/srv/irclogs.ubuntu.com/2009/10/12/#launchpad-dev.txt

lifelessbug 44910500:00
ubot3`Malone bug 449105 in launchpad "logging into sourceforge with openid breaks" [Undecided,New] https://launchpad.net/bugs/44910500:00
mupBug #449105: logging into sourceforge with openid breaks <Launchpad itself:New> <https://launchpad.net/bugs/449105>00:00
mupBug #449105: logging into sourceforge with openid breaks <Launchpad itself:New> <https://launchpad.net/bugs/449105>00:00
mupBug #449105: logging into sourceforge with openid breaks <Launchpad itself:New> <https://launchpad.net/bugs/449105>00:00
lifelessoh joy00:01
lifelesshappy mup00:01
wgrantlifeless: I think it's SF.net misbehaving.00:06
wgrantlifeless: It's only half-respecting the delegation.00:06
* thumper away doing emails00:06
lifelesswgrant: happens with apache2-mod-openid00:06
lifelesswgrant: which I was trying to setup last week, looks exactly the same00:07
wgrantlifeless: The proprietary Canonical fork, or an open one?00:07
lifelessapt-get install00:07
wgrantAha.00:07
* wgrant checks the source.00:07
lifelesswgrant: what proprietary canonical fork?00:07
wgrantlifeless: https://launchpad.net/apache-openid00:08
lifelesswgrant: looks to be all handed back to me00:10
lifelesswgrant: oh, and I think its a different code base anyhow - there are *two* apache openid modules00:11
lifelessyeah, thats mpopenid, not openid00:11
lifelesswgrant: (mod python openid is what we use, apache2-mod-openid is written in C I think)00:12
wgrantAh.00:12
wgrantThe author of libopkele seems to have had a vendetta against whitespace.00:16
lifeless http://freshmeat.net/projects/libopkele/ ?00:25
wgrantThat one.00:26
lifelessyou're tracking down a root cause?00:27
wgrantI am.00:27
lifelesscool, thank you00:36
* mwhudson lunches01:01
wgrantlifeless: c-i-p bug. It's not present in the copy that I of course didn't retrieve from lp:launchpad's history, because I'm not allowed to do that.01:10
wgrantlifeless: All the declarations in the page are fine, but c-i-p now provides XRDS for OpenID 1.0 and 1.1, as well as 2.0 (which it always has)01:10
wgrantThe XRDS overrides the declarations on the page itself, and doesn't specify the identity to delegate to.01:11
wgrantJust the server.01:11
lifelessc-i-p ?01:14
lifelessoh, canonical-identity-provider?01:14
lifelesswgrant: I presume you'll put the details into my bug?01:15
wgrantlifeless: Yep, doing so now.01:15
lifelessthanks!01:15
wgrantI was very confused for a while before I noticed the XRDS link.01:16
wgrantBecause identical delegation code worked fine elsewhere.01:16
lifelesskfogel: http://www.pmease.com/features/screentour/03:17
lifelesskfogel: /nice/ 'tour' presentation03:18
spmwgrant: that buildd issue - *should* be sorted now03:27
wgrantspm: Thanks. What was the issue?04:03
spmfirewall upgrade and subtle funkies thereof04:03
wgrantLovely.04:04
spmexcluding the ones I did - which always worked perfectly. natch - I've never known a firewall upgrade/change to go perfectly. ;-)04:07
spmmy mates at $job-1 went thru 3-5 iterations of pain trying to work with the gateway provider (as it so happens $job-2 :-) )04:08
spmwhere each iteration is around 5-8 hours from midnight04:08
wgrantUrgh.04:08
spmand that was *just* the final firewall to the corporate network04:08
spmnot including the 4 or so DMZ ones just for their webserver farm04:09
spmeach of which has 13-17 active interfaces. \o/04:09
spmI *so* don't miss working at $job-2 :-D04:09
=== abentley1 is now known as abentley
=== ursula is now known as Ursinha
adeuringgood morning08:34
mrevellMorning!09:24
wgrantbigjools: I need some hints on writing tests for the ddeb stuff.10:07
bigjoolsheh10:07
bigjoolsyou mean you didn't write the tests first? :)10:08
wgrantNo, no, I am writing them first.10:08
wgrantI'm not that crazy.10:08
bigjoolsok10:08
bigjoolsyou need to figure out all the places in the code where we're likely to need to keep a ddeb change in parallel10:09
bigjoolsI hope this is mostly in BPPH10:09
wgrantIt should be.10:09
bigjoolsand if not we should move code into BPPH10:09
bigjoolsI suspect nascentupload will be a pig10:09
gmbDoes anyone mind if I set fire to checkwatches? No? Good.10:09
bigjoolsand PackageUpload too10:10
=== mthaddon` is now known as mthaddon
bigjoolsonce you figure that out, write tests to match your expected behaviour, in fact you can probably just alter the existing ones on BPPH for copy, override etc10:10
wgrantnascentupload gets a bit hard, as BPR is immutable but BPRs will now need to refer to other BPRs created in the same upload. I guess I'll need to create the DDEBs first.10:10
bigjoolshmmm interesting, what is making BPRs immutable?10:11
wgrantDB and Zope permissions.10:11
wgrantNo user has more than SELECT+INSERT on BPR.10:11
wgrantWhich seems sane.10:12
bigjoolswell we can change it if it makes sense10:12
wgrantActually, that can't be right.10:12
wgrantBecause queue overrides modify the BPR, don't they?10:12
* wgrant looks again.10:12
bigjoolsnascentupload runs zopeless10:12
=== mrevell changed the topic of #launchpad-dev to: This is Launchpad Development Channel | Week 1 of 3.1.10 | PQM is OPEN | https://dev.launchpad.net/ | Get the code: https://dev.launchpad.net/Getting | On-call review in #launchpad-reviews | Use http://paste.ubuntu.com/ for pastes | This channel is logged: http://irclogs.ubuntu.com/
wgrantOK, so queued is the only real thing that can UPDATE BPR. I guess uploader could be allowed to.10:14
wgrantBut anyway.10:14
bigjoolswgrant: you probably won't need to10:14
bigjoolschange the sql config I mean, the uploader won't have committed anything so you're only updating the storm copy10:15
wgrantIt will have flushed it.10:15
wgrantBecause storeInDatabase does queries of some kind.10:15
bigjoolsbah10:15
wgrantI've tested this; it is problematic in practice.10:15
bigjoolsok10:15
bigjoolsfix security.cfg then10:15
wgrantOK.10:16
wgrantNow, the tests of new nascentupload functionality involve checking state at various points over a sequence of about 5 uploads. I think it might be neatest to have it all as one big doctest, although they seem to be losing favour...10:17
jmlhello.10:17
bigjoolsyeah, I'm unsure of which is the best direction, but possibly a combo of both10:18
bigjoolshi jml10:18
bigjoolsdid you know there was a TV station called JML in the UK? :)10:18
jmlno, I didn't.10:18
jmlor maybe mrevell told me once10:19
jmlor twice10:19
bigjoolswgrant: celso was moving towards using more unit tests as they are usually more focused on the actual test conditions10:19
mrevellbigjools: To my shame, I've told him all about the products of JML.10:19
mrevellrepeatedly10:19
mrevell:)10:19
bigjoolswgrant: there is one doctest that's an end-to-end test though, but it's currently disabled10:20
bigjoolsmrevell: ha :)10:20
wgrantbigjools: Now, this nascentupload-supertest will also end up testing for correct supersedure and overriding in the course of what it needs to do. So it's not strictly nascentupload-only. Ew.10:21
bigjoolswgrant: that will be already tested somewhere (I forgot where)10:22
wgrantbigjools: So I should integrate the more specific DDEB tests alongside the existing tests for those methods?10:23
bigjoolswgrant: yes, I think so, that would be the first place I'd expect to see them I think.10:23
mwhudsonjml: good morning11:17
jmlmwhudson, hello11:17
jmlmwhudson, I just sent you an email.11:17
mwhudsonjml: so you did11:17
mwhudsonjml: thanks for the mail11:18
jmlmwhudson, np.11:18
=== statik` is now known as statik
=== jml changed the topic of #launchpad-dev to: This is Launchpad Development Channel | Week 1 of 3.1.10 | I am Zero OOPS and So Can You! http://is.gd/4fkLl | PQM is OPEN | https://dev.launchpad.net/ | Get the code: https://dev.launchpad.net/Getting | On-call review in #launchpad-reviews | Use http://paste.ubuntu.com/ for pastes | This channel is logged: http://irclogs.ubuntu.com/
bigjoolsthumper: still up?12:43
=== mrevell is now known as mrevell-lunch
wgrantbigjools: I see you are working on a branch to allow customisation of the first PPA's name. Does it still default to 'ppa'?13:01
bigjoolsno13:05
=== bigjools is now known as bigjools-phone
wgrantbigjools: That seems like a bad idea. The display names that people come up with are bad enough.13:06
jmlgary_poster, good morning13:09
gary_posterjml, good morning.13:11
jmlgmb, if I wanted to learn about how our external bug tracking system works and what it does & doesn't do, where'd be the best place for me to look?13:22
gmbjml: Well, there's actually not all that much in-depth documentation for it (I've been meaning to write some). But there are some wiki pages that are relevant. Let me find them for you.13:23
gmbjml: Also, if you want to hurt yourself. there's lib/lp/bugs/externalbugtracker/* and lib/lp/bugs/scripts/checkwatches.py. The latter is a bit of a mind-bender.13:23
jmlgmb, heh :)13:24
jmlgmb, maybe a good idea is for me to read what docs there are & have a chat w/ you sometime after.13:25
gmbjml: Hmm, turns out that my dream of relevant wiki pages was optimistic... I can only find https://help.launchpad.net/FeatureHighlights/BugWatches, which says nothing useful.13:25
gmbjml: Although, I did write this just the other day, which should tell you a bit about what's wrong with the current setup: https://dev.launchpad.net/Bugs/CheckwatchesNG13:26
jmlgmb, thanks.13:26
=== bigjools-phone is now known as bigjools
bigjoolswgrant: they are *Personal* PAs, and it can't be any worse than "ppa" ... !13:46
=== mrevell-lunch is now known as mrevell
wgrantbigjools: The old default ppa/'PPA for Some User' seems better than what people are coming up with now. There are no examples for what is appropriate.13:52
bigjoolswhy do you think it's necessary to do that though?  I'm not against it, just keen to know why.13:52
bigjoolsas an aside, I want to change the interface description to make it obvious that the name appears in the URL and the description in the GPG key13:53
wgrantI bet a lot of people are going to start naming their first PPA with their username.13:54
wgrantIs the display name used in the OpenPGP key? I thought it was 'Launchpad PPA for $owner'... but I haven't checked the code.13:54
bigjoolsfrom memory it is13:54
bigjoolsbut I've been wrong before :)13:54
wgrantUrgh. You're right.13:57
wgrantThat sounds like a bug.13:57
wgrantAs it's now shared between all PPAs.13:57
bigjoolshmmm good point13:57
=== EdwinGrubbs2 is now known as EdwinGrubbs
=== barry` is now known as barry
barrygary_poster: ping14:51
gary_posterbarry: hey are you taking today off?14:52
barrygary_poster: nope, swapping it.  how about yourself?14:52
gary_posterbarry: me too, cool!  salgado has today off though.  on call will ping you14:52
barrygary_poster: np14:52
=== barry` is now known as barry
gary_posterbarry: yo, let's start15:09
gary_posterbarry, want to go to another channel?  #launchpad-sprint or something silly?15:09
barrygary_poster: +115:10
gary_postercool15:10
=== stub1 is now known as stub
=== intellec` is now known as intellectronica
rockstarbarry, am I safe to assume there is no call today?16:12
barryrockstar: yep16:12
* rockstar has been listening to smooth jazz for too long...16:13
* jml has been listening to _The Score_ for too long.16:17
barryrockstar, abentley: https://code.edge.launchpad.net/~launchpad/launchpad/python-migration16:33
barryrockstar, abentley i'm deleting that branch, so nm ;)16:35
rockstarbarry, okay.16:35
rockstarbarry, I think abentley is off on a national holiday today.16:36
barryrockstar: oh right.  anyway the branch had a big warning that said just: MemoryError16:36
rockstarbarry, eep.16:36
barryrockstar: might be reproducible by: branching off of stable; pushing w/no stacking (or with a --stacked-on that lp will deliberately ignore); wait for the puller to bomb16:37
rockstarbarry, yeah, I could see how that might be it.16:38
rockstarI wonder if there's an open bug about that.16:38
=== andrea-bs_ is now known as andrea-bs
leonardradeuring, i have a question about bug 40212616:47
mupBug #402126: top level publications must be public <lazr.restful:New> <https://launchpad.net/bugs/402126>16:47
ubot3`Malone bug 402126 in lazr.restful "top level publications must be public" [Undecided,New] https://launchpad.net/bugs/40212616:47
mupBug #402126: top level publications must be public <lazr.restful:New> <https://launchpad.net/bugs/402126>16:47
mupBug #402126: top level publications must be public <lazr.restful:New> <https://launchpad.net/bugs/402126>16:47
adeuringleonardr: yes?16:48
leonardrpresumably you tried to publish the hwdbapplication object as a top-level object, and it didn't work16:48
leonardrwhat was the failure?16:48
adeuringleonardr: I must admit that I can't remeber the details...16:48
adeuringleonardr: give me some time to reproduce.16:49
leonardradeuring, sure16:49
leonardrrockstar, i have a question about bug 32630716:55
mupBug #326307: Need @property equivalent for the API <api> <lazr.restful:Triaged> <https://launchpad.net/bugs/326307>16:55
ubot3`Malone bug 326307 in lazr.restful "Need @property equivalent for the API" [Low,Triaged] https://launchpad.net/bugs/32630716:55
mupBug #326307: Need @property equivalent for the API <api> <lazr.restful:Triaged> <https://launchpad.net/bugs/326307>16:55
mupBug #326307: Need @property equivalent for the API <api> <lazr.restful:Triaged> <https://launchpad.net/bugs/326307>16:55
leonardrwhat is the accessor method you'd like to publish as a property?16:55
leonardri'd like an example for a doc i'm writing16:55
rockstarleonardr, looking.16:56
rockstarleonardr, hm, it's been a while since I wanted that.  Lemme look around.16:57
rockstarI'm sure I can find something.16:57
leonardrrockstar, thanks16:57
rockstarleonardr, how about IBranch.getPullURL exposed as IBranch.pull_url17:06
leonardrok, great17:06
adeuringleonardr: I believe this patch re-enables HWDBApplication as a top-level publication (with access restriction): http://paste.ubuntu.com/291685/ Results in this error: http://paste.ubuntu.com/291687/17:20
leonardradeuring: can you just except link_name from launchpad.View protection?17:26
=== EdwinGrubbs is now known as Edwin-lunch
leonardrit might be better not to show hwdb at all, but then it wouldn't show up in the wadl17:27
adeuringleonardr: any hint for how to exempt link_name from the <require> rule? A simple addition of "allow attrinbutes="link_name" leads to the error "Failed to load application: Conflicting configuration actions"17:31
leonardradeuring: unfortunately the only way i've found to do that is to restrict every field *except* the one you want to allow17:53
leonardrthere's a slight chance gary will have a better idea17:53
adeuringleonardr: that would be good. Defining the complete set of restscited attributes would not be very convenient...17:54
=== bigjools is now known as bigjools-afk
gary_posteradeuring: no, you can either arrange the interfaces to suit your needs better, or define the set, I'm afraid.18:16
gary_posterthat's something we could theoretically improve.  if you wanted to put a bug into foundations I could see how much support there is for it.  describing what you want in a clear, compelling, and general way would probably help the sales pitch when I present it to others.  Think of it as trying to sell it to your team leads. :-)18:16
adeuringgary_poster: OK; I'll add a comment to bug 40212618:17
ubot3`Malone bug 402126 in lazr.restful "top level publications must be public" [Undecided,New] https://launchpad.net/bugs/40212618:17
mupBug #402126: top level publications must be public <lazr.restful:New> <https://launchpad.net/bugs/402126>18:17
mupBug #402126: top level publications must be public <lazr.restful:New> <https://launchpad.net/bugs/402126>18:17
mupBug #402126: top level publications must be public <lazr.restful:New> <https://launchpad.net/bugs/402126>18:17
gary_posterheh, mup + ubot3` == excitement!18:17
gary_posteradeuring: it's more of a problem in the zcml spelling18:18
adeuringgary_poster: yes, maybe18:18
gary_posterIOW, we would be redoing some zope bits to try to do what you want.  I think. :-)  eh, write up what you think you want and we'll see.  ;-)18:18
mrevellNight all18:26
gary_posterbac, are you approving CPs?  Is anyone to your knowledge?18:47
gary_poster(and hi btw :-) )18:47
bacgary_poster: i am not.  my RM powers expired a while back.  you'll have to talk to flacoste18:48
gary_posterbac ok thanks.  I figured, but was worth a check.  (flacoste's out on nat. holiday afaik)18:49
* rockstar lunches19:06
=== EdwinGrubbs2 is now known as EdwinGrubbs
maxbbarry: You appear to have removed all my commentary on why particular tests failed20:16
barrymaxb: dang.  sorry for the hamfisted wiki editing.  i'll try to revert that20:17
barrymaxb: btw, you are very welcome to join gary_poster and myself on #launchpad-sprints!20:17
maxbs/sprints/sprint/20:18
barryright, sorry.  and changes reverted20:18
EdwinGrubbsbarry: hi, can you mark this mp as reviewed so I can use "ec2 land" on it? https://code.edge.launchpad.net/~edwin-grubbs/launchpad/bug-430708-registry-windmill-layer/+merge/1314820:31
barryEdwinGrubbs: done20:33
EdwinGrubbsthanks20:33
rockstarmwhudson, when you're around, I'd like to chat with you about upgrading stacked branches.20:36
mwhudsonrockstar: ok20:45
mwhudsonrockstar: i'm definitely not caffeinated enough for that yet, will let you know :)20:46
rockstarmwhudson, okay.  I'm assuming we'll have to be on the phone together for it.20:46
thumperrockstar: are you working today?21:07
rockstarthumper, why wouldn't I be?21:07
thumperrockstar: the email said it was a public holiday21:08
rockstarthumper, what email?21:08
thumperthe staffing email21:08
rockstarthumper, ah, well, I guess I should have read that then.  :)  Yes, I'm indeed working today.21:08
* thumper runs to drop off Maia21:08
mwhudsoni guess someone should look at the launchpad failures with bzr.dev at some point21:14
rockstarmwhudson, yeah, I keep wondering that.  I get those emails and wonder if maybe one of use should respond about them.21:17
mwhudson:)21:20
mwhudsonthumper did an accurate summary of the failures a while ago21:21
mwhudsonthe part that's really sigh-inducing is the plugin api versioning nonsense21:21
thumpermwhudson: skype?21:22
mwhudsonthumper: as usual i'm online21:22
mwhudsonthumper: did you try calling?21:23
wgrantIntriguing. Breaking LP OpenID is Low, I see.21:51
lifelesswgrant: ?22:22
wgrantlifeless: The bug I diagnosed yesterday was marked as a duplicate, and the original became Low.22:37
thumperrockstar: I guess we are done then22:38
rockstarthumper, oops. I didn't mean to click the hang up button.22:38
thumper:)22:38
thumpernp22:38
rockstarStupid touchpad sensitivity...22:38
lifelesswgrant: meep :(22:44
lifelesswgrant: bug number?22:44
lifelessfound it22:47
* mwhudson -> into town, back online in a bit23:52

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!