carlos> SteveA: I think daf already filed that bug report
[04:40] <stub> The exception ended up in zope.app.rdb, but it might just be SQLOS not driving the API correctly
[04:40] <carlos> sorry
[04:40] <carlos>  /s/SteveA/stub/
[04:43] <carlos> stub: forget that, I think daf talked about it already but the bug was not filed
[04:43] <stub> Thats good - the duplicate-of field in Malone got lost a while ago ;)
[04:44] <carlos> X-)
[04:49] <carlos> SteveA: Now that I'm implementing the final import into the database, some doubts come into my mind
[04:49] <carlos> (doubts, not questions :-P)
[04:50] <carlos> SteveA: usually, it could be a long process (several minutes)
[04:51] <carlos> that will make the connection to timeout before getting the answer
[04:53] <SteveA> yeah, I figured that might be the case.  here's what we do.
[04:53] <SteveA> 1. implement it as you are doing so,  check it all works properly with small imports
[04:55] <carlos> ok
[04:56] <SteveA> we have a choice: re-do it as a script, or make it not depend on being connected and return an "import id" and allow you to check the import id at a url.
[04:56] <SteveA> I think it might just work as a script.
[04:56] <SteveA> we'll run the script as a long-running process that just imports the next thing, then imports the next thing after that, and if there is nothing else, waits for a minute or so before looking again
[04:58] <carlos> Do we have a procedure about how to execute scripts?
[04:59] <carlos> or it's just a kind of fork & exec?
[05:00] <SteveA> the thing is, we don't want to run this script when it is already running
[05:00] <SteveA> my first thought is to make the script a long-running process, and to have it made so that if you run the script while it is already running, in the same working directory (with the same configuration perhaps), then it doesn't actually run a new one
[05:02] <carlos> makes sense
[05:02] <SteveA> so, first, we get it totally working
[05:02] <SteveA> btw, I will be on stage, presenting this on the middle saturday of the conference
[05:02] <SteveA> so I really do want it to work!
[05:02] <SteveA> and then, we optimize with this script
[05:03] <SteveA> I want to see it totally working on dogfood before we work on the script
[05:05] <carlos> sure, don't worry
[05:16] <sabdfl> stub: does the product db update include a production code update?
[05:17] <lifeless> sabdfl: yes
[05:17] <sabdfl> strange
[05:17] <lifeless> sabdfl: I haven't seen your email
[05:17] <lifeless> stub: I've rebuilt
[05:18] <elmo> lifeless: ?
[05:18] <sabdfl> launchpad.ubuntu.com is definitely not running the latest code
[05:22] <elmo> lifeless: I have 2.4 debs ready for python, psycopg, twisted and egenix.  let me know if/when I'm okay to install
[05:22] <lifeless> sabdfl: I haven't done a code drop, I've been in the baz meeting :)
[05:22] <lifeless> elmo: cool.
[05:22] <sabdfl> ok
[05:33] <stub> lifeless: ?
[05:33] <BradB> sabdfl: I'll be checking in a quick prototype shortly of what I had in mind for improving the search UI for the bug listing. Will you have a chance to take a look in about 10-15 mins?
[05:34] <lifeless> stub - I rebuilt my db
[05:34] <stub> And everything is magically working? Or is something still bitching about the Person table being missing?
[05:41] <lifeless> I had before reporting the failure
[05:42] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: added prototype search UI for improving the bug listing UI (patch-874)
[05:46] <lifeless> stub: ping
[05:46] <sabdfl> BradB: possibly!"
[05:46] <lifeless> is the authserver down ?
[05:46] <sabdfl> seems we have a problem with the auth server
[05:46] <sabdfl> stub, elmo could you reset the authentication server?
[05:47] <elmo> done
[05:48] <SteveA> spiv is back on monday
[05:48] <BradB> sabdfl: if you have a moment to check it out, let me know. there was a couple of minor points to note.
[05:49] <sabdfl> BradB: is it committed
[05:49] <BradB> sabdfl: yeah (see dilys message a few lines back :)
[05:49] <sabdfl> ah, i see ok
[05:49] <sabdfl> i was mentally looking for dilys to be telling me about batching, because that still seems busticated
[05:49] <BradB> s/seems/is/ :)
[05:50] <sabdfl> __len__ on a resultset is doing something new and different and broken :-)
[05:50] <SteveA> really?  worked for me...
[05:50] <BradB> SteveA: You didn't check in a test for it though, which means you probably didn't run the sqlobject test suite either before you checked it in.
[05:51] <SteveA> I didn't change sqlobject
[05:51] <SteveA> I don't have rights to do that
[05:52] <BradB> What did you change then? :)
[05:52] <BradB> It was sqlobject, I could have swore (if not the official repo, but still...)
[05:52] <lifeless> elmo: go ahead
[05:52] <SteveA> I added a __len__ attribute to SelectResults that does:  lambda self: self.count()
[05:53] <carlos> SteveA: xqf.sf.net wants to use rosetta _now_ to translate it
[05:53] <SteveA> carlos: that is great
[05:53] <carlos> SteveA: well, some of them, they are talking about it now
[05:53] <carlos> they are asking about what they should do
[05:53] <carlos> they como from Jordi Mallach
[05:53] <carlos> already an alphatester of rosetta
[05:53] <BradB> sabdfl: When's a good time to spend five minutes discussing improving the bug UI to make it easier to find bugs? (for both avoiding duplicate reports and for just getting where you want to go quickly)
[05:53] <elmo> lifeless: installed.  nb, python2.4 isn't the default (obviously?)
[05:54] <carlos>  /s/como/come/
[05:54] <SteveA> BradB, sabdfl: even when I remove my patch, I get the same problem calling results.count()
[05:55] <BradB> What problem?
[05:55] <SteveA> I am confident that my patch has not caused the problem I'm seeing in sql result
[05:55] <lifeless> elmo: oh, how do I make it ?
[05:55] <SteveA> >>> from canonical.lp import initZopeless
[05:55] <SteveA> >>> from canonical.launchpad.database import Language
[05:55] <SteveA> >>> Language.select()
[05:55] <SteveA> <sqlobject.main.SelectResults object at 0x40d9e2cc>
[05:55] <SteveA> >>> Language.select().count()
[05:55] <SteveA> Traceback (most recent call last):
[05:55] <SteveA>   File "<stdin>", line 1, in ?
[05:55] <SteveA>   File "/stuff/code/Launchpad/launchpad/lib/sqlobject/main.py", line 1255, in count
[05:55] <SteveA>     count = self.accumulate('COUNT(*)')
[05:56] <SteveA>   File "/stuff/code/Launchpad/launchpad/lib/sqlobject/main.py", line 1251, in accumulate
[05:56] <SteveA>     return conn.accumulateSelect(self,expression)
[05:56] <SteveA> AttributeError: 'ConnectionDescriptor' object has no attribute 'accumulateSelect'
[05:56] <SteveA> >>>
[05:56] <stub> lifeless: If you are getting that error (relation person missing), then either the database is not building correctly or your program is connecting to the wrong database. bob2 was getting exactly the same error btw., so I expect you are doing whatever he was.  Running 'psql -d launchpad_dev -c "\d person"' will tell you if the table exists.
[05:56] <elmo> lifeless: the default?  I don't think we can sanely, that means changing _everything_ to use python2.4  which is a lot more than just psycopg, egenix and twisted
[05:56] <BradB> SteveA: are you running that on a db backend that supports count() ops?
[05:57] <elmo> can't you just invoke your stuff as /usr/bin/env python2.4 instead of /usr/bin/env python ?
[05:57] <BradB> it's not immediately obvious to me what initZopeless means WRT to backend
[05:57] <lifeless> stub: it doesn't.
[05:57] <lifeless> elmo: hmm. I'll talk to ddaa
[05:57] <stub> lifeless: I'll need the output of 'make' then
[05:57] <stub> lifeless: You run your code as 'python2.4 whatever.py' instead of 'python whatever.py'
[05:58] <lifeless> stub running testing script
[06:00] <carlos> elmo: https://rosetta.shuttleworthfoundation.org/ is not working
[06:00] <carlos> did you moved it to point to the dogfood server?
[06:01] <elmo> I haven't touched it
[06:02] <lifeless> stub mailed
[06:05] <stub> lifeless: The postgresql account that is being used is not a PostgreSQL super user.
[06:06] <carlos> Proxy Error
[06:06] <carlos> The proxy server received an invalid response from an upstream server.
[06:06] <carlos> The proxy server could not handle the request GET /.
[06:06] <carlos> Reason: Error reading from remote server
[06:07] <lifeless> stub: how do I fix that ?
[06:07] <carlos> sabdfl, SteveA: with the launchpad to dogfood change, what happens with rosetta? should It be moved to dogfood also?
[06:08] <stub> sudo -u postgresql psql -d template1 -c "alter user lifeless createdb createuser"
[06:08] <ddaa> stub: what's up with dogfood atm?
[06:09] <stub> nothing?
[06:09] <ddaa> he does not seem to be answering anymore (I guess I will end up having a timeout)
[06:09] <Kinnison> @ Gina completed.
[06:09] <Kinnison> real    1m22.452s
[06:09] <lifeless> sudo: no passwd entry for postgresql
[06:09] <stub> c/postgresql/postgres
[06:09] <SteveA> daf: hello
[06:09] <ddaa> stub: you mean "nothing is up, I know that and am working on it" or "there is no problem I am aware of"?
[06:10] <ddaa> oh, nm, it's answering again
[06:10] <lifeless> ok, running make again
[06:11] <stub> Mmm.... must have had a few long running requests and the server blocked. I saw the same thing, so it isn't your net connection. But it was fixed by the time I had logged onto mawson.
[06:11] <ddaa> BTW, the passwd I set up for david.allouche@canonical.com on prod does not seem to be working on dogfood either...
[06:13] <ddaa> neither do any of the passwords I have in my books for david@canonical.com :-(
[06:13] <BradB> ddaa: could you email lp@ and tell us how to get gnarly?
[06:14] <ddaa> baz get rocketfuel@canonical.com/gnarly--devel
[06:14] <ddaa> BradB: why are you asking?
[06:15] <BradB> ddaa: Because I don't want my tree-lint to output stuff about a symlink that point to a non-existent file. :)
[06:15] <ddaa> I'll update dist.
[06:15] <Kinnison> update the development config
[06:15] <ddaa> yup
[06:16] <BradB> ddaa: If you could follow that up with an email to lp@ (so that you don't have to answer the question five or six different times in here), that'd be great. :)
[06:16] <lifeless> stub much happier
[06:16] <ddaa> okay, in exchange for that, can you please pester stub until I can log into prod and dogfood?
[06:17] <ddaa> ;-)
[06:23] <stub> ddaa: I've emailed you a fresh password
[06:23] <ddaa> stub: thanks, I'll have a look
[06:26] <lifeless> ddaa: what did you do to build zope with python 2.4 ?
[06:26] <ddaa> I do not quite remember...
[06:26] <lifeless> I need to do it now.
[06:27] <ddaa> Basically, I did apt-get source python-twisted, then looked at some of deps I needed, did a python2.4 setup.py
[06:29] <ddaa> I remember I needed to hack a bit on egenix-base to disable some FREELIST optimisation that caused crash with python2.4
[06:29] <lifeless> thanks.
[06:33] <dilys> Merge to rocketfuel@canonical.com/dists--devel--0: add gnarly to devel configs of buildbot and launchpad (patch-36)
[06:39] <ddaa> stub: thanks, it's working
[06:40] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: More deb/udeb stuff and an optimisation on sourcepackage scanning (patch-875)
[06:40] <Kinnison> thanks babe
[06:40] <Kinnison> Am I safe to kick off a *BIG* gina run on mawson?
[06:43] <Kinnison> stub: ^^
[06:44] <dilys> New Malone bug #90: "RCSTypeEnum and RCSNames belong in dbschema.py", submitted by David Allouche
[06:44] <dilys> https://dogfood.ubuntu.com/malone/bugs/90
[06:48] <carlos> SteveA: ok, the import form seems to be working now (except for a small problem I'm fixing now)
[06:48] <carlos> SteveA: what's the policy about the SQLObject's attributes update from browser.py?
[06:48] <carlos> I have an SQLObject and I want to do object.attribute= foo
[06:49] <SteveA> ok
[06:49] <carlos> at this moment I need to duplicat that field in the interface so it's also present in the edit interface
[06:49] <SteveA> ok
[06:49] <carlos> and add also it to the .zcml
[06:49] <carlos> to allow its update
[06:50] <carlos> a method is not needed, but an attribute I don't see any other way to do it
[06:50] <SteveA> that sounds okay
[06:50] <SteveA> you can say set_attributes="foo" in zcml too
[06:50] <carlos> I only know that way to allow attributes
[06:51] <SteveA> you can use set_schema too
[06:51] <carlos> that's why I'm not sure if it's the best way to do it
[06:51] <SteveA> but, use what you know
[06:51] <carlos> ok
[06:54] <carlos> SteveA: I suppose I should use the: launchpad.Edit permission now, right?
[06:54] <dilys> Merge to rocketfuel@canonical.com/dists--devel--0: full production-5 update (patch-37)
[06:54] <carlos> We have at the moment:
[06:54] <carlos> <content class="canonical.launchpad.database.POFile">
[06:54] <carlos>       <require permission="zope.Public"
[06:54] <carlos>           interface="canonical.launchpad.interfaces.IEditPOFile"
[06:54] <carlos>           set_attributes="header pluralForms rawimportstatus"/>
[06:54] <carlos>     </content>
[07:00] <carlos> SteveA: I don't understand this error:
[07:01] <carlos> ValueError: Unknown SQL builtin type: <class 'canonical.lp.dbschema.Item'> for <Item FAILED (4) from <class 'canonical.lp.dbschema.RosettaImportStatus'>>
[07:03] <Kinnison> I'm gonna assume it'll be okay for me to get gina going on mawson
[07:04] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: added implementation of prototype bug id/searchtext widget (patch-876)
[07:05] <Kinnison> elmo: can I get some up-to-date keyrings please for a hoary import?
[07:06] <elmo> uh?
[07:07] <elmo> up-to-date compared to what?  i.e. what are you using now?
[07:07] <Kinnison> whatever I had way back when you gave me a set of keyrings at the soyuz sprint
[07:07] <Kinnison> and I know there've been keychanges since then
[07:07] <elmo> there's a copy on chinstrap
[07:07] <elmo>  /srv/keyring.no-name-yet.com/ or so
[07:07] <Kinnison> okay ta
[07:09] <SteveA> carlos: how did you get that error?
[07:09] <carlos> SteveA: the trace shows it when launchpad tries to commit the transaction
[07:10] <carlos> potemplate.rawimportstatus = \
[07:10] <carlos>                             RosettaImportStatus.FAILED
[07:10] <carlos> that's the instruction I'm executing
[07:10] <SteveA> use RosettaImportStatus.FAILED.value
[07:11] <carlos> checking...
[07:11] <carlos> perfect
[07:11] <carlos> thanks
[07:12] <SteveA> we'll improve this in sqlobject at some point
[07:12] <SteveA> by making a kind of ENUM field
[07:13] <carlos> but this works: if template.rawimportstatus == RosettaImportStatus.PENDING:
[07:13] <carlos> so I suppose it's only a problem when using the value to change a variable, right?
[07:16] <SteveA> yes
[07:17] <SteveA> and, it should work.  I think it should be co-erced into an int/
[07:17] <SteveA> but a better solution overall is to make the ENUM
[07:17] <lifeless> a clue for thins ?
[07:17] <lifeless> zope.configuration.xmlconfig.ZopeXMLConfigurationError: File "/home/robertc/source/canonical/buildbot/launchpad/site.zcml", line 18.4-18.55
[07:17] <lifeless>     ZopeXMLConfigurationError: File "/home/robertc/source/canonical/buildbot/launchpad/override-configure.zcml", line 2.2-2.56
[07:17] <lifeless>     ZopeXMLConfigurationError: File "/home/robertc/source/canonical/buildbot/launchpad/override-includes/+mydatabase-configure.zcml", line 14.2-14.57
[07:17] <lifeless>     IOError: [Errno 2]  No such file or directory: '/home/robertc/source/canonical/buildbot/launchpad/lib/canonical/rosetta/sql.zcml'
[07:17] <lifeless> make: *** [run]  Error 1
[07:18] <SteveA> well...
[07:18] <SteveA> I expect +mydatabase-configure.zcml has an <include ... /> directive in it
[07:18] <SteveA> that refers to the sql.zcml file that is in rosetta
[07:21] <Kinnison> stub: ping?
[07:22] <lifeless> right. but what changed : what should I change it to ?
[07:22] <lifeless> Kinnison: dude, its 5:22am
[07:22] <Kinnison> stub: I want to start the gina import but this is gonna really hammer the db and librarian for around four hours. Will that be okay?
[07:22] <Kinnison> lifeless: Oh I guess his presence an hour ago made me assume he was readying himself for Mataro
[07:23] <lifeless> ;)
[07:24] <Kinnison> Aah well, noone wanted mawson to run quickly right?
[07:24] <daf> lifeless: if there was a "baz file-history" command, you could find out what happened to it :)
[07:26] <Kinnison> Well guys, mawson will be a touch busy now as it imports hoary main,restricted,universe for i386,powerpc,amd64
[07:26] <Kinnison> :-)
[07:26] <SteveA> who added the mail and launchpad-database files in package-includes?
[07:26] <Kinnison> (or not; yay for sourcepackage addition threw an error)
[07:27] <SteveA> stub / BradB|lunch: either of you add those files?
[07:29] <carlos> SteveA: should I link to the import queue page from a particular place?
[07:30] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: add missing import to auth app (patch-877)
[07:53] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: add fmt:date fmt:time and fmt:datetime stuff for use in tales (patch-878)
[07:56] <Kinnison> My name is Daniel and I am *SPESHUL*
[07:56] <lifeless> yes, yes you are
[07:56] <Kinnison> lifeless: we gonna see you this w/e at all?
[07:58] <carlos> daf: hey, I didn't saw you
[07:58] <daf> hi carlos
[07:59] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: fix sourcesource browser (patch-879)
[07:59] <lifeless> Kinnison: probably not
[07:59] <Kinnison> o/~ Guns don't import packages. Gina does! WooWooWooooo summon the soyuz
[07:59] <Kinnison> lifeless: Oh well. I'll see you in Spain then dude.
[08:00] <carlos> daf: the pair programming you suggested is ok for me
[08:00] <daf> ok, cool
[08:00] <dilys> Merge to 	rocketfuel@canonical.com/launchpad--production--1.5: cherry pick fix sourcesource browser (patch-1)
[08:02] <dilys> Merge to rocketfuel@canonical.com/buildbot--devel--0: Use the unix process backend. (patch-76)
[08:02] <dilys> Merge to rocketfuel@canonical.com/gnarly--devel--0: Add missing import. (patch-2)
[08:02] <carlos> wow, pqm is really busy atm...
[08:03] <daf> yeah
[08:04] <carlos> daf: I'm going to take a break, will be back in about 30 minutes
[08:04] <daf> ok
[08:05] <dilys> Merge to rocketfuel@canonical.com/buildbot--devel--0: Stick description and title into the database. (patch-77)
[08:17] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: Finished the basic rawimport queue edit (patch-880)
[08:20] <dilys> Merge to rocketfuel@canonical.com/buildbot--devel--0: properly detach jobs when the slave is lost (patch-78)
[08:26] <Kinnison> Mmm phantom
[08:41] <BradB> lifeless: What's the status of Launchpad running on an upgraded Z3?
[08:45] <carlos> daf: I'm ready
[09:17] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: fixed a bug in the bug id/text search when no matching bugs are found (patch-881)
[09:33] <daf> BradB: if I do somesqlobjectthing.foo = "'; DROP DATABASE launchpad", it's going to escape the ', right?
[09:35] <BradB> Yeah
[09:35] <BradB> kiko-afk, sabdfl: Where's the UI to assign one or more maintainers to a product or source package?
[09:51] <kiko-afk> BradB, it's obtained via lucille/gina, and not done via the website [yet?] 
[10:49] <dilys> Malone bug #12 fixed for package malone: Adding a follow-up to a comment fails
[10:49] <dilys> https://dogfood.ubuntu.com/malone/bugs/12
[10:50] <BradB> sabdfl: ping
[10:51] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: made followup widgets identical to top-level comment widgets (patch-882)
[11:08] <BradB> daf, carlos: Is https://dogfood.ubuntu.com/malone/bugs/26 fixed?
[11:09] <carlos> not yet
[11:09] <BradB> ok
[11:09] <carlos> I think stub fixed the problem I had to fix them
[11:09] <carlos> but I hadn't time to test it
[11:10] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: disambiguated product name in bug listing (patch-883)
[11:11] <dilys> Malone bug #46 fixed for product Malone: bug assigment product name ambiguous
[11:11] <dilys> https://dogfood.ubuntu.com/malone/bugs/46
[11:15] <dilys> Malone bug #86 fixed for product Malone: Bug listing needs "Quick Searches"
[11:15] <dilys> https://dogfood.ubuntu.com/malone/bugs/86
[11:15] <dilys> Malone bug #80 fixed for product Malone: Bug titles/descriptions need to be text searchable
[11:15] <dilys> https://dogfood.ubuntu.com/malone/bugs/80
[11:31] <dilys> New Malone bug #91: "Bug listing clickability needs to be easier to use", submitted by Brad Bollenbach
[11:32] <dilys> https://dogfood.ubuntu.com/malone/bugs/91
[11:32] <BradB> daf: ISTR dilys doesn't notify the channel on rejects. Would it be possible to wire this up to make it so? #launchpad'ers will want to see that someone just rejected a bug they reported.
[11:33] <daf> sure
[11:33] <BradB> thanks
[11:33] <daf> BradB: I'm busy right now
[11:33] <BradB> no prob
[11:34] <daf> I've dumped the code at http://muse.19inch.net/~daf/dump/dilys-malone