StevenKlifeless: I guess we can look forward to drowning in scriptactivity warnings about oops-prune when your branch gets deployed, then. :-(00:00
lifelesswhy? we will delete its entreis00:01
StevenKmwhudson: My change only means bin/ec2 won't use AMIs created by you -- that's all.00:04
mwhudsonStevenK: oh right00:04
mwhudsonStevenK: sorry for the noise00:04
lifelessthis is perhaps a little to eager, as poolie just added himself00:07
lifelessI certainly think non-staff need to be removed00:07
StevenKmwhudson: It's perfectly fine. :-)00:09
StevenKlifeless: Well, to be fair, I JFDI'd, rather than communicated about it.00:09
lifelessI think thats fine00:11
lifelessjust suggesting perhaps it needs a tweak00:11
poolieStevenK, i was only suggesting to remove non staff00:23
poolieConfigurationConflictError: Conflicting configuration actions00:23
poolie  For: ('protectName', <class 'lp.services.salesforce.proxy.Voucher'>, u'__str__')00:24
poolie    File "/home/mbp/launchpad/lp-branches/work/configs/development/salesforce-configure-normal.zcml", line 16.4-18.400:24
poolieis this my fault?00:24
wgrantWhat did you change?00:24
pooliei exposed a new method on bug00:25
poolienothing to do with sf00:25
wgrantYour indentation is buggy, but agreed.00:26
wgrant295 of the diff00:26
wgrantYou have a two-space indent.00:26
poolieso does the code i copied :)00:27
pooliebut i will fix it00:27
wgrantSo it does :(00:27
poolieam i allowed to put blank lines around those declarations?00:28
pooliei think it would be more readable00:28
wgrantVWS is discouraged but not forbidden, unless lifeless sees you doing it.00:28
pooliei thought it was required between class members?00:28
pooliebetween _methods_00:29
poolieaccording to pep800:29
wgrantBetween methods, sure.00:29
wgrantThat's not discretionary VWS.00:29
wgrantDiscretionary VWS is strongly discouraged.00:29
pooliei'll leave it then00:30
poolieindents are fixed00:30
pooliecould you read the diff for me?00:30
wgrantWhat produces that error?00:30
wgrantStarting a dev appserver? Starting the test suite?00:30
poolierunning the tests00:34
wgrantIs that the right branch?00:36
wgrantIt's an odd name.00:36
poolieis the branch in which that error occurred00:37
wgrantRight, but it's in a directory named 'work', which suggests it's either old or wrong or polluted with bad ZCML or config symlinks or something like that.00:37
poolieit's a colo workspace00:37
pooliei have just one tree00:37
lifelesshuwshimi: I'm glad to see you writing up UI patterns; I'm a bit surprised that we don't have such things already00:44
huwshimilifeless: Yeah, it's something I've been meaning to do for ages :)00:45
lifelessheh, I know that feeling :)00:45
rick_hhuwshimi: link?00:47
huwshimirick_h: https://dev.launchpad.net/UI/Patterns00:47
huwshimirick_h: It's not much more than a stub at the moment00:47
rick_hcool, I was looking for something like that the other day so very cool00:48
rick_hI think it was http://www.amazon.com/Designing-Web-Interfaces-Principles-Interactions/dp/0596516258/ref=sr_1_10?ie=UTF8&qid=1321231747&sr=8-10 that had some good things about creating documented sets of ui "modules" and describing interactions and such00:49
huwshimiI certainly would like to come up with a pattern for how to write that kind of documentation too.00:51
rick_hsec, checking if I got it on kindle, pdf, etc00:51
=== lifeless changed the topic of #launchpad-dev to: https://dev.launchpad.net/ | On call reviewer: - | Critical bugtasks: 275
lifelesswho is OCR today ?01:10
lifelessI want the bot to know this01:10
StevenKlifeless: And how does the bot deal with sick days/days off, etc? :-)01:36
pooliehi huw, i'm happy to see that too01:38
pooliehuwshimi, i do wonder a bit how much of that is describing the current state vs endorsing it01:42
poolieeg green links01:42
pooliewallyworld_, could you reread https://code.launchpad.net/~mbp/launchpad/678090-affected-count/+merge/81108 for me some time?01:42
* wallyworld_ looks01:42
wallyworld_poolie: looks like a nice change. are feature flags reset after each test? i normally use "with FeatureFixture ". perhaps also the flag name could be declared as a constant02:00
wallyworld_poolie: the tests all seem to set the fflag. are there any tests without the fflag?02:01
pooliethere are thousands of tests without the flag :)02:02
pooliesome of them covering this02:02
pooliei don't have any tests that assert what the default flag value is02:02
poolienot sure that would be worthwhile02:03
mwhudson"are tests reset after each test" -- yes02:03
poolieflags too02:03
mwhudsonif not, i'm coming after you with a rusty pickaxe :-)02:03
mwhudsonah yes02:03
pooliethat would be insane if not02:04
pooliewallyworld_, re making it a constant02:05
pooliemaking it a constant pointing to a string would feel kinda half baked i think02:05
poolieperhaps the features should change to be really registered as objects not tuples02:05
pooliethen the code can mention the object02:05
pooliewhich would make it harder for them to be orphaned02:06
pooliei guess for out of process lookups and maybe tal we would still use just a symbolic name02:06
pooliecould do that separately though i think02:06
wallyworld_poolie: i was just talking about the tests02:07
wallyworld_poolie: i think i got confused. the flag seems to be to used to revert to the old behaviour? ie by default, without the flag, dupes are counted, which is the new behaviour?02:08
wallyworld_hence wouldn't all the 1000's of  existing tests need to be changed?02:09
wallyworld_to count differently02:09
pooliethis is a 'get out of jail' that restores the old less useful behaviour if it turns out this is too expensive02:09
pooliewallyworld_, there were no previous tests about counting affected users across dupes02:10
wallyworld_yes,  but my point is that i would have expected to see the flag used in the "1000's of existing tests" to make them work properly02:10
poolienone of the affectsmetoo tests create dupes02:10
wallyworld_ah ok02:10
poolieaside from the ones i adedd02:10
wallyworld_so, back to my original question :-) shouldn't some of the new tests added be with and withou tthe fflag?02:11
pooliebut the other tests reassure me that my change hasn't broken the bugs page when you file dupes, look at dupes anonymously, etc02:11
poolieoh, like i should copy eg test_other_users_affected_count and do it with and without the flag?02:11
pooliemaybe i should02:11
wallyworld_yes, otherwise you are not testing the new functionality02:11
pooliesorry no eg test_counts_affected_by_master02:12
pooliemost of the tests do try the new behaviour02:12
pooliethey explicitly set the flag to '' ie false, ie 'not disabled'02:12
wallyworld_ah, i misread that. because normally setting the fflag to *anything* equates to true, my eyes misread the tests02:13
wallyworld_i missed the ''02:13
wallyworld_whenever i see a fflag set, my mind assumes it is being set to a value which evaluates to true02:14
wallyworld_poolie: so do you really need to set the flag to ''?02:15
pooliei want to also do a followon which makes them properly typed02:15
poolieie if its declared bool, you get a real bool02:15
lifelessI am skeptical02:16
pooliewallyworld_, i just thought it was more explicit to say "this is the behaviour i expect when this is enabled"02:16
lifelessI think that will make them more fragile.02:16
wallyworld_poolie: what i mean is, i the default is false, why have effectively a noop at the start of the test?02:16
wallyworld_had the opposite effect on me :-)02:16
poolielifeless, one concrete probelm is that they suffer from bool('false') == True02:17
wallyworld_but it was my misreading02:17
lifelesspoolie: OTOH they can't raise TypeError02:17
wallyworld_poolie: fwiw, most of the other tests i see and also write myself do not set the fflag at the start of the test if the default is required02:18
lifelesspoolie: I'd be happy with the editor page being taught that a bool entry has a 0 length string to indicate False02:18
poolieyou're saying you don't want bad input data to cause a typeerror at the point the flag is used?02:18
poolieit will be less fragile if it's just given some arbitrary interpretation, as it is at the moment?02:19
lifelessamongst other concerns yes; I think there is a good leanness is having it be None-or-a-string02:19
lifelessits a clear common substrate02:19
lifelessNo matter what crap gets into the table, that substrate won't fail02:20
wallyworld_poolie: i've already had this argument with lifeless :-)02:20
lifelessif you have a type system, you, by definition need to handle date entered with one type being read out with a different type02:20
lifelessbah, my commas are in terrible places there02:21
poolielifeless, i'm pretty sure there's already some code that does cast-to-bool and some that does =='on'02:21
pooliethis seems poor02:21
lifelessI'm happy - delighted even - for the editing UI - to have such complexity in it, and I'm happy for helpers to work with what the substrate returns.02:21
pooliei agree with being robust about invalid stuff in the table02:22
mwhudsonis there some magic to installing launchpad-dependencies on oneiric?02:24
mwhudsoni get launchpad-dependencies : Depends: libsvn-dev but it is not going to be installed02:24
poolieyou need ppa:bzr too02:24
poolieand maybe universe?02:24
mwhudsonah, i'll try with the bzr ppa02:25
lifelesspoolie: My constraints are:02:25
poolieno it's in main02:25
mwhudsoni have universe enabled02:25
lifeless - the substrate Must Not Fail02:25
lifeless - Ever02:25
lifeless - single pages can die, thats tolerable02:26
mwhudsonE: Unable to correct problems, you have held broken packages.02:26
lifeless - the editing page -should- never fail, but we could tolerate a failure there02:26
mwhudsonthanks apt, how about telling me which02:26
wgrantmwhudson: sudoa pt-get install libsvn-dev?02:26
wgrantYou have to drill down manually.02:27
wgrantBut it'll eventually tell you.;02:27
pooliethanks for being clear about the constraint02:27
pooliei agree with that02:27
mwhudsonwgrant: a seemingly endless excercise in dependency chasing02:27
wgrantmwhudson: Yes02:27
mwhudsonisn't sorting that out what apt is for? :)02:27
lifeless - the type system has to be clearly a hint - the underlying data -is- untyped and I don't want folk to forget that02:27
lifeless - fin02:27
pooliei guess you can try aptitude02:27
lifelessIf you want to design something up that works within those constraints, I'll be happy to be a sounding board and/or review the design for you.02:28
poolienot today02:28
pooliei will take you up on it later02:28
lifelessPersonally, I wouldn't do it. I'd write helpers that work on the substrate rather than changing the substrate.02:28
mwhudsonwgrant: ah02:29
mwhudson krb5-multidev : Depends: libkrb5-3 (= 1.9.1+dfsg-1ubuntu1.1) but 1.9.1+dfsg-1ubuntu2.1 is to be installed02:29
pooliei was just going to add a function get_flag_whatever(name)02:29
mwhudsonwhich is a bit strange02:29
pooliedoes the casting to the type declared for that flag02:29
* mwhudson disables -proposed02:29
pooliethis is what most callers want02:29
poolieit can squelch errors if we want02:29
StevenKmwhudson: arch-indep has been published before arch-dep02:29
pooliei'm not planning to change the db or anything02:29
poolieseparately, in the editor, things that are bools, which seems to be the majority, can be shown as such02:30
mwhudsonStevenK: ah, so the archive/my mirror is broken?02:30
StevenKI'd check your mirror first02:30
StevenKBut the pulse may have happened between publications as it were02:30
wgrantStevenK, mwhudson: Except that that's not meant to happen any more.02:30
StevenKThere is that, too02:30
wgrantThat's why the publisher takes 90 minutes.02:30
StevenKPerhaps it doesn't work over pockets02:30
wgrantThat shouldn't matter, since apt should see the right version in some other pocket.02:31
wgrantAnd pick that.02:31
mwhudsonnz.archive.ubuntu.com does not have a record of working 100% of the time02:31
StevenKAh, citylink02:31
StevenKHeh heh02:31
lifelessmwhudson: the US mirror works pefectly :P02:31
poolieif you add the real archive at the end of sources.list02:31
poolieit will read from your nearby mirror if possible and otherwise fall back02:32
ajmitchStevenK: you say that like it's a bad thing :P02:32
StevenKajmitch: But it is? :-P02:32
mwhudsondisabled -proposed and switching to archive.ubuntu.com and still the same nonsense02:33
lifelessmwhudson: 'lxc-create'02:33
wgrantmwhudson: You haven't held krb5-multidev or something?02:34
wgrantmwhudson: apt-get install krb5-multidev02:34
wgrantWhat does it whine about?02:34
mwhudson krb5-multidev : Depends: libkrb5-3 (= 1.9.1+dfsg-1ubuntu1.1) but 1.9.1+dfsg-1ubuntu2.1 is to be installed02:34
mwhudsonwgrant: as before02:34
wgrantapt-cache policy krb5-multidev02:34
mwhudsoni do have some pinning, but no holding afaict02:34
mwhudsonwgrant: http://paste.ubuntu.com/737890/02:35
wgrantWell then.02:35
wgrantWhy is security pinned over updates?02:36
wgrantThey're naturally 500.02:36
mwhudsonbecause i followed some stupid instructions i suspect02:36
mwhudsonto whit: https://wiki.ubuntu.com/Testing/EnableProposed02:36
wgrantThe first sign that something was up should have been the pre-Lucid screenshot.02:37
wgrantThat's never a good sign.02:37
StevenKmwhudson: 'narsil' ?02:37
mwhudsonStevenK: lotr/tolkien references in machine names seemed like a good idea a decade ago02:37
wgrantmwhudson: So, pin proposed below 500 if you want, but I'd recommend not pinning up updates/security.02:39
mwhudsonwgrant: yeah, i deleted the !proposed sections02:39
wgrantAlthough I just run with -proposed unpinned. It's hardly worse than running a dev series...02:39
wgrantThat reminds me, I must upgrade to precise soon.02:41
lifelessright, tomorrow I write the untestable launchpadlib glue, and then we have fast oops pruning done.03:21
lifelessone more tick in the box.03:21
huwshimiok, tal hackers, in this paste: http://paste.ubuntu.com/737935/ the function "view/style_for_add_milestone" returns a style, instead I want it to return a class. I have modified the function, but I'm not sure how to modify the tal to return the class when there are already classes on the <a>.04:16
pooliehuwshimi, i would guess you need to define it as a python expression that returns the whole string04:23
poolieeg putting the other classes into style_for_add_milestone04:24
pooliewallyworld_, so what was your overall verdict?04:27
mwhudsonhuwshimi: you can say things like tal:attributes="class string:bish bash bosh ${view/other_classes}"04:28
mwhudson(if that's what you're asking for, not really sure)04:28
poolieoh right04:28
wallyworld_poolie: looks ok to me, my misreading of the fflag aside. my personal preference is not to set the flag if the test is expecting the default behaviour04:29
pooliei have mixed feelings about the test04:30
pooliei think i will leave it for now04:30
wallyworld_ok. the actual business logic looks ok i think04:31
huwshimimwhudson: Thank, that worked perfectly04:50
mwhudsonhuwshimi: np04:50
poolieStevenK, if nothing falls over after that ec2 landing i will update the image later in the week05:01
huwshimiA pretty trivial CSS review if someone's keen: https://code.launchpad.net/~huwshimi/launchpad/style-removal-one/+merge/8209605:27
wgrantlol bigpond05:30
wgrantAfternoon bigjools.05:30
bigjoolscrappiest internets evar05:30
StevenKWelcome to Australia05:30
bigjoolsexcept when I was in Brazil and the connection went every time it rained05:30
StevenKAnd it's bigpong, so it's even worse.05:31
bigjoolshey, it's free with the apartment05:31
StevenKThere is a reason for that.05:31
* bigjools looks for slab of meat to throw on bbq05:32
StevenKYou could always visit wallyworld_ and suffer through using TPG.05:32
bigjoolsI was at his place last night05:32
StevenKBrisbane is only 31, disappointing05:33
StevenKIt hit 38 here today05:33
wgrantIn November? I doubt that.05:33
wgrantExcept you're right.05:33
bigjoolsnot that I'd know inside the airconned car and airconned shopping mall05:33
StevenKDon't doubt me, I'm bloody melting05:33
bigjoolsis this all you guys talk about before EU wakes up? :)05:34
StevenKNot usually05:34
bigjools38 is not pleasant05:35
StevenKAnd no air-con here05:35
* wgrant submits a new definition of "unwise" to the dictionary.05:36
bigjoolsdoesn't your equipment go bleargh?05:36
bigjoolswgrant: s/unwise/tight fisted/  :)05:36
StevenKbigjools: Rented apartment, I have no choice05:36
wallyworld_StevenK: tpg is good :-P05:43
StevenKs/\(good\)/\1 for nothing/05:43
wallyworld_have you ever used it?05:43
StevenKI have used both their dial-up and broadband service, yes05:44
StevenKBut I have not been a customer of theirs05:44
wallyworld_for me it's fast and cheap05:44
wallyworld_never have drop outs05:44
StevenKWait until you call support05:44
StevenKHeh heh05:45
wallyworld_why would i wan tto do that?05:45
wallyworld_never had a need to in over 4 years05:45
StevenKJust because you never have had to doesn't mean you won't need to.05:45
wallyworld_use them for mobile too. heaps of data and calls for not much cash05:45
wgrant$ grep wallyworld.*tpg.*Ping\ timeout irclogs/Freenode/#launchpad-dev.log | wc -l05:46
wallyworld_why should i care about that latency? i don't play games05:46
StevenKwgrant: Pointless use of wc -l :-P05:46
* StevenK waits for "Silence!"05:47
wgrantI didn't see anyone talk about latency...05:47
StevenKA ping timeout is not latancy05:47
wallyworld_sorry, i misread05:47
wgrant(well, unless >5min latency still counts as latency)05:47
wgrantWhich it probably doesn't, unless you're on Mars.05:47
wallyworld_but won't it still timeout if i suspend the laptop to move rooms etc05:48
wallyworld_so those figures are bogus05:48
wgrantAh, if you run your IRC client on a laptop, indeed.05:48
wallyworld_i only use my laptop05:48
wallyworld_i honestly have no issue with tpg05:49
wallyworld_bigjools: found a slab of meat yet?05:49
wgrantpeople I know who used to use them got terrible speeds and dropouts.05:49
bigjoolswallyworld_: oh yes05:49
StevenKAnd some beer?05:49
wallyworld_wgrant: my speed is around 10Mbps and I'm a fair way from the exchange05:50
bigjoolsand some sparking rose05:50
StevenKIs it room temperature beer? :-P05:50
bigjoolsStevenK: who tf drinks that!05:50
bigjoolscellar temperature is more ideal05:50
StevenKThe British, if you believe the jokes05:50
wgrantwallyworld_: Not bad.05:50
wallyworld_i guess i'm one of the lucky ones05:51
bigjoolsat least we have beer, rather than watery piss05:51
wallyworld_you mean Fosters? you lot own that now05:51
poolie38C beer would certainly be unpleasant05:51
bigjoolsI am proud to say I've never drank Fosters05:51
poolieStevenK, i bought a Vornado fan last year when my ac was broken05:51
bigjoolslager.... <shudder>05:51
poolieit is very quiet and compact05:51
poolieswitching to using only my laptop helped tooo05:54
StevenKbigjools: Except the link in that article is a 40405:54
bigjoolsStevenK: ha!05:54
pooliealso "How to Work from Home Without Going Insane"05:54
* StevenK stabs lg.com05:57
bigjoolswallyworld_: I was up at Chapel Hills today06:06
wallyworld_how'd it go?06:06
bigjoolsgreat school06:06
bigjoolsnice area06:06
bigjoolssome lovely houses!06:07
StevenKbigjools: s/s$// doesn't match anything :-P06:07
bigjoolsStevenK: silence06:07
wallyworld_bigjools: yes, it is a great area. Mt Cootha being so close is a bonus for sure too06:08
bigjoolswallyworld_: yeah, some great trails06:08
* wallyworld_ has to take the poor dog to the vet now to get his tail looked at06:09
bigjoolswallyworld_: oops06:10
bigjoolsso, I buy a $19 recharge for a prepay SIM and end up with $50 credit. This is odd.06:10
wallyworld_bigjools: we can't stand the blood being cast everywhere when he wags it. it looks like a scene from CSI06:10
StevenKbigjools: Sounds right.06:10
wallyworld_bigjools: that $50 is an inflated value06:11
bigjoolson the plus side, 1GB of data06:11
wallyworld_you really need to look at how many minutes talk time it buys you06:11
wallyworld_it's very confusing here in Oz06:11
bigjoolsor how much data it gives me so I can use Skype :)06:11
wallyworld_yep :-)06:11
StevenKbigjools: Which provider?06:11
* wallyworld_ goes to round up the dog06:11
bigjoolsStevenK: Virgin Mobile06:12
StevenKAh, neat.06:12
StevenKThere is a "Virgin Usage Meter" app on the market too06:12
bigjoolsoh cool06:12
bigjoolswell, the phone has built-in meters06:12
StevenKThat one screen-scrapes from Virgin's site just in case the two don't agree06:13
bigjoolsStevenK: I just realised the humour in "Virgin Usage Meter"06:13
StevenKAre you going to rate it down for not doing what the app title suggested? :-P06:14
wgrantHmm, now, the ultimate question.06:45
wgrantDid he suspend his laptop, or was it Telstra...06:46
bigjoolsBigpong is up and down like a whore's knickers07:19
wgrantLike I said.07:23
=== wgrant changed the topic of #launchpad-dev to: https://dev.launchpad.net/ | On call reviewer: - | Critical bugtasks: 277
stubMust suck when your cow-orkers in the third world have better 'Net than you do.07:28
stubbigjools: You on holiday or is this a more permanent move?07:30
bigjoolsstub: yes07:30
bigjoolsstub: holiday for now...07:31
jtvstub: don't forget—cow-orkers _in disaster zones_ in the third world.07:33
nigelbbigjools: Where are you? :)07:34
nigelbi.e. where are you vacationing07:34
nigelbjtv: cow-workers?07:34
bigjoolsnigelb: Brisbane07:34
nigelbbigjools: Neat. :)07:34
jtvnigelb: yes, people who ork cows.07:35
bigjoolsI ate a cow last night07:35
nigelbis that a figure of speech?07:36
bigjoolsnot at all - you should have seen the size of the steak I had07:36
jtvnigelb: how so?  Julian ate a cow last night, and presumably used a nife and an ork.07:36
jtvIt all makes so much sense.07:36
nigelbjtv: You are milking this for all its worth aren't ya? :D07:36
bigjoolsjtv: BA-DUM-TISH!07:36
nigelbbigjools: Gah, Now I want steak.07:36
jtvnigelb: Ooh I see what you did there, with the cow theme!07:37
nigelbThis is turning into a cow-themed punland.07:37
stubjtv: Well, I was thinking on New Zealand. But I guess it holds true here too ;)07:38
nigelbI remember a joke sometime back to the likes of "Next time you take a vacation, go to an island with poor internet conectivity. But you already live in New Zealand"07:39
nigelbjtv: We'll be in closer time-zomes next weekend!07:41
jtvOh, right!07:41
nigelbI'm sad I can't make it to Bangkok yet :|07:41
=== rvba changed the topic of #launchpad-dev to: https://dev.launchpad.net/ | On call reviewer: rvba | Critical bugtasks: 277
wgrantstub: Thanks.08:30
wgrantlifeless: Want to recheck https://code.launchpad.net/~wgrant/launchpad/observer-db/+merge/81104?08:30
adeuringgood morning08:52
jtvhi adeuring08:53
adeuringhi jtv!08:53
jtvStevenK, wgrant: I'm trying to get back into bug 876594.  That unwanted rejection is generated by UploadHandler, right?08:54
_mup_Bug #876594: rejected builds for synced packages send mail to Debian maintainer <derivation> <qa-untestable> <regression> <Launchpad itself:In Progress by jtv> < https://launchpad.net/bugs/876594 >08:54
jtvhi mrevell09:19
nigelbMorning mrevell09:20
mrevellHey there09:20
=== danilo_ is now known as danilos
=== allenap` is now known as allenap
stub1rvba: https://code.launchpad.net/~stub/launchpad/oops/+merge/8211910:34
=== stub1 is now known as stub
rvbastub: k10:34
mvohttps://dev.launchpad.net/PatchSubmission <- this page says that step #1 is to go to #launchpad-reviews, but that channel is apparently no longer in use - could someone update the wiki please (it looks like I don't have permissions to do so)10:47
rvbamvo: sure, I'll do it.10:48
mvoso I worked on lp:~mvo/launchpad/maintenance-check-precise to make sure that the "apt-cache show pkgname" "Supported: 18m" info is accurate for precise, this required some tweaks to the code. I wonder if someone could run the testsuite for me on this branch (or trigger it in ec2 or somesuch). I have a local LP VM, but its really outdated. my QA for the code was simply to ensure that for lucid it produces the same output as it did before10:51
rvbamvo: Let me have a look.10:57
rvbamvo: could you please write a MP explaining the rationale for this?10:59
mvorvba: sure, I will note in the MP that I have not run the testsuite locally yet11:01
rvbamvo: cool.11:02
mvoI put it here: https://code.launchpad.net/~mvo/launchpad/maintenance-check-precise/+merge/8212511:21
mvoand have some lunch now11:21
* mvo waves11:21
mvoI set it to WIP as long as no testsuite was run against it11:22
rvbamvo: you can set the status as "Needs Review". We usually run the test suite (in ec2) after we've reviewing the mp.11:23
jelmerlifeless: hi11:51
jelmerhi rick_h12:01
jtvBuildbot just failed: hung twisted thread or something along those lines, in the branch puller.12:09
jelmerfun :(12:11
=== benji changed the topic of #launchpad-dev to: https://dev.launchpad.net/ | On call reviewer: rvba, benji | Critical bugtasks: 277
rvbaMorning benji.13:07
benjigood morning rvba13:07
mvorvba: aha, thanks! I set it back to needs-review now14:01
deryckMorning, all.14:02
deryckHi, rick_h_.  Welcome to Launchpad!14:03
deryckofficially and all ;)14:03
deryckrick_h_, do you have access to the Canonical wiki now?14:04
mrevellOh, hey rick_h_, is it your first day? Welcome!14:04
rick_h_yes, been working through it this morning14:04
rick_h_couple of glitches, but progressing14:04
deryckrick_h_, can you point me at the doc you're working through?14:04
abentleyrick_h_: Good morning, and welcome to the squad.14:07
rick_h_abentley: thanks14:07
jcsackettmorning, folks.14:07
deryckMorning, jcsackett.14:09
rvbabenji: I'll review jtv's mega lint branch…14:15
=== salgado is now known as salgado-lunch
benjirvba: sorry, I was AFK visiting the alter of coffee in my kitchen; yeah that sounds good14:22
jtvrvba: thanks for picking it up14:22
nigelbbenji: wow, you have an altar for that? Neat.14:24
flacosterick_h_: welcome to Launchpad!14:26
deryckDoes anyone know what's up with db-devel failed build?  buildnot is getting on my last nerve.14:26
jtvallenap: looking at test_stalePidFile.  You've got a comment in there saying it's safe as long as pids don't wrap around… did you mean it's safe as long as pids don't get reused?14:49
rick_h_thanks flacoste14:51
allenapjtv: I'll take a look...14:53
allenapjtv: Yes, but from what I could discover, Linux doesn't reuse pids until the counter wraps.14:54
jtvallenap: upon consideration I think that's what's meant.  TBH I think the world would fall apart quickly if pids got re-assigned quickly.  :)14:54
allenapjtv: Indeed! :)14:55
jtvCall it our dirty little secret.14:57
jtvStrange thing is, I'm suddenly getting a DeprecationWarning there instead of a UserWarning about deprecation.14:57
deryckjelmer, hi.  Have you seen the db test failures?  Are these related to your work?15:13
jcsackettsinzui: are you free to mumble?15:18
jelmerderyck: no, which failures?15:25
jelmerderyck: I haven't landed anything in quite some time15:25
deryckjelmer, flacoste thought it might have been your bzr-builder updates.  but now he suspects his own work. ;)15:26
deryckjelmer, so un-ping.  Sorry. :)15:26
flacostederyck: i don't suspect my change at all15:29
flacosteit is something bzr related15:29
deryckflacoste, ah, sorry.  I misunderstood.15:29
flacosteTestPullerMasterIntegration' object has no attribute '_lock_actions'15:30
flacostelook likes an API change15:30
flacostecould be poolie change also15:30
flacostethat run upgrade before tests15:30
jelmerhmm, I think I've seen that failure before as part of a spurious test failure15:34
jelmerI'll have a look15:34
jelmerderyck: do you know for how long this has been failing?15:34
jelmeror is it just a one-off thing?15:34
deryckjelmer, we've only had this one failure like this in the builder.15:35
flacostederyck: but the these tests always fail for you locally, no?15:36
sinzuijcsackett, https://code.qastaging.launchpad.net/private-project-test/15:37
deryckflacoste, I'm trying against current db-devel now to confirm.  I was on an old rev, before the last 4-6 revs, to see a clean state....15:37
deryckflacoste, and was failing then.  but trying on current tip now.15:37
jelmerflacoste: thanks15:38
jelmerI'm pretty sure I've seen this as part of a spurious test failure before15:38
jelmernot sure what the background of it is though, it seems pretty odd for a simple operation like that to take more than 20 seconds15:39
mrevellHey sinzui! On a project-group overview page, should all member projects be listed? If yes, is there a delay between adding a project to a group and it showing up?15:41
nigelbthe launchpad planet could use some love :)15:42
nigelbIts a pain to look at now :)15:42
sinzuimrevell, I think there is a 15-60 minute delay. The list is cached since it is prone to timeout15:43
mrevellalo ^^^ does that help?15:43
mrevellthanks sinzui15:43
flacostei still have this OperationalError: fe_sendauth: no password supplied15:45
flacostewhenever i attempt to run DatabaseLayer tests15:45
deryckflacoste, ok, so these tests do pass for me locally running latest db-devel.15:49
deryckflacoste, should we force the build?15:49
flacostederyck: might as well15:50
flacosteor disable the tests15:50
deryckflacoste, ok.  thanks.  if it fails again, we can disable.  cool?15:51
deryckok, thanks15:51
deryckalright, build force.  land 'em if you got 'em.15:52
=== salgado-lunch is now known as salgado
=== salgado_ is now known as salgado
=== deryck is now known as deryck[lunch]
=== beuno is now known as beuno-lunch
=== rvba changed the topic of #launchpad-dev to: https://dev.launchpad.net/ | On call reviewer: benji | Critical bugtasks: 277
flacostewhere are the deployment reports nowadays?17:14
flacostervba: how is the qa of rev 14289?17:15
rvbaflacoste: I'm on it.17:15
flacostervba: great, thanks17:15
rvbaflacoste: qa-ok17:17
=== Ronnie1 is now known as Ronnie
=== deryck[lunch] is now known as deryck
=== beuno-lunch is now known as beuno
lifelessgood morning18:15
jtvmorning lifeless18:32
* jtv double-takes18:32
jtvHmm… morning?18:32
jtvStupid timezones.18:33
Ronniewhat version of python-launchpadlib is advised. the 1.6.0 from the 10.04.3 repo or the 1.9.9 from pip ?18:42
lifelessnewest you have available is the general recommendation18:46
flacostelifeless: would you have any idea where to start debugging http://pastebin.ubuntu.com/738453/18:46
flacostelifeless: i suspect i missed a step in the pgbouncer transition18:46
lifelessflacoste: local?18:47
flacostei'm basically unable to run any tests invovling the DB18:47
lifelesswell, the fixture sets up phbouncer itself18:47
lifelessthats not using pgbouncer, I don't think.18:47
lifelessI suspect your pg install isn't setup correctly18:47
flacosteor it might be the oneiric upgrade18:47
flacostei don't have anything else than LP in pg18:48
lifelesswhat pg version are you runniung with ?18:48
lifelessor yeah try the initial setup script again18:48
lifelessdoes psql launchpad_ftest_template work ?18:51
flacostelaunchpad-database-setup seems to have done the trick18:51
flacostethe tests are now running18:51
benjilifeless: I'm wondering why my DB branch (https://code.launchpad.net/~benji/launchpad/bug-877195-db) isn't showing up on http://lpqateam.canonical.com/qa-reports/deployment-db-stable.html and I suspect it's because I forgot to target the original MP to db-stable and resubmitted the MP after you and stub had approved the other and I haven't sought you to approve the new one.  Can you confirm that theory?19:30
lifelessbenji: mp's aren't needed for the deployment report19:31
lifelessbenji: if they are present they are used as an additional source of bug links, is all19:31
lifelessbenji: the deployment report depends on:19:32
lifeless - the instance running19:32
lifeless - the code being deployed to said instance19:32
lifelesshas your code reached db-stable, and is the running staging revno >= the revno your branch landed in?19:32
benjilifeless: yes to having reached db-stable, and no to the second (https://staging.launchpad.net reports the revision before mine)19:35
benjiso, I assume that means I need to ask/wait for staging to be updated19:35
lifelessas your rev may be the reason its not deployed (if the patch fails)19:38
lifelessthe logs are on devpad to let you check this19:38
abentleyderyck: has adeuring's feature-flag-in-javascript stuff landed?19:42
adeuringabentley: I received a positive PQM message at 1900 UTC19:42
adeuringerm, 1800 UTC19:43
abentleyadeuring: Ah, well if it was today, I'll need to merge devel.19:43
adeuringabentley: right, it landed just recently -- testfix hassles...19:44
deryckabentley, what adeuring said ;)  Sorry was otp.19:46
abentleyadeuring: That adds a list of "beta_features" to the IJSONRequestCache?19:50
adeuringabentley: this list was already been added, but the list is empty by default19:51
abentleyadeuring: I see.  I was hoping for a list of active feature flags, like that at the bottom of the page: Features: {'baselayout.careers_link.disabled': None, 'hard_timeout': None,19:52
adeuringabentley: that's what you get19:52
abentleyadeuring: where do I get it?19:53
adeuringbut a list, not a dict19:53
adeuringabentley: enable the feature flag and view any bug listing19:53
abentleyadeuring: So this feature of including feature flags itself is enabled by a feature flag?19:53
adeuringabentley: well, to acutally see a non-empty list, you need my recent branch19:53
adeuringabentley: no. Take my branch that landed two hours ago, and enable the LEP feature19:54
abentleyadeuring: And this shows all feature flags, unlike the empty list already present?19:55
adeuringabentley: it is a bit more complicated:19:55
adeuring1. the view needs a non-empty property beta_features, as BugTaskSearchListingView19:56
adeuring2. you must enable this feature19:56
abentleyadeuring: So if beta_features is non-empty, it lists all active feature flags?19:56
adeuringabentley: let's mjumble, that's easier19:57
benjilifeless: since I can find no reference to my revision (11143) in the log (sourcherry/2011-11-11-staging_restore.log), I assume no attempt has been made to deploy it to staging; what is the best course of action to get it deployed?19:57
lifelessbenji: how long ago did it pass buildbot ?19:58
lifelessbenji: if < 1 hour, wait.19:58
lifelessif > - check with los a19:58
benjilifeless: it got buildbot love on Friday, so I'll talk to a losa about it19:59
benjiooh, I found something lifeless: the log for deploying the revision (sourcherry/2011-11-12-staging_restore.log) is very unhappy but in a way that is meaningless to me.  The first (real) error seems to be this bit: https://pastebin.canonical.com/55752/20:06
deryckabentley, adeuring -- I'm going to change work locations here in about 20 -30 minutes and will be offline for 30 minutes while I do.20:12
abentleyderyck: ack20:12
rick_h_benji: can you do me the fav of checking https://code.launchpad.net/~rharding/launchpad/bugfix_814696/+merge/8221420:14
rick_h_and hand holding and all that new guy stuff :)20:14
benjirick_h_: gladly20:15
abentleyderyck: http://people.canonical.com/~abentley/visible-information.png20:15
lifelessbenji: ahha20:16
lifelessbenji: slony 1 vs slony 2 restore, I bet20:16
deryckabentley, oh wow.20:16
deryckabentley, is that chrome or firefox?20:16
abentleyderyck: FF20:16
lifelessbenji: this is the first staging restore post slony 2 AFAIK20:16
abentleyderyck: It actually looks kinda cool, but...20:16
deryckabentley, ok, ack.  I'll take a look here shortly.20:16
benjilifeless: interesting20:16
lifelessabentley: if that was deliberate, wow :)20:16
benjirick_h_: looks good; we're not friendly to the tab characters though, so you should replace it with spaces (you can clean up the other tabs in that general area while you're at it if you want)20:21
rick_h_ok, will do20:22
benjirick_h_: I can't decide if that usb-only monitor on your blog is awesome or insane; possibly both20:26
rick_h_benji: I'll bring it in Jan20:26
rick_h_the insane part is now I'm thinking about getting another :)20:26
benjirick_h_: unfortunately I won't be there in January, fortunately I'll be distracted by having a new baby20:27
rick_h_oh, congrats20:27
rick_h_know how that goes20:27
rick_h_ok, so I removed all the tabs, found one in another file and fixed that for the why not factor20:28
rick_h_where would I go from here?20:28
rick_h_while the diff is regen'ing20:28
benjisounds good, just push the tweaks and you can mark the MP as approved20:28
benjiif there were more substantial changes you could ask for another review20:29
rick_h_you marked it as "Approve" so what am I marking as approved?20:29
abentleyadeuring: So it turns out that list is actually a list of feature flags that have been used, rather than all feature flags.20:30
abentleyadeuring: the idea of re-purposing your data is now more appealing.20:31
adeuringabentley: even better :)20:32
benjirick_h_: at the top20:33
rick_h_ah, got it thanks20:33
adeuringderyck: could you review this mp: https://code.launchpad.net/~adeuring/launchpad/bug-sorting/+merge/82217 ?21:29
deryckadeuring, indeed, I can.21:29
adeuringderyck: greatm thanks!21:30
deryckadeuring, there's no diff on the mp.21:37
adeuringderyck: gahh... did a "bzr push" before a "bzr commit". The current version of the branhc is not pushed to LP21:38
deryckadeuring, ah, now I see a diff.21:43
deryckadeuring, r=me21:46
adeuringderyck: great, thanks!21:46
deryckadeuring, I'm not crazy about that isinstance stuff, but I see that's the pattern in that bit.21:47
adeuringderyck: do you mean the "isinstance(col, tuple)"? I know that this is duplicated, but to together iwth the "if orderby_col.startswith("-")" we need to handle for cases. Hard to not have some sort of duplication21:49
adeuringmorning huwshimi21:50
deryckmorning huwshimi21:52
deryckadeuring, it's not the duplication.  just doing type checks in general also bother me.21:52
adeuringderyck: I understand ;) I could change the other entries in the mapping, making the isinstance() calls obsolete, but I'll postpone this to the next branch related to bug search sorting21:54
deryckadeuring, ok, cool.  It wasn't a big deal to me since the surrounding code was like this; just noting it.21:55
deryckadeuring, But I welcome later updates to clean this up if time allows.21:55
=== benji changed the topic of #launchpad-dev to: https://dev.launchpad.net/ | On call reviewer: - | Critical bugtasks: 277
sinzuiStevenK, lp/soyuz/model/archive.py addArchiveDependency() looks promising and lp/bugs/model/bug.py addNomination() is helpful too. But about this last example, maybe an losa cannot fix a nomination22:15
wgrant>>> urlparse.urljoin('https://api.qastaging.launchpad.net/', 'devel', 'foo/bar')22:18
wgrantjcsackett: ^^ there's your problem22:18
flacostelifeless: do you have any reservation on turning on incremental diff on for everyon?22:20
flacoste(see email thread)22:20
wgrantflacoste: Incremental diffs, or the inline MP diffs?22:22
wgrantThey are different things for very similar purposes.22:22
flacosteinlin MP diffs22:22
flacostewhich is displaying an incremental diff22:22
lifelessflacoste: so, I know that loggerhead times out22:23
wgrantRight, but there's the separate incremental MP diff feature.22:23
lifelessits an inconsistent UI22:23
lifelessflacoste: its probably ok to turn on22:23
wgrantIt also doesn't have a spinner or error handling at all.22:23
lifelessflacoste: fortunately I should be able to do timeout oopses for loggerhead in a few hours work \o/22:23
wgrantAnd it's blue instead of green.;22:23
deryckrick_h_ is fixing the green link bug as his first bug fix.22:27
deryckhi wgrant and lifeless, btw :)22:27
thumperhi deryck22:27
deryckhi thumper22:27
lifelessderyck: hi :)22:27
wgrantEvening deryck.22:27
wgrantlifeless: Could you have a final look over https://code.launchpad.net/~wgrant/launchpad/observer-db/+merge/81104 before I land it?22:51

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