[01:28]  * wallyworld hates "An updated diff will be available in a few minutes. Reload to see the changes. " My reload button is all worn out
[01:28] <StevenK> You still have F5
[01:29] <wallyworld> StevenK: oh, i was wondering what that keyboard button was. The "F5" has all worn off  from overuse :-P
[01:31] <wallyworld> and another review for someone https://code.launchpad.net/~wallyworld/launchpad/remove-mp-relatedbugs/+merge/52069
[01:46] <thumper> :(
[01:46] <thumper> dumb windows email clients
[01:50] <wgrant> Oh?
[01:50] <wgrant> Dumber than Gmail?
[01:51] <thumper> what does the mail RFC say about encoded email headers?
[01:51] <thumper> we are getting some with windows-cp1252
[01:53]  * wgrant checks 2047.
[01:53] <wgrant> Ah, it permits anything IANA-registered. :(
[01:54] <wgrant> But windows-cp1252 doesn't appear on IANA's charset registry...
[01:54] <thumper> wgrant: meaning?
[01:55] <wgrant> Its proper name is apparently windows-1252, not windows-cp1252
[01:56] <thumper> yeah, I'm getting that one
[01:56]  * thumper sighs 
[01:56] <thumper> my problem, not the code
[01:56] <wgrant> Oh?
[01:56] <thumper> the email['from'] field is encoded with windows-1252
[01:57] <wgrant>  * 2513 Exceptions
[01:57] <wgrant> :(
[01:57] <wgrant> ~500 more than I expected.
[01:57] <wgrant> Ah, MP job breakage.
[01:57] <lifeless> timeouts are down!
[01:57] <thumper> Content-Type: text/html; charset=windows-1252
[01:57] <lifeless> -woo-
[01:57] <thumper> that is what the email says
[01:57] <lifeless> ScopedCollection:CollectionResource:#person-page-resource is new
[01:57] <wgrant> thumper: That's valid.
[01:57] <thumper> but the headers are encoded with that charset too
[01:57] <lifeless> as is Product:+filebug-show-similar
[01:58] <wgrant> thumper: Hm, do you have the librarian URL for that email?
[01:58] <thumper> it's spam
[01:58] <wgrant> 88/176 for BugTask:+index. That's nice.
[01:58] <thumper> http://launchpadlibrarian.net/65554538/9a948026-45ee-11e0-86fc-001e0bc3957e.txt
[01:58] <lifeless> freaking awesome
[01:59] <wgrant> thumper: Ugh.
[01:59] <StevenK> lifeless: Aren't you on a plane?
[01:59] <wgrant> I would have thought so.
[01:59] <lifeless> boards at 10 past
[01:59] <lifeless> in wellington
[01:59] <wgrant> Oh, right.
[01:59] <wgrant> thumper: So, catch the lack of encoding and reject, I suppose :/
[02:00] <thumper> wgrant: well, at the moment it gets rejected through the oops :)
[02:00] <wgrant> Unless we have a convenient way to apply UnicodeDammit to MIME.
[02:00] <wgrant> Damn.
[02:00] <wgrant> The exception summary doesn't quite get to single-ocurrence OOPSes yet.
[02:01] <wgrant> Tomorrow!
[02:01] <lifeless> RootObject:+recently-registered-branches is a little slow
[02:02] <lifeless> wgrant: 131 OperationalError: FATAL: Ident authentication failed for user "person-merge-job" FATAL: Ident authentication failed for user "person-merge-job"
[02:02] <lifeless> OOPS-1888REPORTIFSEEN109, OOPS-1888REPORTIFSEEN110, OOPS
[02:02] <lifeless> should be easy
[02:02] <poolie> wallyworld, in https://code.launchpad.net/~wallyworld/launchpad/mp-related-bugtasks-webservice/+merge/52004
[02:03] <poolie> would that also work, and be cleaner, as just a collection link?
[02:03] <lifeless> wgrant: you might like to remind folk setting up jobs that they *must* make a new config and *must* have the losas set it up : we *cannot* [because of oops ids] just use the default config
[02:03] <wgrant> lifeless: Sure, if we had a LOSA :)
[02:03] <poolie> or is that already what it does?
[02:03] <lifeless> wgrant: I say you, because I'm on a plane :>
[02:03] <lifeless> poolie: we should talk about collections vs queries.
[02:03] <lifeless> poolie: It may be my failing, but I rather dislike the billion-collections approach.
[02:03] <poolie> we you and me?
[02:03] <wallyworld> poolie: can collection links take parameters?
[02:03] <lifeless> we you and me
[02:03] <wallyworld> i don't think so
[02:04] <poolie> no, they can't
[02:04] <poolie> but this doesn't seem to take parameters
[02:04] <wallyworld> it takes the user
[02:04] <poolie> oh?
[02:04] <lifeless> bug 726803
[02:04] <_mup_> Bug #726803: not possible to export a method as a collection with bound parameters <lazr.restful:Triaged> < https://launchpad.net/bugs/726803 >
[02:04] <wgrant> thumper: https://lp-oops.canonical.com/oops.py/?oopsid=OOPS-1888BZR102541 needs looking at.
[02:05] <poolie> wallyworld, i thought that meant it depends on the implicit request user?
[02:05] <wallyworld> poolie: we need the user since getRelatedBugTasks() needs it so that it can see what bug tasks are allowed to be exposed
[02:05] <wgrant> thumper: It's a single user running a script against a single branch, AFAICT always looking for the next revno.
[02:05] <poolie> in that case, yes, of course collections can look at that
[02:05] <lifeless> >< http://news.cnet.com/8301-17938_105-20028475-1.html
[02:05] <wgrant> thumper: Happens a few hundred times a day, and gets logged as an OOPS.
[02:05] <thumper> do we know who?
[02:06] <wallyworld> poolie: we set a default user to be the request user but i think we still need to allow other users to be specified?
[02:06] <wgrant> thumper: It was easy enough to track down, but I don't remember.
[02:06] <wgrant> thumper: Regardless, it should not OOPS.
[02:06] <thumper> hmm...
[02:06] <thumper> agreed
[02:06] <thumper> I wonder why it does
[02:06] <wgrant> thumper: Oh, right:
[02:06] <wgrant> user_id: 1689259
[02:06] <thumper> wgrant: is there a bug
[02:06] <wgrant> thumper: I don't believe os.
[02:07] <poolie> wallyworld, oh, i see, it makes an optional parameter
[02:07] <poolie> http://trpdsaya.tumblr.com/post/3463810356/omg-this-app-is-so-restful
[02:07] <poolie> also http://trpdsaya.tumblr.com/post/3600386585/the-url-structure-on-this-site-is-so-pretty for that matter
[02:07] <wallyworld> poolie: yes. i think i have done it the right way
[02:07] <wallyworld> but i'm happy to be told i need to do it another way
[02:08] <lifeless> I think its fine for now
[02:08] <lifeless> its in devel
[02:08] <lifeless> if we want to impove it we can :)
[02:08] <lifeless> later y'all.
[02:08] <poolie> much better to have it there at all
[02:08] <poolie> i was just wondering
[02:08] <wallyworld> np. always good to ask questions
[02:09] <wgrant> StevenK: I wonder if mawson is going to survive publishing all of natty, seriously this time.
[02:09] <StevenK> Haha
[02:09] <wgrant> StevenK: Will parallel a-f with all of the files present OOM it, I wonder?
[02:09] <StevenK> You tell funny jokes
[02:09] <poolie> it would be kind of nicer if it was not a ws.op thing, but rather
[02:09] <wgrant> StevenK: It has finished downloading all the files, so is now calculating lists for a-f. I guess that will take another couple of horus.
[02:10] <poolie> https://api.launchpad.net/devel/~wallyworld/launchpad/mp-related-bugtasks-webservice/+merge/52004/+related-bug-tasks
[02:10] <poolie> and then you could filter it etc
[02:10] <poolie> but that probably needs more infrastructure elsewhere
[02:11] <wallyworld> poolie: you suggesting that that api be added?
[02:12] <StevenK> wgrant: cocobanana and mawson have roughly the same amount of RAM
[02:13] <StevenK> wgrant: *and* cocobanana runs rsyncd, which eats babies.
[02:14] <wgrant> StevenK: I thought cocoplum would have more.
[02:14] <wgrant> Plus cocoplum doesn't have 223GB of DB.
[02:14] <StevenK> No, it has a 400GB archive instead :-)
[02:15] <wgrant> But it doesn't need to keep a big chunk of that in RAM.
[02:15] <wgrant> Also, mawson now has 120GB of that archive :P
[02:15] <StevenK> Haha
[02:15]  * StevenK looks at changing the db user for a garbo class
[02:16] <michaelh1> Hi there.  When a merge request comes in I'd like to do a bzr export, fire it into my build farm, then update the merge request with the automatic build results.  Are there existing tools past launchpadlib I can reuse?
[02:24] <StevenK> Rargh, TunableLoop can't do that?
[02:26] <thumper> michaelh1: not as far as I'm aware
[02:27] <michaelh1> thumper: ta.
[02:29] <StevenK> wgrant: I don't know if the populate job can run under garbo, just due to DB perms
[02:32] <jtv> jcsackett, really still here?  If so, got a review for you!  https://code.launchpad.net/~jtv/launchpad/bug-623391/+merge/52114
[02:36] <jcsackett> jtv: ah no, actually. can't seem to change channel info though. could you change it to take me off?
[02:36] <jtv> thanks StevenK :)
[02:37] <wgrant> StevenK: You know that you can change DB perms, right?
[02:37] <StevenK> Assuming allenap is sleeping
[02:37] <jtv> He should be!
[02:37] <wgrant> Looks like we need to get memcached killed on the lucid_db_lp slave :(
[02:38] <wgrant> And no LOSAs :(
[02:38] <StevenK> wgrant: Sure, but we don't change db perms on prod. Or am I out of date?
[02:38] <wgrant> StevenK: If a script needs more permissions, give it more DB permissions.
[02:38] <wgrant> We grant new permissions during nodowntime rollouts.
[02:39] <StevenK> wgrant: I can do that. I'm still wondering how the heck to test it
[02:42] <wallyworld> thumper: you able to review any of my branches?
[02:42] <wallyworld> https://code.launchpad.net/~wallyworld/launchpad/remove-decoratedbug/+merge/52003
[02:42] <thumper> wallyworld: aye
[02:42]  * thumper looks
[02:42] <wallyworld> and  https://code.launchpad.net/~wallyworld/launchpad/remove-mp-relatedbugs/+merge/52069
[02:42] <wallyworld> thanks :-)
[02:43] <wallyworld> this one should also be good to go since i fixed the issue raised https://code.launchpad.net/~wallyworld/launchpad/mp-related-bugtasks-webservice/+merge/52004
[02:44] <thumper> wallyworld: 710 lines?
[02:44] <wallyworld> which one? isn't the limit 800?
[02:45] <wallyworld> thumper: much/most of the line count is removing stuff
[02:45] <wallyworld> there's lots of red there
[02:53] <poolie> hi thumper
[02:53] <poolie> are you going to fix 728827 by downgrading the message from warning to info or something?
[02:54]  * wallyworld has to take kid to train station
[02:57] <thumper> poolie: yeah
[02:57] <poolie> great
[02:57] <thumper> done even
[02:58] <poolie> sorry, i didn't realize just having a warning message caused an oops
[02:58] <poolie> i guess it makes sense
[02:58] <thumper> neither did I
[02:58] <thumper> I had to hunt to work out where it was happening
[02:58] <wgrant> It only started a few months ago.
[02:59] <poolie> only started giving a warning then?
[02:59] <wgrant> Warnings and errors only became OOPSes then.
[02:59] <poolie> ok
[02:59] <poolie> i wonder how much it is worth documenting that vs just having people gradually know it
[03:12]  * wgrant kicks buildbot-poller
[03:12] <StevenK> wgrant: Hm?
[03:12] <wgrant> StevenK: It pulled devel->stable but did not merge stable->db-devel.
[03:12] <wgrant> Possibly because db-devel was testfixed.
[03:13] <StevenK> Mmmm, the deployment report makes me happy.
[03:13] <wgrant> Yes.
[03:13] <wgrant> But there are several more revs that would be nice.
[03:14] <wgrant> eg. 12522
[03:14] <wgrant> Which is in stable, and cuts a couple of seconds off binary copies.
[03:15] <StevenK> We're just waiting for qastaging, then?
[03:15] <thumper> https://code.launchpad.net/~thumper/launchpad/mail-header-oops/+merge/52156 anyone?
[03:15] <thumper> I'm just pushing the conflict resolution merge
[03:18] <poolie> i will do that
[03:18] <poolie> do the review
[03:22] <poolie> thumper, how does the x-launchpad-original-to header get into the mail?
[03:22] <poolie> by an incoming gateway?
[03:25] <wgrant> Yes, an MTA adds it somewhere.
[03:30] <thumper> poolie: yeah
[03:31] <poolie> i can't retrieve any of the sample urls like http://launchpadlibrarian.net/62571148/916041d6-2558-11e0-b038-001e0bc3957e.txt
[03:31] <poolie> did they expire or something?
[03:33] <thumper> I think they were garbage collected
[03:33] <thumper> I'm not sure exactly how that works
[03:36] <poolie> anyhow i commented
[03:55] <thumper> poolie: I replied
[03:56] <poolie> oh, blah
[03:56] <poolie> i just got a "Your message was rejected" from launchpad-bugs-owner@lists.canonical.com
[03:57] <thumper> :)
[03:58] <thumper> https://code.launchpad.net/~thumper/launchpad/strip-email-attachment-path/+merge/52159 anyone?
[04:07] <wallyworld> thumper: i'll take a look
[04:09] <wallyworld> thumper: with my 3 mp's,  they all follow on from each other. do i have to give them to ec2 one at a time, or can i just land the last one?
[04:09] <thumper> wallyworld: you can just do the last one
[04:10] <wallyworld> thumper: can i approve the middle one even though it says needs fixing - i've addressed the requested change (to export to the "devel" ws api not 1.0) but lifeless i don't think is online atm
[04:11] <thumper> wallyworld: socially, no
[04:11] <wallyworld> yeah thought so :-)
[04:11] <wallyworld> i'll wait
[04:11] <thumper> wallyworld: programatically, yes
[04:20] <wallyworld> jtv: got a question for you if you are around
[04:20] <jtv> wallyworld: I'm around, as long as you're not planning to bother me with questions
[04:21] <wallyworld> oh :-(
[04:22] <wallyworld> jtv: bug 407260
[04:22] <_mup_> Bug #407260: Translations export branch can't be team-owned <lp-translations> <Launchpad itself:Triaged by wallyworld> < https://launchpad.net/bugs/407260 >
[04:22] <jtv> ah that
[04:22] <jtv> (clever of you not to use question marks)
[04:22] <wallyworld> i tried to write a test case to reproduce it
[04:22] <wallyworld> but my test worked
[04:22] <wallyworld> so clearly i'm missing something
[04:22] <jtv> Yes, IIRC the problem is with the branch picker
[04:23] <StevenK> wgrant: It seems I can't do self.getCandidateSPRIDs(0).count() == 0
[04:23] <wallyworld> i tried to adapt test_translations_export_to_branch
[04:23] <wallyworld> and made the branch owner a team
[04:23] <StevenK> wgrant: FeatureError: Single aggregates aren't supported after a  GROUP BY clause
[04:23] <wgrant> StevenK: No, since some of them will fail.
[04:23] <wgrant> That too.
[04:23] <wgrant> But the garbo job will probably never end, since some will fail to unpack.
[04:24] <wgrant> This may put an end to that idea.
[04:24] <StevenK> Until we remove it
[04:24] <jtv> wallyworld: perhaps more precisely, my recollection is that we have no vocabulary for "a branch that I'm *an* owner of"
[04:24] <wallyworld> jtv: oh, so it's a gui problem
[04:24] <wgrant> StevenK: But there may be enough that every run only processes the broken ones, without reaching the end.
[04:25] <wallyworld> jtv: the back end code works regardless, but the gui doesn;t allow the desired branch to be selected
[04:25] <jtv> wallyworld: pretty muchâ€”it only occurs when setting the branch
[04:25] <wallyworld> jtv: thanks. i thought it was a back end problem
[04:25] <wallyworld> but the test worked, so clearly not :-)
[04:26] <StevenK> wgrant: I'll reimplement the finish_at bit then
[04:26] <jtv> wallyworld: Not that I can think of, noâ€¦ we decided not to check for ownership during actual commit, IIRC; as an owner you get to write out a permanent license for LP to commit to it.
[04:26] <wallyworld> jtv: i'm doing this one because a user brought it up on #launchpad and it seems a good one to fix, since it affects a few people and has been around for a while
[04:26] <wgrant> StevenK: How do you plan to preserve this across garbo sessions?
[04:26] <wallyworld> thanks for the clarification
[04:27] <StevenK> wgrant: I don't
[04:27] <wgrant> StevenK: The problem will still exist.
[04:27] <jtv> wallyworld: yes, I've been wanting to do that for ages.
[04:28] <wallyworld> jtv: cool. see, me asking questions wasn't so bad after all :-)
[04:28] <jtv> Only because I wasn't around.
[04:29] <StevenK> wgrant: How many do you expect won't unpack?
[04:30] <wgrant> StevenK: Potentially thousands.
[04:31] <StevenK> That makes me sad.
[04:31]  * thumper is getting the muchies
[04:31] <thumper> or munchies even
[04:31] <StevenK> But it's only 5:30pm?
[04:34] <thumper> yup
[04:34] <thumper> after beer o'clock on a Friday
[04:35] <wallyworld> thumper: have a beer for me too
[04:35] <thumper> wallyworld: ack
[04:35] <wallyworld> or a nice red even
[04:35] <wallyworld> i can't tonight - have to go out in the pouring rain and coach soccer :-(
[04:37] <jtv> StevenK: something I'm wonderingâ€¦
[04:37] <jtv> â€¦if you're still here.
[04:38] <LPCIBot> Yippie, build fixed!
[04:38] <LPCIBot> Project devel build #501: FIXED in 5 hr 9 min: https://hudson.wedontsleep.org/job/devel/501/
[04:38] <StevenK> jtv?
[04:38] <Ursinha> this bot is so happy
[04:38] <StevenK> Haha
[04:38] <jtv> StevenK: as you know I'm generating DistroSeriesDifferences directly based on SPPH/SPR comparison between two DS.
[04:39] <jtv> I create the DSDs with source_version set to SPR.version from the derived series, and parent_source_version set to SPR.version from the parent series.
[04:39] <jtv> Up to one of those can be null, if the difference type is not different_versions.
[04:40] <jtv> (And on a sidenote, I suspect that unique_to_derived_series could be very subtly misleading and that specific_to_derived_series might be better)
[04:41] <jtv> I just thought I'd ask if there is still substantial work to do for the case where a package is only in the parent or only in the derived series.
[04:42] <StevenK> jtv: If it's only in the parent or only in the child, we don't want a DSD
[04:42] <jtv> Err
[04:42] <jtv> Then why are there DSDTypes for those cases?
[04:43] <StevenK> jtv: I have no idea -- but there *no* *point* generating diffs for them
[04:43] <thumper> two or three critical bugs closed today as they code paths no longer exist and oops hasn't happened in ages :)
[04:43]  * thumper EODs
[04:43] <jtv> StevenK: then maybe all you want is to block the things?
[04:43] <jtv> bye thumper
[04:43] <thumper> have a good weekend y'all
[04:43] <jtv> same to you
[04:43] <StevenK> thumper: Have a good weekend
[04:43] <jtv> TGIF & ODIM
[04:44] <jtv> The Yin & Yang of the business world.
[04:45] <jtv> StevenK: I meant, maybe the only thing we really want from packages that only exist on one of the two sides is the ability for the owner to either block or synchronize?
[04:46] <StevenK> jtv: Exactly
[04:46] <jtv> Because if that is the case, and it can be done entirely in SQL, then we have no need to bother the DSD script with those cases.
[04:47] <jtv> The way the script works now, figuring out the differences between a DS and its parent and creating all the DSDs all happens in SQLâ€”none of the data ever gets pulled into the python code.
[05:08] <wgrant> jtv: Does http://paste.ubuntu.com/575313/ mean anything to you? We didn't see anything like that in the old code, but everything looks fine anyway.
[05:11] <jtv> wgrant: are we sure that the old code would have reported output like that in the first place?
[05:11] <wgrant> jtv: I am not entirely sure.
[05:12] <jtv> ISTR the old code being a bit iffy in that department.
[05:12] <wgrant> Anyway, once I've run the publisher twice with this to get timings, I will rerun with the old one to compare the time and diff the trees.
[05:12] <wgrant> Hopefully the next run will not take 24 hours :)
[05:12] <jtv> Well it'll be weekend for you either way won't it?  :)
[05:13] <wgrant> A good point.
[05:16] <wgrant> With luck this weekend will have fewer production incidents than the last.
[05:26] <wgrant> jtv: OutputLineHandler has a bug.
[05:26] <jtv> ?
[05:27] <wgrant> >>> s = OutputLineHandler(sys.stdout.write)
[05:27] <wgrant> >>> s('foo')
[05:27] <wgrant> >>> s('bar\n')
[05:27] <wgrant> foobar>>> s('bar\n')
[05:27] <wgrant> foobar>>> s('bar\n')
[05:27] <wgrant> It never empties the buffer if the input ends with \n.
[05:28] <wgrant> Not critical, but explains some of the ridiculous output I've been getting.
[05:45] <jtv> wgrant: damn, I thought I'd covered that specifically.
[05:45] <jtv> IIRC the \n was stripped out somewhere _before_ it got to the OutputLineHandler
[05:50] <poolie> hi
[05:50] <wgrant> Yay, netsplits.
[05:59] <poolie> wgrant, do you agree with me on https://code.launchpad.net/~thumper/launchpad/mail-header-oops/+merge/52156
[05:59] <wallyworld> jtv: just to check i understand, "you" still handle translation approvals? eg blue squad now on maintenance, and there's a queue entry waiting approval. do i need to do anything?
[06:00] <jtv> wallyworld: do you have a translations veteran on the team?
[06:00] <wallyworld> jtv: our team is me, tim, leonard, and steve but steve is on loan to the red team at the moment
[06:00] <wallyworld> so no :-(
[06:01] <jtv> (By the way, I can't tell you how much it means to me for another developer to tell me there is "a" queue entry awaiting approval)
[06:01] <StevenK> wallyworld: jtv is on red, so he is aware of what I'm working on :-)
[06:01] <wallyworld> jtv: i filtered on the suggested "ANy Project" and "Pot only". was that not correct?
[06:02] <jtv> Yes, it isâ€”and presumably on the also-correct Needs Review as well.
[06:02] <jtv> wallyworld: want to walk through it with me?
[06:02] <jtv> I don't think it'll be very hard.
[06:02] <wallyworld> jtv: yes please, so i can communicate back to the blue team
[06:02] <jtv> Just click on the edit icon next to "no import target selected yet."
[06:02]  * wallyworld clicks
[06:02] <jtv> Well I say "just," but I mean "for starters."  :)
[06:03] <wgrant> poolie: I've wondered that myself. For the field to be inserted after a blank line, our MTA would presumably have to be buggy.
[06:03] <wallyworld> jtv: there is no edit icon
[06:03] <wgrant> poolie: I agree that investigation is required.
[06:03] <wallyworld> for me
[06:03] <jtv> wallyworld: then we haven't added you to the rosetta-admins team yet.  Let me fix that.
[06:03] <wallyworld> ok
[06:03] <poolie> it seems kind of unlikely to me that that bug would be triggered iff the message is spam
[06:03] <poolie> indeed if it's reliably doing that, we should sell it ;-)
[06:04] <poolie> it's possible it's some interaction between a spam scanner and something else
[06:04] <wgrant> It is a reasonable assumption that something with header fields in the body is spam.
[06:04] <wgrant> But I cannot see why *this* field would be there.
[06:04] <poolie> note that we're not checking for a body that looks like it contains headers
[06:04] <poolie> just for that string, anywhere in the body
[06:05] <poolie> i should say so on the mp i suppose
[06:05] <wgrant> Yes.
[06:08] <jtv> wallyworld: I've invited the launchpad-chr team to rosetta-admins, as per my email to the list.  But now we need that invitation accepted.
[06:09] <jtv> I'll email Diogo about it.
[06:09] <wallyworld> jtv: thanks. can we walk through an approval another time?
[06:09] <jtv> Absolutely.  I'll just do this one for now.
[06:09] <wallyworld> i'd still like to learn how to do it
[06:09] <wallyworld> thanks
[06:10] <wallyworld> jtv: i think that's the key thing to know about while on maintenance? anything else critical to be aware of?
[06:14] <jtv> wallyworld: plenty of recurring questions, but for most things IMHO it's best to find out as you go along.
[06:14] <wallyworld> jtv: ok. thanks again
[06:14] <jtv> We have lots of CHR documentation on the internal wiki, so you could try reading through that and following some of the links.  But I wouldn't recommend trying to keep up with all the FAQs.  On-demand caching is the ticket.
[06:17] <wallyworld> jtv: yeah, there was so much material i'd thought i'd try and cheat a bit by asking about the "viewing highlights"
[06:17] <jtv> Sensible.
[06:17] <wallyworld> :-)
[06:35] <huwshimi> Have a good weekend people
[06:57] <poolie> bye huwshimi
[07:30] <jtv-eat> Still no reviewers available?  Got two dependent branches blockedâ€¦ starting with https://code.launchpad.net/~jtv/launchpad/bug-623391/+merge/52114
[09:07] <adeuring> good morning
[09:28] <jtv> Still no reviewers in the house?  Need a critical pair of eyes for https://code.launchpad.net/~jtv/launchpad/bug-623391/+merge/52114
[09:28] <jtv> (And then I'll try to get someone else for the follow-up branch :)
[09:30] <henninge> adeuring: ^ Aren't you reviewing today?
[09:30] <henninge> Hi jtv! ;-)
[09:31] <jtv> hi!
[09:31] <adeuring> jtv, henninge, yes. I'll look at your branch. jtv
[09:31] <jtv> Wunderbar!
[10:11] <maxwell69> hello
[10:13] <maxwell69> I have a problem to build launchpad
[10:13] <maxwell69> when I do maake schema I get an error
[10:14] <maxwell69> with zc.buildout
[10:16] <maxwell69> "couldn't find index page for zc.buildout
[10:18] <maxwell69> I think the program is trying to download zc.buildout but I get ***BLOCKED*** by --allow-hosts
[10:53] <jtv> adeuring: I've been having some system & connection problemsâ€¦ did I miss any questions?
[10:54] <adeuring> jtv: no, and, btw, no complaints so far. nice to read the diff
[10:54] <adeuring> ... but i need some more time ;)
[10:54] <jtv> sure
[11:22] <maxwell69> Hello jtv, I have a problem with the "make schema" command
[11:22] <stub> Give me a name for a marker instance for something that might be a Storm subclass, or an instance of a Storm subclass. Besides IFooStormObjectOrClass.
[11:23] <jtv> maxwell69: I may be able to help, but it's Friday night here and I'm putting out some fires.  :)  What's the problem?
[11:23] <stub> nm. got one.
[11:24] <jtv> stub: a marker interface you meant?
[11:24] <stub> yes
[11:24] <stub> IUseSessionStore is what I'm actually after.
[11:25] <maxwell69> I'm trying to buils launchpad but I have an error:
[11:25] <maxwell69> "couldn't find index page for zc.buildout"
[11:26] <maxwell69> the program try to download zc.buildout but I have this error:  ***BLOCKED*** by --allow-hosts
[11:26] <adeuring> jtv: r=me
[11:26] <jtv> thanks adeuring!  I ought to find someone else to bother with the follow-up branch I guessâ€¦
[11:27] <maxwell69> but my connction wors well and the website is aviable
[11:27] <adeuring> jtv: nah, i'll take that too :)
[11:27] <jtv> adeuring: thanks again!
[11:27] <wgrant> maxwell69: How did you check out the source?
[11:27] <jtv> maxwell69: it sounds horribly like a zope + system setup problemâ€¦  which I wouldn't know much about.
[11:28] <wgrant> maxwell69: Did you use rocketfuel-setup?
[11:29] <wgrant> I suspect not.
[11:29] <maxwell69> wgrant: yes, I followed instructions on https://dev.launchpad.net/getting
[11:31] <jtv> Now, --allow-hosts sounds vaguely familiar
[11:31] <jtv> Maybe I'm just thinking hosts.allow.
[11:32] <wgrant> It should be getting buildout from download-cache.
[11:32] <wgrant> maxwell69: What is in the download-cache directory inside your LP branch?
[11:37] <maxwell69> wgrant: I have "dist" directory
[11:37] <wgrant> maxwell69: Does that contain a few hundred files?
[11:37] <maxwell69> wgrant: and I manually put zc.buildout-1.5.1 dir
[11:39] <maxwell69> wgrant: in dist I just have setuptool***.egg
[11:40] <wgrant> maxwell69: There should be several hundred eggs in there.
[11:40] <wgrant> Is download-cache a symlink?
[11:41] <maxwell69> wgrant: I'm affraid not
[11:42] <wgrant> adeuring: Hi.
[11:44] <adeuring> hi wgrant
[11:44] <wgrant> adeuring: Could you please review https://code.launchpad.net/~wgrant/launchpad/turn-off-parallel-a-f/+merge/52185?
[11:45] <adeuring> wgrant: sure; let me just finish another review for Jeroen
[11:45] <wgrant> adeuring: Sure.
[11:45] <wgrant> Thanks.
[11:53] <maxwell69> wgrant: should I re-execute rocketfuel-setup?
[11:53] <wgrant> maxwell69: What do you have in ~/launchpad/lp-sourcedeps?
[11:54] <maxwell69> There is no lp-sourcedeps...
[11:55] <wgrant> maxwell69: Hmm. Did rocketfuel-setup give any errors?
[11:56] <adeuring> jtv: r=me again
[11:56] <jtv> adeuring: and thanks again
[11:58] <maxwell69> wgrant: there is warnings: DocumentRoot [/var/tmp/archive] does not exist
[11:59] <wgrant> maxwell69: That's fine.
[11:59] <maxwell69> and ./rocketfuel-setup line 362: cd: /usr/local/bin: no such file or directory
[12:00] <wgrant> That would be it.
[12:01] <wgrant> maxwell69: That should exist on fresh systems... could you create it and run rocketfuel-setup again?
[12:02] <maxwell69> ok, I have a call,I"ll do it and tell you what hapend :)
[12:02] <maxwell69> thanks
[12:04] <adeuring> wgrant: r=me
[12:05] <wgrant> adeuring: Thanks.
[12:17] <wallyworld> wgrant: i have a branch against devel which it turns out requires lines added to security.cfg. what's the best way to land it?
[12:17] <wallyworld> do i need to do anything extra?
[12:17] <wgrant> wallyworld: Is it just extra permissions, without new users?
[12:17] <wallyworld> wgrant: yes. just 2 extra select permissions
[12:18] <wallyworld> for an existing user
[12:18] <wallyworld> or section
[12:18] <wallyworld> whatever it's called
[12:18] <wgrant> So, mthaddon says you should talk to LOSAs, but that sort of change is no problem at all.
[12:18] <wallyworld> [merge-proposal-jobs]
[12:18] <wgrant> Since we now grant new permissions on every rollout.
[12:18] <wallyworld> cool. so just land as normal?
[12:18] <wgrant> Yup.
[12:19] <wallyworld> excellent. thanks.
[12:19] <mthaddon> if it's not adding new users, and is just changing existing permissions for users, that's fine
[12:20] <wallyworld> mthaddon: yes, adding public.bugsubscription = SELECT to [merge-proposal-jobs]
[12:21] <wallyworld> so just a new permission
[12:21]  * mthaddon nods
[13:08] <matsubara> jtv, ping
[13:08] <jtv> hi
[13:14] <matsubara> jtv, so, I think we can actually disable the launchpad-chr team, can't we?
[13:14] <jtv> matsubara: can we?  I don't know
[13:15] <matsubara> I think it was used as some form of access control/receive email
[13:15] <matsubara> let me dig that
[13:15] <matsubara> lifeless, hi
[13:15] <bac> hello adeuring.  busy today?
[13:15] <lifeless> hi
[13:15] <matsubara> lifeless, prefix LPNETPREFIX has been loaded. is there any other?
[13:15] <lifeless> hmm
[13:16] <lifeless> that may mean the config tom deployed was slightly incorrect
[13:16] <lifeless> matsubara: could you add servers through to Z?
[13:16] <matsubara> lifeless, yeah
[13:17] <matsubara> lifeless, and those are already handled by oops-tools. A to Z
[13:18] <lifeless> ok
[13:18] <lifeless> we'll need move to AA AB AC etc soon
[13:22] <matsubara> lifeless, really? do we have that many instances up?
[13:23] <matsubara> it'd be nice to get rid of the prefixes by that time as you proposed in the LEP
[13:23] <matsubara> jtv, sorry, I'll dig in and let you know here or by email
[13:23] <wgrant> lifeless has plans.
[13:24] <jtv> matsubara: well, I'm not particularly involved with the launchpad-chr team so all that it would take for me is to add rosetta-admins to another team.
[13:26] <adeuring> bac: hi bac, well, I did a few reviews.
[13:26] <lifeless> matsubara: I think we'll need the monday.
[13:27] <matsubara> lifeless, add the right prefixes to lp-production-config and oops-tools will read from there
[13:28] <matsubara> lifeless, instructions are here https://dev.launchpad.net/Foundations/QA/OopsToolsSetup  to make oops-tools recognize new prefixes (I can do it today, Monday I'll be off)
[13:28] <lifeless> matsubara: can you do me a favour? I'm roaming at the moment - drop me a quick mail with that link
[13:29] <matsubara> lifeless, sure
[13:29] <lifeless> matsubara: and enough info for me to step a losa through it
[13:29] <lifeless> (if its not on the page ;P)
[13:35] <maxwell69> wgrant: Hi again, I created the directory, launched the rocketfuel-setup again, files have been downloaded but when I do "make schema" I have the same error
[13:35] <maxwell69> Link to http://pypi.python.org/simple/ ***BLOCKED*** by --allow-hosts
[13:36] <wgrant> maxwell69: Is download-cache a symlink now?
[13:39] <maxwell69> wgrant: it is a dir
[13:42] <wgrant> maxwell69: rm -r download-cache; rm -r eggs; rocketfuel-get
[13:42] <wgrant> Oh, also rm -r sourcecode
[13:44] <maxwell69> wgrant: inlp-branches/devel ? just to be sure
[13:44] <wgrant> maxwell69: Yeah.
[13:45] <maxwell69> OK, downloading...
[13:49] <maxwell69> wgrant: Ok, I have now: "make: Entering directory `/home/maxwell/launchpad/lp-branches/devel' Missing ./download-cache. Developers: please run utilities/link-external-sourcecode. make: *** [download-cache] Error 1 make: Leaving directory `/home/maxwell/launchpad/lp-branches/devel'"
[13:50] <wgrant> maxwell69: in ~/launchpad/lp-branches/devel, run 'utilities/link-external-sourcecode ~/launchpad/lp-sourcedeps'
[13:57] <maxwell69> Ok now it is a symlink to "lp-sourcedeps/download-cache"
[13:57] <wgrant> Great.
[13:57] <wgrant> make might work now.
[13:59] <maxwell69> wgrant: looks good.
[14:00] <maxwell69> takes a lot of CPU :)
[14:01] <maxwell69> wgrant: sorry, I m a newbie, but i m not dev, juste a sysadmin :)
[14:02] <wgrant> maxwell69: Yeah, the first run of make will take a while do unpack all the eggs.
[14:12] <maxwell69> wgrant: the tail: utilities/shhh.py make -C sourcecode build PYTHON=python2.6 \ 	    LPCONFIG=development make: *** sourcecode: No such file or directory.  Stop. make: *** [compile] Error 2
[14:12] <wgrant> maxwell69: :(
[14:13] <wgrant> maxwell69: mkdir sourcecode, and run link-external-sourcecode again.
[14:14] <wgrant> The missing /usr/local got your tree into a bit of an odd state.
[14:20] <maxwell69> I ran make again:  make[1]: *** No rule to make target `build'.  Stop. make[1]: Entering directory `/home/maxwell/launchpad/lp-branches/devel/sourcecode' make[1]: Leaving directory `/home/maxwell/launchpad/lp-branches/devel/sourcecode' make: *** [compile] Error 2
[14:21] <maxwell69> I have symlinks in the sourcecode dir
[14:25] <bigjools> wgrant: do you have a feel for how much of the publisher (and associated cron stuff) is specific to Ubuntu?
[14:26] <benji> gary_poster: I had it like that but it looked really odd to me (this looks more like the edit and unsubscribe menu options on the bug page), that being said, I don't feel strongly about it; here it is all on one line: http://i.imgur.com/xRhom.png
[14:26] <wgrant> bigjools: The Python? Not at all. cron.publish-*? lol
[14:26] <gary_poster> benji, whee, let's go to #clojure next! ;-)
[14:27] <bigjools> wgrant: :D
[14:27] <wgrant> bigjools: publish-distro/process-accepted work fine on multiple distros.
[14:27] <bigjools> wgrant: you published debian didn;t you?
[14:27] <benji> oh I swear!
[14:27] <wgrant> bigjools: Yes.
[14:27] <bigjools> cool
[14:27] <gary_poster> :-)
[14:27] <bigjools> wgrant: I want to try and set something up so we can spawn off separate runs for publishers for different distros
[14:27] <henninge> abentley: http://paste.ubuntu.com/575490/
[14:28] <wgrant> bigjools: Definitely.
[14:28] <bigjools> wgrant: or DDs will be somewhat interesting
[14:28] <wgrant> bigjools: This may mean you finally get to rewrite cron.publish-ftpmaster in Python.
[14:28] <bigjools> wgrant: YES
[14:28] <wgrant> And cull crap like dsync.
[14:28] <bigjools> there's a blueprint about adding user hooks to it
[14:28] <bigjools> I want to make it generic enough to be re-used
[14:29] <bigjools> and then have a job runner fire off the scripts based on a schedule managed in LP
[14:29] <wgrant> Right.
[14:30] <bigjools> wgrant: can you remember how much of the publisher config is ubuntu-specific?
[14:30] <bigjools> or more to the point did you change much to get debian working
[14:31] <bigjools> I'm trying to work out if any of it "clashes" if you have >1 distro published
[14:32] <wgrant> bigjools: Nothing clashes.
[14:32] <bigjools> sweet
[14:32] <wgrant> But there are restrictions on where they can be.
[14:32] <bigjools> yeah I am mostly thinking of paths
[14:32] <wgrant> eg. archive dependency code assumes that all distros live under http://ftpmaster.internal/$DISTRO
[14:32] <wgrant> And the publisher assumes that everything lives under /some/common/prefix/$DISTRO
[14:33] <wgrant> But assuming you live within or lift those constraints (paths can be easily changed using separate configs, URLs are harder), there are no conflicts.
[14:34] <bigjools> can't solve this with config
[14:34] <bigjools> needs to come from the DB
[14:34] <wgrant> Right.
[14:34] <wgrant> It will need to be sort of like lucilleconfig, except with one path instead of 7 plus 2 other values that never change.
[14:34] <bigjools> I want to remove as much config as possible and have a management form
[14:35] <bigjools> "sort of like lucilleconfig
[14:35] <bigjools> "
[14:35] <bigjools> was your mouth out
[14:35] <bigjools> wash*
[14:35] <bigjools> jeez, type man
[14:35] <wgrant> Also possibly without the ini data.
[14:35] <wgrant> I guess all you need now is a single path, and possibly a machine identifier.
[14:35] <wgrant> And some way to define the base URL for buildds to use.
[14:43] <wgrant> bigjools: Anything else?
[14:43] <bigjools> wgrant: I can't think off-hand, more will come to light when we work through it
[14:43] <wgrant> Yeah, it will be interesting to see what falls out.
[14:44] <bigjools> migrating to the DB will be easy enough
[14:44] <wgrant> The publisher config is still a little bit insane, but it's much better than it was.
[14:45] <wgrant> Anyway, night.
[14:46] <maxwell69> wgrant: I paste it to have a clear view : http://paste.ubuntu.com/575507/
[14:48] <wgrant> maxwell69: 'bzr revert sourcecode/Makefile'
[14:48] <wgrant> And it nears 2am, so I hope somebody else can help you if you have further issue :)
[14:48] <wgrant> issues
[15:14] <henninge> abentley: got a bit distristracted but could try it now
[15:14] <henninge> abentley: that solved it, thanks.
[15:29] <bigjools> wgrant: Lucille is totally gone now right?
[15:29] <bigjools> LucilleConfig I mean
[15:30] <bigjools> meh you're probably asleep
[15:56] <jcsackett> rvba: don't know if you saw, but per your answers and bigjools i've marked your MP r=me.
[15:56] <rvba> jcsackett: yes I saw thanks, I'm waiting to be setup with PQM and I will ec2 land
[16:01] <jcsackett> rvba: cool stuff. :-)
[16:33] <jcsackett> where does lp.dev put oops files?
[16:44] <bigjools> in the log directory I think
[17:02] <jcsackett> leonardr: ping.
[17:06] <maxb> Is there a schedule for when staging restores happen?
[18:30]  * sinzui reconciles stable with db-devel
[18:48] <dobey> bah netsplits
[18:49] <LPCIBot> Project devel build #504: FAILURE in 4 hr 44 min: https://hudson.wedontsleep.org/job/devel/504/
[19:20] <sinzui> pqm hates me. It wont except my branch to fix db-devel and stable.
[19:20] <sinzui> What is wrong with my message: "[testfix][rs=sinzui] Resolve conflicts with stable."
[19:21] <sinzui> I have removed [testfix] and [rs=sinzui] and I still get a failure message with a 0 byte rejection log :(
[19:34] <dobey> hmm, LP recipe build request dialog seems to be broken. it immediately disables the series i want to build for with a "(Pending build)" comment, even though there are no pending builds. :-/
[19:39] <jelmer> dobey: it seems like a regression that was introduced in the last day or two
[19:40] <jelmer> bug 728789
[19:40] <_mup_> Bug #728789: "Request build" dialog indicates builds are pending when they are not and doesn't allow new builds <recipe> <regression> <Launchpad itself:Triaged> < https://launchpad.net/bugs/728789 >
[19:40] <dobey> jelmer: ok thanks
[19:40] <sinzui> I haet pqm
[19:41] <sinzui> hate, dislike, despise.
[19:41] <jelmer> dobey: fwiw the API still works
[19:42] <dobey> jelmer: right, except for the part where it has a quota. and that i'd have to write a script to request specific builds :)
[19:44] <jelmer> dobey: http://pastebin.ubuntu.com/575664
[19:44] <jelmer> dobey: the web UI and the API have the same quota afaik
[19:45] <dobey> jelmer: maybe, but my bot runs as a bot user, and on the web i'm me, so it's not so bad. can we get rid of the quota, or raise it significantly or something?
[19:46] <dobey> and yeah, i've written the code before for tarmac. i guess i should put a script in lptools for requesting builds
[19:58] <jelmer> dobey: I wouldn't object to that :) the current quota mechanism is too rudimentary
[20:22] <benji___> wgrant: am I reading https://wiki.canonical.com/InformationInfrastructure/OSA/LaunchpadProductionStatus#Cowboys%20%28requires%20IncidentReport%29 correctly in that both listed there have been landed and do won't have to be re-cowboyed on the next release?
[20:30] <jcsackett> benji: there's a good chance wgrant is actually here. it's early saturday morning for him.
[20:30] <jcsackett> s/is/isn't/
[20:30] <benji> jcsackett: I forgot to check my time zone decoder ring
[20:31] <jcsackett> if the usual lp deployment process is being followed, yes, the cowboys are already part of the main codebase, so you should be fine.
[21:20] <gary_poster> matsubara, hi.  Does the devpad sync from the losas help the oops tool respond any faster?  It doesn't seem to, but maybe I am being impatient. :-)
[21:22] <matsubara> gary_poster, a little bit. you can request the sync from the prod servers to devpad but then have to wait the cronjob that loads the oops from the filesystem into the db
[21:22] <matsubara> that cronjob runs every 7 min
[21:22] <gary_poster> ack, thanks matsubara
[21:22] <matsubara> np
[22:27] <wgrant> benji: Yes, they landed weeks ago. No need to worry about them this time.
[22:28] <lifeless> wgrant: so to measure lpnet latency
[22:28] <lifeless> wgrant: use the icing
[22:28] <wgrant> lifeless: That is a very good idea.
[22:28] <lifeless> wgrant: grab that a few times, gives you a baseline
[22:29] <benji> wgrant: thanks for the info
[22:30] <wgrant> lifeless: Seems to be 320ms, as expected. So the DC latency is often <50ms, it seems.
[22:30] <lifeless> (now)
[22:30] <wgrant> Right.
[22:30] <lifeless> opstats would let us graph that
[22:31] <lifeless> except ssl
[22:31] <lifeless> bwah
[22:35] <lifeless> I wonder, whats our fastest db page
[22:35] <lifeless> also
[22:35] <jelmer> lifeless: launchpad.net/404 ? :-P
[22:36] <lifeless> whats the open id sso service times
[22:36] <lifeless> jelmer: *db* page
[22:36] <jelmer> lifeless: What do you mean with database page then ? a 404 accesses the database
[22:36] <lifeless> it doesn't show anything useful though
[22:36] <lifeless> its not a successful request
[22:37] <lifeless> wow, page id fail... :createPPA
[22:37] <lifeless> (https://devpad.canonical.com/~lpqateam/ppr/lpnet/latest-monthly-pageids.html)
[22:40] <wgrant> Firefox is not a fan of that page.
[22:41] <lifeless> its the svg
[23:03] <wgrant> Hah. It takes 7 minutes to write out all 142000 records in the natty binary file lists at once, or 10 minutes per batch of 10000.
[23:03] <wgrant> The limit and offset must be fairly expensive...
[23:04] <lifeless> temp table time
[23:04] <lifeless> except don't do that cross xaction
[23:05] <wgrant> It's due to a compound cross-table sort key.
[23:05] <wgrant> I think I might reduce it to a single one, add a buffer, and then do the rest in Python.
[23:14] <wgrant> Real parallel a-f reduces what is a 16 minute process on cocoplum to a 5 minute process on mawson. Excellent.
[23:53] <maxb> I've just clicked in a join request for ~loggerhead-team, with the desire to be able to re-open merge proposals for branches de-merged by the shuffling off of recent trunk to the experimental branch