/srv/irclogs.ubuntu.com/2010/09/17/#launchpad-dev.txt

* thumper attempts to suspend again00:32
lifelesssinzui: I'm going to work on the moderate thing00:33
thumperyay, it worked00:34
lifelessthumper: ping00:37
thumperlifeless: hey00:37
lifelesshttps://bugs.edge.launchpad.net/launchpad-code/+bug/633758 - can has qa?00:37
* thumper gets on it00:38
lifeless(oh, and goo dmorning)00:38
thumpermorning00:39
thumpermy talk went well I think00:39
lifelessthumper: cool00:41
lifelesswhat was it ?00:41
thumperqa done00:41
thumperI was giving a talk to second year software engineering students about our development process00:41
lifelessone of the errors last night on BB : bzrlib.errors.InvalidHttpResponse: Invalid http response for https://xmlrpc.edge.launchpad.net/bazaar/: Unable to handle http code 502: Bad Gateway00:42
thumpera one hour lecture00:42
thumper?!?00:42
lifelessthumper: 'we throw stuff at the wall and hope it works'00:42
lifelessthumper: the 502 is probably an edge deploy.00:42
lifelessthumper: I wonder what bzr version it is, that its using edge00:42
* lifeless waits for https://devpad.canonical.com/~lpqateam/qa_reports/launchpad-stable-deployment.txt to update00:44
lifelesssinzui: is there a list with lots of messags that I can look at the moderation page with ?00:57
=== Ursinha is now known as Ursinha-afk
lifelessEdwin-afk: / jcsackett: https://bugs.edge.launchpad.net/launchpad-registry/+bug/597738 needs QA for the latest landing (rev 11547)01:11
lifelessthumper: https://code.edge.launchpad.net/~lifeless/launchpad/cp/+merge/3576402:08
lifelessthumper: No new code, and I've marked up all the user facing / risk mitigating/correcting changes.02:08
lifelessthumper: I'd like to get this through ec2 today to ask for a CP on monday.02:08
lifelessthumper: ping?04:46
thumperlifeless: hey04:47
lifelesscould you tag your review release-critical please?04:47
lifeless(thanks for doing it)04:47
wgrantWhen's the next release?04:49
thumperlifeless: so this is part of the "try to release more often?"04:49
thumperjust trying to work out your rationale04:50
thumperlifeless: my thinking is based around marking it "release-critical" where it clearly isn't just to cross some 'T's and dot some 'I's seems stupid04:54
thumperif we are changing the process, we should just do it04:55
thumperand not require this fake stamp just for the hell of it04:55
lifelessthumper: yes04:59
lifelessthumper: I suggest; we do this under the current process, and simultaneously point out that the current process is more onerous than the proposed, approved RFWTAD process04:59
lifelessthumper: and so qa-ok'd stuff gets a free pass for CP's.05:00
lifelessthumper: we'll need to tweak some toolchain stuff though - at least for the moment we *need* release-critical stamps to get past PQM.05:00
lifelessand PQM isn't really setup to check 'has been qa-oked'.05:00
lifelesswgrant: three weeks05:00
* thumper sighs05:00
thumperlifeless: so you need the rc to get pqm to land it?05:01
lifelesswgrant: yes05:01
lifelessbah05:01
lifelessthumper: yes05:01
wgrantHm.05:01
thumperok, I'll stamp it05:01
wgrantThere are revs in devel that shouldn't go out onto prod, AFAICT...05:02
lifelesswgrant: yes, my branch isn't all of devel05:02
lifelesswgrant: its all the ones passed by the qatagger05:02
wgranteg. r11552 enabled some unfinished UI on prod.05:02
wgrantThe bug doesn't have a tag, so I don't know its QA status.05:02
lifelesswgrant: I'm doing up to and including 1154605:02
lifelessthats all that has been fully QA'd.05:03
wgrantAha.05:03
wgrantEverything up to there is fairly sane, IIRC. Sounds good.05:03
wallyworldthumper: ping?05:06
thumperwallyworld: aye05:06
wallyworldjust want to check something....05:06
wallyworldanyone can submit a merge proposal, right?05:07
thumperwallyworld: as long as they can see the source and target branches, yes05:07
wallyworldok. there's some ambiguity with some of the comments in a test05:08
wallyworldjust wanted to confirm, thanks05:08
wallyworldi might skype you later if i need to discuss05:08
jtvStevenK: did you ever figure out that broken test?05:19
StevenKjtv: Which one?05:23
jtvStevenK: the one you suspected fakelibrarian of05:24
StevenKjtv: I didn't investigate it, sorry05:24
jtvStevenK: Oh, I thought it was what put db-devel into testfix so I thought it'd be a big thing.05:25
jtvlifeless: are you familiar with LaunchpadScriptLayer?05:32
lifelessNo, but i can be05:33
lifelesswhy?05:33
jtvIt combines ZopelessLayer and LaunchpadLayer, and adds one thing:05:33
jtv        provideUtility(TestMailBox(), IMailBox)05:33
jtvThere's a 4-year-old XXX from Francis saying it should be registered via ZCML in LaunchpadFunctionalLayer.05:34
jtvSorry; the comment says it _is_ registered via ZCML in the LaunchpadFunctionalLayer.05:35
jtvAnyway, this smells of installFixture to me.05:35
jtvI was thinking tests that use this could use ZopelessDatabaseLayer instead, and install fixtures for the test mailbox and if needed, the fake librarian.05:36
jtv(Or if they need the real librarian, they could use LaunchpadLayer—though that provides memcached which I'm not sure scripts are currently supposed to access)05:40
lifelessso05:41
lifelessI'd like to move all our 'layers' to use fixtures, decoupling stuff05:41
lifelessspecifically05:41
lifelesslp:python-fixtures05:41
lifelessI need to tweak the API a little more though05:41
=== poolie_ is now known as poolie
lifelessjtv: 16:41 < lifeless> so05:47
lifeless16:41 < lifeless> I'd like to move all our 'layers' to use fixtures, decoupling stuff05:47
lifeless16:41 < lifeless> specifically05:47
lifeless16:41 < lifeless> lp:python-fixtures05:47
jtvlifeless: thanks for repeating—my IRC connections are extra-weird today.  Yes, fixtures instead of layers would be good.  I've gotten used to reserving setUp for that sort of thing—as opposed to the creation of repetitive test data.05:49
pooliehello jtv!05:49
jtvhi poolie!05:49
jtvHow's things?05:49
pooliegood; i thought my drive had died but it was just a loose cable05:50
pooliethough it's a bit annoying to lose time to carefully working this out05:50
jtvLife's little rollercoasters05:50
jtvI'm _hoping_ I've got a similar problem with my backup drive. :)05:50
pooliehow do i provoke an oops again?05:51
lifelesspoolie: do something wrong05:51
lifelesspoolie: :P05:51
lifelesspoolie: ++oops++05:51
poolieah, and it appears in the comment05:52
pooliethanks!05:52
poolieand i see the memcache feature was used! yay05:52
lifelesspoolie: it was ?05:52
lifelessoh yes, it will get evaluated05:52
poolieaccording to the comment in view-source:https://edge.launchpad.net/bzr/ it was queried05:52
pooliei don't know if it had any effect05:52
lifelessyes, any page that uses memcache will query it06:00
=== almaisan-away is now known as al-maisan
lifelessback later06:15
=== al-maisan is now known as almaisan-away
=== almaisan-away is now known as al-maisan
=== al-maisan is now known as almaisan-away
=== almaisan-away is now known as al-maisan
henningejtv: I have converted the import code but still some tests are failing.08:25
jtvhenninge: what problems are you hitting?08:25
henningejtv: let me have a closer look, 'll get back to you in a few minutes.08:26
henningejtv: http://paste.ubuntu.com/495141/08:31
henningejtv: I have not yet worked with Karma.08:31
henningejtv: Is that something updateTranslation does, too, which I need to do manually now?08:31
jtvhenninge: shall I run you through the conceptual model of karma?08:32
jtvWhat's happening here is that the karma we assign has changed a bit08:32
henningejtv: that would be nice08:32
jtvIn a nutshell, karma can be assigned in various contexts (somewhat similar to our queue entries, really).08:33
henninge(only 5 failing tests in lp.translations, btw)08:33
jtvAdditionally, each assignment of karma is in a particular category.08:33
jtvIn the database, you might thus have a karma record that says something like "henning earned 5 karma points today in category marking-a-bug-as-fix-released, for the gim."08:34
jtvupdateTranslation assigns karma.08:34
jtvIt does so in Mysterious Ways.08:34
henninge;-)08:35
jtvI did not try to reconstruct and emulate exactly how, when & what.08:35
henningebut it is still done by ... sumitSuggestion/approve/... ?08:35
jtvInstead, we now have a model that's relatively easy to track (see POTMsgSet, though I added an implementation method to POTemplate)08:35
jtvYes.08:35
jtvWell, not "still" since those messages are new.  :)08:36
jtv*methods08:36
henningeyes, but they are the replacements for updateTranslation08:36
jtvOff the top of my head, we award 3 kinds of karma:08:36
jtv * Submitting a suggestion.08:36
jtv * Reviewing.08:36
jtv * Having your suggestion approved by someone else.08:36
henningejtv: ok, I see two of those in the traceback08:37
jtvThose are now neatly separated, with rules like "don't assign review/approved karma for someone approving their own translations."08:37
henningewhere are they defined?08:37
jtvLook for awardKarma in potmsgset.py08:37
jtvOne in submitSuggestion, and two in approveSuggestion.08:38
jtvCome to think of it, I think I forgot to do that in approveAsDiverged!08:38
jtvAnd clearCurrentTranslation.  That's review karma for the reviewer, I think (assuming that there really are suggestions).08:39
henningeok08:41
henningemaybe I need to add that. I use approveAsDiverged in the import if the incumbent ;) message was diverged.08:42
jtvThen that should probably award karma in the same way approveSuggestion does: only if the message was not previously current and the reviewer is not the translator.08:44
jtvhenninge: maybe updateTranslation would be a fit for what you're doing.  That also follows the "if it's diverged, and the new translation is not identical to the shared one, keep diverged" rule.08:49
henningejtv: oh, I have to add the "is not identical" ...08:50
henningejtv: and that is not something "approveAsDiverged" does?08:52
henningediverged = suggestion.is_diverged08:52
henningejtv: ^ Can a suggestion actually be diverged?08:52
mrevellHi09:04
adeuringgood morning09:10
jtvhenninge: a suggestion can be diverged _in another template_.09:19
jtvhenninge: as for approveAsDiverged: that method is for explicitly diverging a message.  It always diverges (except where you're actually trying to set the shared translation as a diverged translation).09:20
jtvbut setCurrentTranslation checks if the current message is diverged and if so, maintains divergence.  (Again of course, except if the new translation is identical to the current shared one)09:20
henningejtv: I don't like setCurrentTranslations because it expects strings ... ;(09:27
henningewhich is kind of dumb, considering that I already have a message, and msgstr objects and all that.09:28
henningethat's why I'd prefer using approve09:28
henningeapproveSuggestion09:28
jtvhenninge: shouldn't be that hard to change—it's only used by tests, and other methods that can look up the POTranslation objects.09:28
jtvI'm not saying you have to use it, but have a look if it serves your needs otherwise.09:29
henningejtv: well, "approveSuggestion" is really only a wrapper around setCurrentTranslation09:30
henningeplus karma assignment09:30
henningejtv: So, I guess I don't need to check for divergence - submit suggestion will DTRT?09:31
jtvsubmitSuggestion can't09:32
jtvBut I see what you mean—approveSuggestion maintains divergence.09:33
jtvI do see one other thing that approveSuggestion probably should do: disable a diverged message in the same pofile that may be masking the suggestion.09:34
jtvIt shortcuts the case where the suggestion is already current.  It probably shouldn't, though of course not shortcutting does mean that the karma assignment needs a bit of extra care.09:35
jtvBut yes, this means that in principle approveSuggestion is good enough for imports.  No need for approveAsDiverged.09:36
jmlrockstar, https://dev.launchpad.net/LEP/BetterPrivacy, https://dev.launchpad.net/LEP/BugzillaComponents and https://dev.launchpad.net/LEP/BuildFarmScalability all spring to mind10:07
henningejtv: I think approveSuggestion may have gotten the Karma handling wrong.11:26
henningejtv: this is from the failing test: http://paste.ubuntu.com/495203/11:26
jtvhenninge: I can see how that's different from what approveSuggestion does, but how is what approveSuggestion does wrong?11:29
henningejtv: well, different from documentation == wrong11:29
henninge;-)11:29
jtvThis pastebin leaves me with the impression that updateTranslation gets it wrong11:29
henningethat is possible11:30
henningejtv: should a translator not get karma twice -- once for suggesting and once for suggestion being approved?11:31
jtva translator, yes11:31
jtvbut if it's the same person approving that same translation?11:31
jtvThat seemed wrong to me.11:31
henningejtv: so, what is karma trying to show?11:32
jtvIt is trying to reward people for doing useful things.11:32
jtvI'm trying to find something wrong with my own earlier reasoning, but not finding anything so far.11:33
jtvYes, it does mean that a reviewer can hope to gain more karma from translating in translation mode—_if_ they can get those suggestions reviewed by someone else.11:34
jtvI don't think that's a bad thing.11:34
henningeA translator that translates 10 strings that are all approved by a reviewer will get twice as much karma as a reviewer that enters the same translations directly.11:34
jtvI think that's actually rewarding good behaviour.11:35
jtvAt first glance it seems to increase the risk of those translations staying unapproved.  But the author doesn't get the extra karma until they are.11:35
jtvSo a proper, well-motivated karma whore will try to get them reviewed.11:36
henningeIn total, a suggested-approved translation will deal out three times as much karma as a directly-approved will.11:36
jtvTo different people.11:37
henningejtv: I am just finding reasons, trying to understand. I am not against your reasoning.11:37
jtvThat's fine.11:37
henningeSo I should update the test.11:37
jtvI think so, yes.11:37
jtvBTW dogfood is being upgraded to lucid.  It'll take about an hour.11:38
henningeSo far, imported translation would give the uploader "translationsuggestionapproved" karma, now they will get "translationsuggestionadded" karma.11:39
henningeAre they weighed differently?11:39
jtvDon't know.11:39
* henninge consults launchpad_dev11:40
henningejtv: http://paste.ubuntu.com/495216/11:42
jtvhenninge: thanks for digging that up!11:43
henningelooks like we have some more karma actions ...11:43
jtvBTW another way of looking at the current karma design is, "karma shouldn't reward people for joining a translation team just to translate."11:43
=== al-maisan is now known as almaisan-away
deryckMorning, all.12:04
=== mrevell is now known as mrevell-lunch
jmlderyck, good morning12:14
jmlI'm going to head out to Foyles and see if I can pick up a book on queueing theory12:26
=== mrevell-lunch is now known as mrevell
=== matsubara-afk is now known as matsubara
=== Ursinha-afk is now known as Ursinha
jmlhello14:41
bigjoolsdoes the .dev instance run with multiple request threads?14:45
salgadobigjools, yes14:49
bigjoolssalgado: cool thanks14:50
bigjoolsnow I have a fighting chance of re-creating my concurrency issue14:51
=== almaisan-away is now known as al-maisan
bigjoolssalgado: do you have any clue how I could write a test case that simulates simultaneous requests in 2 threads?15:05
salgadobigjools, I think the test suite runs single threaded15:06
bigjoolsyeah, that's why I asked :)15:06
jmlbigjools, I guess it all depends on how the test browser works15:09
bigjoolsjml: I've only ever seen it manage single-threaded stuff15:10
bigjoolsjml: I've re-created an issue locally where I need to bash the copy button on the PPA copy page as fast as I can15:11
jmlbigjools, well, I don't know how it works. It might just be that all you need to do is spawn two threads each of which has its own test browser15:11
jmlbigjools, hmm.15:12
jmlbigjools, in that case, you might not even need to worry about the test browser15:12
bigjoolsjml: that's one way but it seems as though it might not be deterministic15:12
jmlbigjools, you are using threads, kiss determinism goodbye15:12
bigjoolsthat's why I said "simulate" in my original question :)15:12
jmlbigjools, then what's the problem with spawning two threads and using the test browser?15:13
jmlbigjools, or, actually15:13
jmlbigjools, you don't need the test browser, you just need to have instantiated view objects pointing to the same store15:13
bigjoolsI'm starting to think that a test case is not worth it15:13
jmlbigjools, hang on15:13
jmlbigjools, is this for debugging the problem or writing a test?15:13
bigjoolsa test15:13
bigjoolsI know what the problem is, I think15:13
bigjoolsthe two transactions don't clash at all, so don't trigger a serialization issue15:14
bigjoolsbut the checks that are made work once one of them is committed15:14
jmlbigjools, otp15:14
bigjoolsso I have an evil plan to make them clash some how15:14
bigjoolsok15:15
rockstarjml, did you see my LEP?15:32
jmlrockstar, I saw that it exists. still otp.15:33
rockstarjml, ack.15:35
jmlbigjools, do you know off the top of your head what times the Walter de Cantelupe is closed on Sundays?15:52
bigjoolsjml: late15:54
bigjoolsusually ~11 I think15:54
bigjoolsbut if you're going to be late I'd call and let him know15:54
jmlbigjools, istr Martin saying that it's closed for some time during the afternoon15:55
jmlbigjools, I'll be arriving at ~6pm15:55
bigjoolsjml: I'd call him then, yes it is closed in the afternoons, I can't remember what time it opens15:56
jmlbigjools, will do.15:56
jmlbigjools, thanks15:56
bigjoolsnp15:57
jmlgrr.16:20
=== matsubara is now known as matsubara-lunch
jmlDo you know what would make most technology better?16:42
jmlActually working.16:42
=== benji is now known as benji-lunch
=== beuno is now known as beuno-lunch
=== deryck is now known as deryck[lunch]
jmlrockstar, I just scribbled all over https://dev.launchpad.net/Code/MergeQueues/LEP17:14
jmlrockstar, basically adding more details in the requirements.17:14
rockstarjml, great, I'll take a look.17:14
rockstarjml, ah, yeah, most of those details we've already addressed in my notes at https://dev.launchpad.net/Code/MergeQueues but I'm happy to LEPify them.17:20
jmlrockstar, yeah. I see the notes cover similar ground in story format.17:21
jmlrockstar,17:21
jmlrockstar, but I want to get it down in a requirements format17:22
rockstarjml, absolutely.  abentley and I were just going over the recipe LEP and reviewing what else needs to be done, so I'm identifying what is important about a LEP.17:22
jmlrockstar, cool.17:23
jmlrockstar, also, I'm really keen to see thoughts on the full cycle of any given feature: how do people find out about it; how does one use it for the very first time; how does one know it's safe to try out; what do the inner cycles look like; how does one  stop using it? etc17:24
rockstarjml, yeah, and I think we're going to identify much of that with our UI mockups.17:24
jmlrockstar, agreed17:24
rockstar(I have a list of them I need to do)17:24
jmlrockstar, it's easy to let UI mockups focus on the inner bits of the workflow rather than a more broad context, so just flagging now17:25
rockstarjml, yeah, I actually like to address the more broad context in UI mockups.  They should be a general idea, not the final UI.17:26
jmlrockstar, also, I noticed on your notes page that you're thinking of having something on the branch page that says the branch isn't managed by a queue17:26
rockstarThe final UI gets sorted as you actually implement the feature and find out what works.17:26
jmlrockstar, for a project that uses feature branches, that's going to be irrelevant for most branches.17:26
=== salgado is now known as salgado-lunch
=== matsubara-lunch is now known as matsubara
* gmb EoWs. Night folks.18:01
=== benji-lunch is now known as benji
=== beuno-lunch is now known as beuno
=== salgado-lunch is now known as salgado
=== al-maisan is now known as almaisan-away
=== Ursinha is now known as Ursinha-lunch
lifelessbac: hi, around?19:35
bachi lifeless19:36
lifelessrev 11552 on stable19:36
lifelessI'm not clear if that is ok to deploy to lpnet now, or not.19:36
baclifeless: b/c it is qa-untestable or b/c of pre-requisite branches19:38
lifelessbecause of the description19:38
lifeless'show xx on lpnet'19:38
lifelessI don't know if you intend that to wait for the next release19:38
lifelessor if showing it on Monday would be ok.19:38
baclifeless: Monday is great19:39
lifelessok, cool19:39
lifelesswe're not qa'd all the way up to it, but as things move forward...19:39
baclifeless: we should have rolled it in 10.09 but it slipped my mind19:39
lifelessUrsinha-lunch: hiya19:39
lifelessbac: feature flags will help :)19:39
baclifeless: yes, they just weren't ready at the time...19:39
baclifeless: why do you do the endwith test at line 27 of https://code.edge.launchpad.net/~lifeless/launchpad/registry/+merge/3577419:42
lifelesscompleteness19:43
=== EdwinGrubbs is now known as Edwin-lunch
baclifeless: but when would it not pass?19:43
lifelessoh ahh19:43
lifelessthats a bug, I meant to put %3f or whatever in there.19:44
lifelessI'll come back to it, its already in ec2 land - thanks for noticing.19:44
bacok, i wasn't sure if it was a bug or you were doing something clever19:44
lifelessvery very clever19:45
lifelessso clever you put a tail on it and call it a ferret19:45
=== Ursinha-lunch is now known as Ursinha
Ursinhalifeless, hi20:27
lifelessUrsinha: hi there20:38
lifelessuhm, /me pages in20:38
lifelessUrsinha: oh right, tagger stuff - how goes it? Anything blocking you?20:38
Ursinhalifeless, I'm not blocked on anything right now20:45
Ursinhalifeless, more trying to figure out how to solve the mess unique branches for several bugs introduced to the scenario :)20:46
lifelessheh, sorry ;)20:48
lifelessgrah20:57
lifelessec2 fail20:57
lifeless File "./test_on_merge.py", line 50, in setup_test_database20:57
lifeless   con = psycopg2.connect('dbname=template1')20:57
lifelesspsycopg2.OperationalError: could not connect to server: No such file or directory20:57
lifeless       Is the server running locally and accepting20:57
lifeless       connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"?20:57
lifelessok, is the AMI bust or something?21:55
lifelessI'm getting pg not connected errors consistently on ec2land21:56
lifelessmars: ^21:56
lifeless# Run all tests. test_on_merge.py takes care of setting up the21:56
lifeless# database.21:56
lifeless/var/launchpad/test/bin/py -t ./test_on_merge.py -vv "--subunit -vvv"21:56
lifelessTraceback (most recent call last):21:56
lifeless...21:56
lifeless File "./test_on_merge.py", line 50, in setup_test_database21:56
lifeless   con = psycopg2.connect('dbname=template1')21:56
lifelesspsycopg2.OperationalError: could not connect to server: No such file or directory21:56
lifeless       Is the server running locally and accepting21:56
lifeless       connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"?21:56
=== matsubara is now known as matsubara-afk
maxbWhat is the status of Python 2.6 / datacentre lucid upgrades these days?22:51
Ursinhalifeless, help me with something here22:51
Ursinhalifeless, latest deployed revision from stable on lpnet was 11515, now the script states it's 1154622:53
Ursinhawhile True: head -> desk22:54
lifelessUrsinha: heh22:58
Ursinhalifeless, no, not funny22:58
lifelessUrsinha: looking at stable branch ?22:58
Ursinhadeployments.check_status() is returning that revision as latest deployed22:58
lifelessthats right22:58
Ursinhaproduction-stable22:58
lifelessits looking at the branch22:58
lifelessits correct; it goes live Monday22:58
Ursinhawait, what goes live Monday22:59
lifelesscheck LaunchpadProductionStatus22:59
* Ursinha looks22:59
lifelessjelmer: you need to update that page now you've landed your branch on production-devel23:00
lifelessUrsinha: under 'requested cherrypicks'23:00
Ursinhalifeless, oh23:00
Ursinhaso production-stable already has that23:01
Ursinha?23:01
lifelessyes23:01
Ursinhaah23:01
Ursinha*whew*23:01
lifelessworking as designed23:01
UrsinhaI was about to kill someone23:01
lifelessits not live on the appserver because doing it friday isn't the smartest thing for a big merge like that23:01
lifelessnot when we were low-losa to boot23:01
lifelessjcsackett: ping23:02
lifelesshttps://bugs.edge.launchpad.net/launchpad-registry/+bug/59773823:02
lifelessthats the next blocking unqa'd commit23:02
jcsackettthe one for bugs?23:02
lifelessjcsackett: rev 1154723:02
lifelessjcsackett: but the process looks at a whole bug, not one branch-for-it.23:02
lifelessjcsackett: its qa-needstesting at the moment, if the bugs bits are doing what they should on staging/edge, please qa-ok it23:03
Ursinhado that and I will run the script23:03
jcsacketti can ping EdwinGrubbs about the branch he landed to QA; the remaining hasn't reached fix-committed yet. Blueprints should be through buildbot shortly.23:03
lifelessjcsackett: 11547 is what I care about right now.23:04
Ursinhajcsackett, so it's part of a larger fix?23:04
lifelessin general its a bit odd to do what this bug is doing23:04
sinzuiI QAed bugs yesterday and reported bugs.23:04
lifelesssinzui: so should it be qa-bad ?23:04
sinzuino23:04
sinzuiit is still an improvement23:04
lifelessso its qa-ok?23:04
EdwinGrubbsjcsackett: my branch was qa'ed fine. I didn't update the bug since it is is just going back to qa-needstesting when the next bugtask is completed.23:04
sinzuiyes23:04
Ursinhacool23:05
lifelessEdwinGrubbs: it needs to go to qa-ok to deploy23:05
jcsackettEdwinGrubbs: okay, i thought you already had done that.23:05
lifelessotherwise it blocks all commits after it23:05
UrsinhaEdwinGrubbs, you can tag the bug when your branch lands23:05
* sinzui found to pre-existing issues that must be fixed to say bridging-the-gap is done23:05
* jcsackett is confused about the QA process.23:05
EdwinGrubbsoh, I forgot about that.23:05
jcsackettare we not supposed to use needstesting anymore?23:05
lifelessjcsackett: yes, you should. You should also test asap.23:05
Ursinhajcsackett, yes, of course :) but you're saying qa-needstesting for a branch23:06
sinzuiWake up, QA23:06
lifelessjcsackett: https://devpad.canonical.com/~lpqateam/qa_reports/launchpad-stable-deployment.txt23:06
Ursinhayou landed a branch that fixes your bugtask in that bug, that lands, it's marked qa-needstesting for your landing23:06
sinzuinotice staging updates QA...QA is the fastest way to unblock the team23:06
lifelessor on edge23:06
jcsackettokay, that all jives with what i thought. have caught bits and pieces of conversations today that have left me off kilter. :-P23:06
EdwinGrubbsjcsackett: I just treated this bug differently since it has multiple bugtasks. I didn't think about how it needs to go to qa-ok temporarily even though it will go to qa-needstesting when the next branch lands.23:06
lifelessbecause edge runs stable23:06
sinzuiWe were 3 hours from exceeding our limits 3 days ago23:06
lifelessEdwinGrubbs: I would have done multiple bugs for this thing23:07
lifelessbecause the work going on in each one is different23:07
EdwinGrubbsyes, hindsight is 20/2023:07
lifelessnever too late to change ;)23:08
lifelessanyhow yes, this isn't a big issue23:09
lifelesswe're changing process23:09
lifelessso we're going to have muscle memory to unlearn and so on23:09
jcsacketti'll keep this in mind tomorrow--the blueprints task should be ready to qa then.23:09
sinzuiI think the registry team can take ownership of EmailAddress if the foundations team have given up using the table for SSO23:10
lifelesswgrant: please qa https://bugs.edge.launchpad.net/soyuz/+bug/56449123:12
lifelesswgrant: and https://bugs.edge.launchpad.net/soyuz/+bug/56633923:12
Ursinhalifeless, so I cannot mark that bug as qa-ok?23:13
lifelessUrsinha: I don't know23:14
lifelessUrsinha: hang on, which bug :)23:14
Ursinhahehe23:14
Ursinhahttps://bugs.edge.launchpad.net/launchpad-registry/+bug/59773823:14
Ursinhathe one holding the script23:14
lifelessjcsackett: so23:14
lifelessnvm23:15
lifelessUrsinha: yes, sinzui said its ok23:15
Ursinhahehe, cool23:15
* sinzui just changed is since his team are too frightened to say it is okay23:15
Ursinhaoh, I've done that too23:16
lifelessUrsinha: I expect it to stop at 1155623:16
Ursinhalet's see23:16
sinzuiUrsinha, while I promised flacoste that I would let my team do QA, I need to schedule work. I qa every morning, even weekends, and I mark something bad if it really is bad23:16
Ursinhaeven weekends23:16
sinzuiUrsinha, The delay is waiting for a registry developer to do his part23:17
sinzuiI have to review projects on the weekend too23:17
Ursinhalifeless, that's correct, it stopped at 1155623:17
lifelessUrsinha: still, thats another good batch of revs to land23:19
Ursinhayeag23:19
Ursinhayeah23:19
lifelesssinzui: all teams are going to have to tune their processes, I wouldn't stress about it23:20
lifelesssinzui: it'll become very clear the shorter the pipeline to production gets23:20
lifelesssinzui: and once we stop edge autodeployments in favour of QA on stagingqa23:21
sinzuilifeless, I think there is resistance to interrupting a task. When you are coding the a branch waiting for to QA your work, you must be will to stop coding when that work come to QA. Our team hit the QA limit twice last release Bac had to remind me that staging updated a few days ago so that he cold move a card23:24
lifelesssinzui: that makes sense to me; but the price for avoiding that interrupt is a massive team wide interrupt every month, which everyone agrees they like less23:28
sinzuiAs you can image, I have a very set pattern in the morning, lunch, and evening. I look like I am on top of things, but I just schedule repetitive tasks like QA and question every few hours23:30
lifelesssinzui: I think if everyone looked for QAable stuff at the start of work, after lunch, and before signing off, we'd be in good shape23:31
lifelesshttps://bugs.edge.launchpad.net/launchpad-project/+bugs?field.tag=qa-needstesting,qa-bad23:32
lifelessthats one thing for every 2 people: not terrible, but not brilliant either23:33
lifelessjcsackett: https://bugs.edge.launchpad.net/launchpad-registry/+bug/57675723:33
jcsackettlifeless: qaing it as we speak, actually. :-P23:34
sinzuilifeless, we are doing that one now23:34
lifeless\o/23:34
sinzuijinx23:34
lifelessis there some other channel where you coordinate this?23:34
sinzuilifeless, @lp-registry explaining how I pick a victim to test with can be a delicate subject23:35
jcsackettlifeless: when i have no earthly idea how to set up the conditions to test i bug sinzui in another channel. :-)23:35
sinzui#lp-registry23:35
lifelesspublic or private?23:35
sinzuilifeless, private23:36
UrsinhaI'm heading for the weekend23:37
Ursinhawill return later23:37
=== Ursinha is now known as Ursinha-afk
lifelesssinzui: I'll hang out there if its ok; I may every now and then point stuff to this channel (I think we do way to much on the quiet). But you can tell me if I'm wrong :)23:37
lifelessUrsinha-afk: ciao23:37
lifelessQuestion:+huge-vocabulary 50.67 <-99% under23:43
lifelesswheee23:43
jcsackettalright, that's both of the bugs i can hit qa'ed. off to the weekend.23:43
lifelessciao23:43
jcsackettg'night.23:43
=== jcsackett is now known as jcsackett|afk

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