=== JasonO- is now known as JasonO | ||
=== Ursinha is now known as Ursinha-afk | ||
pitti | Good morning | 06:39 |
---|---|---|
pitti | test hacking day! | 06:39 |
pitti | https://lists.ubuntu.com/archives/ubuntu-devel-announce/2012-December/000999.html | 07:01 |
elfy | morning pitti | 07:05 |
pitti | hey elfy | 07:07 |
jibel | good morning! | 07:40 |
pitti | bonjour jibel | 07:40 |
jibel | guten Morgen pitti | 07:42 |
jibel | ready to hack? | 07:42 |
pitti | oui, je suis! | 07:42 |
jibel | great! | 07:43 |
jibel | So, I did a little count this morning | 07:43 |
jibel | I checked https://jenkins.qa.ubuntu.com/view/Raring/view/AutoPkgTest/ | 07:43 |
jibel | and there are 63 packages with tests and 11 of them are failing | 07:44 |
jibel | let see how much we can add or fix in a day | 07:44 |
pitti | I want to work on some failing ones today | 07:44 |
pitti | and I'm happy to review/test/sponsor contributor tests | 07:44 |
jibel | pitti, which failing package will you start with? | 07:47 |
jibel | I can give a try at update-manager | 07:47 |
jibel | https://jenkins.qa.ubuntu.com/view/Raring/view/AutoPkgTest/job/raring-adt-update-manager | 07:48 |
jibel | from the log there is only 1 failing test, looks like a low hanging fruit | 07:48 |
jibel | https://jenkins.qa.ubuntu.com/view/Raring/view/AutoPkgTest/job/raring-adt-update-manager/44/ARCH=i386,label=adt/artifact/results/dsc0t-nose-tests-stdout | 07:48 |
jibel | perfect for me :) | 07:48 |
jibel | I'll file a bug first for this failure | 07:50 |
pitti | jibel: it's not quite that simple, I'm afraid; it fails with two different missing packages on i386/amd64 | 07:50 |
pitti | but it might still be simple after all, I haven't looked into it yet | 07:50 |
jibel | pitti, right, never underestimate mvo's code ;) | 07:51 |
jibel | I filed bug 1089793 | 07:57 |
ubot5 | Launchpad bug 1089793 in update-manager (Ubuntu) "testOriginMatcherWithVersionInUpdatesAndSecurity failed withAssertionError: None == None : no package 'ecryptfs-utils' installed " [Undecided,New] https://launchpad.net/bugs/1089793 | 07:57 |
jibel | so now, I'll prepare a fresh testing environment with prepare-testbed from lp:auto-package-testing | 07:59 |
jibel | $ ./bin/prepare-testbed amd64 | 07:59 |
dholbach | good morning | 08:03 |
pitti | hallo dholbach | 08:03 |
dholbach | happy automated testing day! | 08:03 |
pitti | happy hacking day! /me gets his axe | 08:03 |
dholbach | :-) | 08:03 |
jibel | Hey dholbach | 08:04 |
dholbach | pitti, jibel: at 9 UTC I'm going to interview didrocks for the ubuntudev hangout when do you think we should do a automated testing hangout? | 08:04 |
pitti | I can do a hangout for about 1 minute | 08:05 |
pitti | then google mutes me | 08:05 |
jibel | dholbach, anytime is good | 08:05 |
dholbach | pitti, wow | 08:05 |
jibel | pitti, did you try the nexus with android | 08:05 |
jibel | ? | 08:05 |
dholbach | is that on raring? :-( | 08:05 |
dholbach | that sucks | 08:05 |
pitti | yes, on raring; not sure whether that matters | 08:05 |
pitti | jibel: I guess I could kill raring on the nexus and install android, yes | 08:06 |
jibel | we did hangout with didrocks for hours | 08:06 |
dholbach | it was similar when I interviewed Laney - but it wasn't after 1 minute, but 3 times in a 1h hangout | 08:06 |
jibel | I also must set g+ hangout preferences to low bandwidth despite I have a 100Mbps connection | 08:07 |
pitti | did anyone try killing pulse before, or something like that? | 08:12 |
pitti | I'm going to look at the eternal udisks2 hang | 08:18 |
dholbach | maybe somebody else could later on demo a few things? | 08:19 |
dholbach | it might be good to have a bit of a schedule and if we could do something in the European start of the day and maybe something in the US start of the day | 08:20 |
pitti | jibel: I updated https://wiki.ubuntu.com/QATeam/RequiredTests to grab udisks2, and noted you down for update-manager | 08:21 |
jibel | pitti, ack | 08:23 |
jibel | so it's interesting, adt-run fails when running update-manager tests from the source tree when it runs dh_auto_build | 08:23 |
jibel | it uses the wrong version of python, despite the override in the rules file | 08:24 |
pitti | err, not amused; running the udisks Luks tests causes a kernel oops | 08:34 |
jibel | ah, that explains the test hangs forever | 08:37 |
pitti | I'll check if it also happens in kvm | 08:37 |
jtaylor | is there a time limit on how long autopkgtests can run? | 08:41 |
jtaylor | e.g. is 8 hours too much? | 08:41 |
pitti | I'd say yes; there is only so much capacity we have in teh DC | 08:41 |
jtaylor | there should probably be a restriction specifier for slow tests? | 08:41 |
jtaylor | so that one can run those less often | 08:42 |
jtaylor | in dep8 | 08:42 |
pitti | autopkgtest defaults to a timeout of 2.7 h (10.000 s) for tests | 08:42 |
pitti | that does not include the build though, if you have a test with build-needed | 08:43 |
pitti | jtaylor: well, we can't run them less often -- they need to run on each package and dependency upload, otherwise they wouldn't do what we want them to do (prevent regressions from landing in ubuntu) | 08:46 |
pitti | jtaylor: perhaps you can disable some very expensive tests for dep8, and then they can run manually every now and then? | 08:47 |
jtaylor | couldn't you run the fast tests on every change and the slow test once a week if there was a change? | 08:48 |
pitti | we don't have that kind of smarts ATM | 08:48 |
jtaylor | or at certain milestone points | 08:48 |
jtaylor | jenkins can do that easily | 08:48 |
pitti | wel, not in ADT; we can run them as a different kind of test | 08:49 |
pitti | right | 08:49 |
dkessel | good morning | 08:52 |
dkessel | i want to help with the simple compile/link/run tests | 08:52 |
dkessel | i'm starting at the top of the list, libatk-dev | 08:57 |
pitti | dkessel: great! | 08:57 |
jibel | pitti, dholbach and all, I created http://pad.ubuntu.com/testing-hackfest-20121213 to keep track of our progress | 09:00 |
pitti | ah, so we won't use the wiki page? | 09:02 |
jibel | it's more dynamic and easier to use for the day, then we'll update the wiki page for history | 09:05 |
pitti | ack | 09:06 |
jibel | mvo, about bug 1089808, I can just remove the dh_auto_build call from the override in debian/rules, right? | 09:08 |
ubot5 | Launchpad bug 1089808 in update-manager (Ubuntu) "debian/rules build failed with ImportError: No module named DistUtilsExtra.command" [Undecided,New] https://launchpad.net/bugs/1089808 | 09:08 |
mvo | jibel: let me check | 09:14 |
mvo | jibel: yeah, that should be ok | 09:15 |
dkessel | i am trying to follow the steps on http://developer.ubuntu.com/packaging/html/auto-pkg-test.html | 09:16 |
dkessel | if i run | 09:17 |
dkessel | bzr branch ubuntu:libatk1.0-dev | 09:17 |
mvo | jibel: are you on it or should I just do it and upload? | 09:17 |
dkessel | i get: ERROR: no branch: »bzr+ssh://bazaar.launchpad.net/+branch/ubuntu/libatk1.0-dev/ | 09:18 |
jibel | mvo, I was about to do it, but go for it | 09:21 |
jibel | dkessel, I think it's bzr branch ubuntu:atk1.0 | 09:23 |
jibel | dkessel, atk1.0 is the source package for libatk1.0-dev | 09:24 |
dkessel | oh, right | 09:24 |
jibel | according to apt-cache showsrc libatk1.0-dev | 09:24 |
dkessel | thx | 09:24 |
* dkessel takes notes on new commands :) | 09:25 | |
jibel | mvo, I'm on bug 1089793, you can save an upload if you wait a bit. | 09:27 |
ubot5 | Launchpad bug 1089793 in update-manager (Ubuntu) "testOriginMatcherWithVersionInUpdatesAndSecurity failed withAssertionError: None == None : no package 'ecryptfs-utils' installed " [Undecided,New] https://launchpad.net/bugs/1089793 | 09:27 |
mvo | jibel: sure, I will just commit | 09:27 |
mvo | jibel: and wait for your ack | 09:27 |
mvo | jibel: so just to double check, its calling py2 even when run with --with=python3? | 09:29 |
jibel | mvo, yes | 09:29 |
mvo | jibel: ok, commited | 09:29 |
jibel | mvo, we had a similar issue last week with unity and fixed it with the same override http://bazaar.launchpad.net/~online-accounts/unity-scope-gdrive/trunk/view/head:/debian/rules | 09:31 |
jibel | mvo, does apt always re-reads dpkg-status when Cache.open() is called? | 09:57 |
jibel | mvo, I'm on test_update_origin.py in update-manager | 09:57 |
jibel | this test testOriginMatcherWithVersionInUpdatesAndSecurity | 09:57 |
jibel | these lines (105-122) http://paste.ubuntu.com/1429365/ | 09:57 |
jibel | and pkg._pkg.current_ver is always None, despite the record is set to 'installed' in dpkg status | 09:58 |
mvo | jibel: it should be | 09:59 |
* mvo scratches head | 09:59 | |
dholbach | jibel, awesome | 10:05 |
dholbach | pitti, didrocks used his tablet for the hangout as well :) | 10:09 |
didrocks | well, the sound is cutting down on my laptop machine after a minute in a hangout for some unknown reason, so yeah, have to use a tablet :) | 10:10 |
pitti | yeah, same for me | 10:10 |
didrocks | weird that not everyone is impacted | 10:10 |
dholbach | didrocks, laney had a similar problem on tuesday, but it wasn't after a minute - "just" 3 times in an one hour hangout | 10:12 |
didrocks | 20 minutes in a average is 20 times better! :-) | 10:12 |
didrocks | quite reliably under a minute here :/ | 10:12 |
didrocks | s/in a/on a/ | 10:12 |
dholbach | I'm not impacted as I'm still on quantal on my laptop :) | 10:13 |
dholbach | I know, cheating, I know :-P | 10:13 |
didrocks | dholbach: and you tell that without being ashamed of it? :-) | 10:13 |
didrocks | I see, I see ;) | 10:13 |
dholbach | I use raring elsewhere :) | 10:13 |
didrocks | heh | 10:13 |
jibel | mvo, I did the following test http://paste.ubuntu.com/1429399/ and current_ver is always None, any idea? | 10:24 |
mvo | jibel: give me some minutes, I should have this ready in a bit | 10:29 |
mvo | jibel: eh, I mean, I should be able to look in a bit, just need to finish a mail | 10:29 |
smartboyhw | Hi guys: Automated testing hacking right? | 10:31 |
pitti | hello smartboyhw; yes | 10:32 |
smartboyhw | pitti, OK so what can I hack? | 10:32 |
pitti | smartboyhw: if you want to start with something small, I guess adding a compile/link/run autopkgtest to a library is a good start | 10:33 |
pitti | smartboyhw: see https://wiki.ubuntu.com/QATeam/RequiredTests for a list of library packages that need one | 10:33 |
smartboyhw | pitti, OK | 10:33 |
pitti | let me remove the ones which got tests in the meantime | 10:33 |
pitti | ok, done | 10:36 |
pitti | I dropped the libs which have tests from https://wiki.ubuntu.com/QATeam/RequiredTests and removed stale locks | 10:36 |
mvo | jibel: hm, could that be that there is simply the "Architecture:" missing? that is fixed in current trunk | 10:36 |
pitti | dkessel: so you are working on libatk, right? | 10:37 |
pitti | dkessel, smartboyhw: if you "grab" a package, please note so on http://pad.ubuntu.com/testing-hackfest-20121213, so that we avoid overlapping | 10:37 |
mvo | jibel: I think thats it, I can upload and it should be green again, the fix is in trunk but not uploaded | 10:38 |
jibel | mvo, that's it | 10:39 |
dkessel | pitti: yes, libatk | 10:39 |
* jibel hugs mvo | 10:39 | |
mvo | jibel: cool, I will upload then | 10:39 |
dkessel | pitti: i tried, but i can't access it - "Either you have not been granted access to this resource or your entitlement has timed out. Please try again." | 10:39 |
* mvo hugs jibel back :) | 10:39 | |
pitti | dkessel: ah, odd; anyway, I'll add a note for you | 10:40 |
pitti | dkessel: this should work though standard SSO | 10:40 |
pitti | jibel: edit wars! | 10:41 |
dkessel | pitti: not for me - tried logging out and in again... don't have any problems accessing e.g. launchpad... | 10:41 |
* mvo hugs jibel again, just because | 10:42 | |
dkessel | i get an exception in adt-run when running my test with it: http://pastebin.com/DewiGnWd | 10:44 |
pitti | it's trying to call /usr/lib/pbuilder/pbuilder-satisfydepends-classic | 10:45 |
pitti | dkessel: please sudo apt-get install pbuilder | 10:45 |
pitti | it's recommended by autopkgtest, apparently you disabled recommends in apt? | 10:46 |
didrocks | pitti: oh, you're using -classic as well? I have to use that for using some options that are display in --help by unimplemented in other flavors | 10:46 |
pitti | didrocks: it's what autopkgtest calls | 10:47 |
didrocks | ok, so we are aligned on that at least :) | 10:47 |
* pitti uploads an udisks2 which should avoid the kernel crash and succeed again | 10:49 | |
didrocks | pitti: hum, nice crash. I was telling: this is what I use in my cowbuilder to prepare the source package, but as the other build-deps or maybe not available, I'm doing it with --force-version and --continue-fail | 10:51 |
pitti | I looked ath the software-center failures a while ago, and it seems there's a gazillion different failures; both due to real bugs, and also some because of the test environment | 10:51 |
pitti | mvo: ^ does it make sense to fix those, or will s-c be obsoleted by the unity dash soon anyway? | 10:51 |
dkessel | pitti: apt-cache show autopkgtest says "Recommends: apt-utils" for me... not pbuilder. this is on precise... | 10:52 |
mvo | pitti: it will be obsoleted, but I guess its still worth exploring, let me create a VM for it - raring I assume? | 10:53 |
pitti | mvo: right | 10:53 |
* mvo run ./bin/prepare-testbed | 10:53 | |
pitti | mvo: but I figure you can reproduce most of the failures by simply calling debian/tests/whatever locally | 10:53 |
pitti | dkessel: ooh | 10:53 |
pitti | dkessel: I'm afraid autopkgtest is fairly buggy on precise still | 10:53 |
pitti | dkessel: so you might want to run prepare-testbed and run the full test in kvm | 10:54 |
mvo | pitti: interessting, its doing ok for me, the testsuite is run before each upload but maybe my box is the only system :/ | 10:54 |
pitti | dkessel: and for the first test you can just call debian/tests/yourtestname and ensure that succeeds, doesn't print out anything on stderr, and exits with 0 | 10:54 |
pitti | mvo: in fact, it seems the recent runs fail even before that | 10:55 |
pitti | mvo: a while ago the tests actually ran, with two dozen errors | 10:55 |
pitti | now it just says "please run sudo update-apt-xapian-index", and then aborts | 10:55 |
smartboyhw | pitti, I might try libatk1.0-dev | 10:56 |
pitti | smartboyhw: dkessel is working on that one | 10:57 |
pitti | smartboyhw: see http://pad.ubuntu.com/testing-hackfest-20121213 | 10:57 |
smartboyhw | pitti, ok let me choose another one | 10:57 |
pitti | smartboyhw: gdk-pixbuf perhaps, or pango? | 10:57 |
pitti | hang on, pango ought to have one already | 10:58 |
pitti | ah, it's missing an XS-Testsuite: header; I'll fix that | 10:58 |
smartboyhw | pitti, uh is there anyone who are working on libarchive-dev? | 10:58 |
pitti | smartboyhw: we have one already, see https://jenkins.qa.ubuntu.com/view/Raring/view/AutoPkgTest/ | 10:58 |
* pitti updates wiki page | 10:59 | |
smartboyhw | pitti, please update:( | 10:59 |
pitti | done | 10:59 |
dkessel | pitti: i have run the test already and verified the results. now i am stuck on how to run my local branch on the testbed | 10:59 |
pitti | dkessel: did you run "prepare-testbed -r raring" already? | 11:00 |
dkessel | do i just have to run that from the package source directory? | 11:01 |
jibel | dkessel, no, first branch the project auto-package-testing: bzr branch lp:auto-package-testing | 11:02 |
dkessel | pitti: ^ | 11:02 |
jibel | dkessel, then cd auto-package-testing | 11:02 |
jibel | dkessel, and prepare a test environment with: | 11:02 |
smartboyhw | pitti, well then: How about libsecret-1-dev? | 11:02 |
dkessel | jibel: ok, i did the preparation steps yesterday | 11:02 |
pitti | smartboyhw: sounds good! I'll make a note in the pad | 11:03 |
smartboyhw | pitti, :) | 11:03 |
jibel | dkessel, ok, did you upload a branch with your tests on bzr or its on your local disk? | 11:04 |
dkessel | jibel: i have done the preparation, but how do i run my test for libatk on it? | 11:05 |
dkessel | oh. on my local disk | 11:05 |
pitti | dkessel: push it to LP under e. g. lp:~dkessel/ubuntu/raring/atk1.0/autopkgtest | 11:09 |
pitti | dkessel: and then you can call run-adt-test -r raring -b lp:~dkessel/ubuntu/raring/atk1.0/autopkgtest | 11:10 |
pitti | dkessel: sorry, run it like that to avoid building the package again: | 11:10 |
pitti | run-adt-test -r raring -b lp:~dkessel/ubuntu/raring/atk1.0/autopkgtest atk1.0 | 11:10 |
pitti | (this is all in the documentation, btw) | 11:10 |
jibel | dkessel, or you can start the VM with: ./bin/run-adt-test -kl | 11:11 |
jibel | this will start the test env and log you in | 11:12 |
jibel | then copy your files from your local drive | 11:12 |
jibel | we should definitely add an option to copy local tests or use any VCS | 11:13 |
pitti | yeah, that'd be convenient | 11:13 |
dkessel | pitti: ./bin/run-adt-test -r raring -b lp:~dkessel/ubuntu/raring/atk1.0/autopkgtest atk1.0 fails with: ./bin/run-adt-test: 2: /home/daniel/workspace/auto-package-testing/bin/../etc/config: distro-info: not found | 11:18 |
pitti | dkessel: ah, please instaoo the "distro-info" package | 11:19 |
pitti | "install" | 11:19 |
jibel | dkessel, it is documented here http://bazaar.launchpad.net/~auto-package-testing-dev/auto-package-testing/trunk/view/head:/doc/USAGE.md | 11:20 |
=== _salem is now known as salem_ | ||
dkessel | jibel: thanks. i read http://developer.ubuntu.com/packaging/html/auto-pkg-test.html and that is missing distro-info | 11:23 |
dkessel | jibel: if run-adt-test just outputs "2012-12-13 12:24:06: Info: Cleaning up " - does that mean the test has completed without errors? | 11:24 |
pitti | check "echo $?", it should be 0 | 11:25 |
jibel | dkessel, if it's all you got from the run it's suspicious | 11:25 |
pitti | err yes, it should certainly output two screenfuls of pages before that | 11:25 |
jibel | dkessel, what's the full output, and yeah return status as pitti said | 11:25 |
dkessel | jibel, pitti: that is the full output.... | 11:26 |
jibel | dkessel, can you run again with -d and paste the output | 11:27 |
jibel | ? | 11:27 |
dkessel | apperently, it tries to load the image from /tmp. well - i rebooted since yesterday ;) ok, i rerun the prepare step.... but maybe it should fail with an error if it does not find the image ;) | 11:29 |
jibel | dkessel, right, could you file a bug against https://bugs.launchpad.net/auto-package-testing please ? | 11:30 |
jibel | you can change the location of the image cache in the config file. | 11:31 |
smartboyhw | pitti, a question: Would it be easier for me to use the tests available originally in libsecret? | 11:35 |
jibel | mvo, should apt-clone testsuite run as root or normal user? | 11:35 |
pitti | smartboyhw: if you can convince the upstream tests to run against the installed library instaed of the ones in the source tree, that's ideal of course | 11:36 |
pitti | smartboyhw: but that is orthogonal to adding a compile/link/run test, so you might want to do both (one debian/tests/upstream and a debian/tests/compile) | 11:36 |
smartboyhw | pitti, ok | 11:37 |
pitti | smartboyhw: but as long as the upstream tests run during the package build, they are not really that important to be run again in autopkgtest | 11:37 |
pitti | it cannot hurt of course, so if you can make it work easily, please do :) | 11:37 |
dkessel | jibel: i filed bug 1089885 | 11:47 |
ubot5 | Launchpad bug 1089885 in Auto Package Testing "run-adt-test silently fails when test image is missing" [Undecided,New] https://launchpad.net/bugs/1089885 | 11:47 |
jibel | dkessel, thanks | 11:48 |
smartboyhw | pitti, sorry for asking (quite stupid): So in http://people.gnome.org/~stefw/libsecret-docs/c-examples.html#c-schema-example it says I need to have a header..... so I need to create another header right? | 11:48 |
jibel | update-manager is green again \o/ | 11:48 |
pitti | smartboyhw: no, I don't think you'll need that for a simple compile test; this isn't installed anywhere | 11:52 |
smartboyhw | pitti, then I need my OWN schema...... | 11:52 |
* smartboyhw does not know how to do it....Beginner C guy here | 11:52 | |
* jibel -> lunch | 11:56 | |
pitti | jibel: enjoy! I guess you won't need the intro session :) | 11:57 |
pitti | dholbach: ok, I have an intro prepped | 11:57 |
pitti | jibel: udisks is green again, too! | 11:59 |
pitti | ♩ another one bites the dust ♫ ♬ | 11:59 |
pitti | grabbing u-d-common next | 11:59 |
dkessel | pitti: libakt's debian/control said i should edit debian/control.in instead... so i did that - now run-adt test fails with: "no such directory: ./debian/control" | 11:59 |
pitti | dkessel: that's right, you need to edit debian/control.in and then run "debclean" | 12:00 |
pitti | dkessel: hm, does debian/control exist? | 12:00 |
dkessel | pitti: ... and then commit and push again i guess | 12:00 |
pitti | right | 12:00 |
pitti | but ensure that the XS- header is in debian/control as well; debclean should do that, and then you should see it in bzr diff | 12:01 |
pitti | ATTENTION PLEASE | 12:01 |
pitti | I prepared a short intro about today's hackfest; for those who know about autopkgtest or are already knee-deep in it this won't be of much use, but if we have any newcomers here I'll do it now as dholbach announced in his blog | 12:02 |
pitti | so can you please wave if you are here for the hackfest? | 12:02 |
* dkessel (needlessly) waves :) | 12:04 | |
* pitti hugs dkessel | 12:04 | |
pitti | smartboyhw as well, but he's also already working on a test | 12:04 |
pitti | ok, so let's cut this short | 12:05 |
pitti | dkessel, smartboyhw: I just want to ensure that you have all the links | 12:06 |
pitti | this document describes all the details about how to create and run tests, including a detailled example of a compile/link/run library test: | 12:06 |
pitti | http://developer.ubuntu.com/packaging/html/auto-pkg-test.html | 12:06 |
pitti | If you create a test, I suggest that for the first iterations you just run the debian/tests/yourtestname script on your local machine until that succeeds. Once it does, you need to run the test in a virtual machine, to ensure that it also works in that environment and you specified all necessary test dependencies. | 12:06 |
pitti | If you start working on a package, please add a note to our coordination pad, to avoid stepping on each other's toes: | 12:06 |
pitti | http://pad.ubuntu.com/testing-hackfest-20121213 | 12:06 |
pitti | and finally, you can see the current tests and their status here: | 12:07 |
pitti | https://jenkins.qa.ubuntu.com/view/Raring/view/AutoPkgTest/ | 12:07 |
pitti | *tock tock*, is this thing on? | 12:09 |
dkessel | yup | 12:09 |
dkessel | smartboyhw - can you access http://pad.ubuntu.com/testing-hackfest-20121213 ? | 12:10 |
dkessel | becaus i can't | 12:10 |
mvo | jibel: \o/ | 12:13 |
* dkessel waits.... and waits... | 12:16 | |
dkessel | how can i stop the test image from upgrading the kernel each time it starts? | 12:16 |
pitti | there's no switch for that, I'm afraid | 12:17 |
pitti | it's an unfortunate day, it seems we didn't get fresh cloud images today | 12:17 |
smartboyhw | dkessel, hmm i can:P | 12:21 |
dkessel | pitti: :( hm, i did debclean, commit and push - and it still complains about missing "debian/control" when i run: ./bin/run-adt-test -r raring -b lp:~dkessel/ubuntu/raring/atk1.0/autopkgtest | 12:22 |
pitti | dkessel: right, so do you have a debian/control ? | 12:24 |
pitti | (in your source package branch) | 12:24 |
pitti | dkessel: what's the branch you pushed this to? | 12:25 |
pitti | I'll have a look on LP | 12:25 |
pitti | oh, got it, https://code.launchpad.net/~d-kessel/ubuntu/raring/atk1.0/autopkgtest | 12:25 |
pitti | ok, the file is there | 12:26 |
* pitti runs it himself | 12:26 | |
smartboyhw | pitti, hmm is there more easier packages to work with autopkgtest? I found this one quite hard, if I were to test it's functions | 12:26 |
pitti | smartboyhw: poppler-dev should be more straightforward; you need to add a test dependency to a packge that ships a PDF, and can then try to open that with libpoppler, and check a few basic properties with it | 12:28 |
smartboyhw | pitti, OK | 12:28 |
pitti | smartboyhw: that should provide quite a nice smoketest, and poppler doesn't have the kind of complicated callback API that libsecret has | 12:28 |
pitti | smartboyhw: but even with libsecret, note that you don't need to do anything complicated | 12:29 |
pitti | smartboyhw: any simple library call will do, which proves that you can compile against the library | 12:29 |
pitti | dkessel: I get | 12:29 |
pitti | chmod: cannot access ‘/tmp/tmp.eVTEXdaDCI/ubtree0-ubtree/debian/tests/build’: No such file or directory | 12:30 |
pitti | adt-run: unexpected error: failed to chmod +x /tmp/tmp.eVTEXdaDCI/ubtree0-ubtree/debian/tests/build | 12:30 |
pitti | dkessel: which is because your debian/tests/control says "Tests: build" | 12:30 |
pitti | dkessel: but you named your script debian/tests/atk1.0-dev_test | 12:30 |
dkessel | aaaah so thats that that line means.... | 12:30 |
dkessel | :) | 12:30 |
pitti | smartboyhw: poppler-dev should be more straightforward; you need to add a test dependency to a packge that ships a PDF, and can then try to open that with libpoppler, and check a few basic properties with itdebian/tests/atk1.0-dev_test debian/tests/build" | 12:30 |
pitti | erk, what was that | 12:30 |
smartboyhw | pitti, yes I get it... | 12:30 |
pitti | smartboyhw: sorry, weechat failure | 12:31 |
dkessel | pitti: i still wonder why you get a different error message | 12:31 |
pitti | dkessel: so I suggest to do bzr mv debian/tests/atk1.0-dev_test debian/tests/build | 12:31 |
pitti | dkessel: as this is the atk1.0 package, there is no doubt which package you are testing :) | 12:31 |
pitti | dkessel: yeah :( I'm on raring, you are on precise; as I said autopkgtest on precise still has a number of bugs, that might be the problem | 12:32 |
pitti | dkessel: perhaps you can try "wget http://archive.ubuntu.com/ubuntu/pool/main/a/autopkgtest/autopkgtest_2.2.3ubuntu1_all.deb" | 12:32 |
pitti | dkessel: and then "sudo dpkg -i autopkgtest_2.2.3ubuntu1_all.deb" | 12:32 |
pitti | dkessel: the current versino should work just fine on precise | 12:33 |
pitti | dkessel: ah no, I'm talking gibberish, sorry | 12:33 |
pitti | dkessel: if you created a raring VM, it'll use autopkgtest in that VM, not locally | 12:33 |
* dkessel running kernel upgrade :) | 12:34 | |
smartboyhw | pitti, ok i will change my focus to poppler-dev now | 12:35 |
dkessel | pitti: can you try running it again? i fixed the file name and it still fails with "missing debian/control" here... | 12:36 |
pitti | dkessel: sure! running | 12:37 |
mvo | silly question - I have my testbed now but how do I log into it to try to figure out how to fix the tests in a more interactive way | 12:37 |
pitti | mvo: what I usually do is "run-adt-test -sk ..." | 12:37 |
pitti | mvo: that will install all the deps, run the tests, and then keep the VM running, and print an ssh command to log into it | 12:38 |
mvo | aha, nice | 12:38 |
=== Ursinha-afk is now known as Ursinha | ||
pitti | mvo: or "run-adt-test -sl" and then do the apt-get source, install deps etc. yourself | 12:38 |
pitti | mvo: (-s is just for speed if you have lots of RAM; -k vs. -l is the important bit here) | 12:38 |
=== yofel_ is now known as yofel | ||
pitti | but -s rocks; it's great to see 200 MB of dependencies getting unpacked and configured in like 3 seconds | 12:39 |
smartboyhw | pitti, the sad thing is: I can't seem to think of any packages that includes a PDF file in mind:P | 12:39 |
mvo | sweet | 12:39 |
pitti | dkessel: Fortschritt! http://paste.ubuntu.com/1429597/ | 12:39 |
pitti | smartboyhw: grep '\.pdf' /var/lib/dpkg/info/*.list | 12:40 |
dkessel | pitti: i guess i have to get interactive too... | 12:41 |
pitti | dkessel: so cups-filter's /usr/share/cups/data/default-testpage.pdf should do | 12:41 |
pitti | dkessel: your debian/tests/control depends on libglib2.0-dev, but you want libatk1.0-dev :) | 12:42 |
pitti | copy & paste FTW | 12:42 |
smartboyhw | pitti, hmm the poppler website is very short of documentation:P | 12:42 |
pitti | dkessel: you did create a raring VM, not a precise one, right? | 12:42 |
dkessel | oopsie | 12:43 |
dkessel | pitti: yes, a raring vm | 12:43 |
pitti | smartboyhw: file:///usr/share/doc/libpoppler-glib-dev/html/poppler/PopplerDocument.html#poppler-document-new-from-file | 12:44 |
pitti | smartboyhw: this looks quite reasonable | 12:44 |
pitti | smartboyhw: i. e. one call to load a file and get a PopplerDocument | 12:45 |
pitti | smartboyhw: then maybe check file:///usr/share/doc/libpoppler-glib-dev/html/poppler/PopplerDocument.html#poppler-document-get-n-pages for the expected number of pages | 12:45 |
smartboyhw | pitti, OK | 12:45 |
pitti | smartboyhw: and that should actually be enough for a c/l/r test | 12:46 |
smartboyhw | pitti, OK | 12:46 |
pitti | smartboyhw: then add a dependency to cups-filters for /usr/share/cups/data/default-testpage.pdf, and there you go :) | 12:46 |
smartboyhw | :) | 12:46 |
mvo | pitti: the tests are running now, I keep you updated | 12:49 |
pitti | jibel: hm, right now jenkins runs udisks amd64 three times in parallel; comment est-ce possible? | 12:49 |
pitti | mvo: you got beyond the "run xapian-update" thingy? | 12:50 |
jibel | pitti, ce n'est pas possible, it's a UI bug | 12:50 |
dkessel | pitti: it should work now - but i can't verify because of the error in the VM... | 12:50 |
* pitti turns the crank on dkessel's branch | 12:50 | |
dkessel | *sigh* | 12:51 |
jibel | it uses a blinking ball for all the jobs with the same name in the latest builds panel | 12:51 |
pitti | jibel: yes, but I see it on three different slaves on the left side | 12:51 |
pitti | alderamin, aldebaran, and wazn | 12:52 |
pitti | dkessel: Bazinga! | 12:52 |
pitti | dkessel: thank you for having shopped at Ubuntu QA Labs Inc. Please collect your free beer at the counter now! | 12:53 |
pitti | dkessel: so I'll merge and upload this, and commit it to Debian too | 12:53 |
jibel | pitti, on the left side, there is 1 for amd64, 1 for i386 and a virtual job with a role of controller for jenkins internal mechanics and which will consolidate the results of all the architectures | 12:53 |
dkessel | pitti: yay! thanks for the help - I'll have lunch before the beer though :) | 12:55 |
jibel | I restarted udisk because of this hashsum mismatch error again | 12:56 |
pitti | jibel: ah, thanks; I was about to restart it because of the hash sum error, and saw that it was already running | 12:57 |
jibel | I don't know why it happens soooo frequently | 12:57 |
pitti | ah, c'est vert a nouveau! | 12:57 |
* pitti donne une accolade à udisks2 | 12:58 | |
pitti | dkessel: I add -Wall -Werror for extra nitpickyness, if you don't mind | 13:02 |
smartboyhw | Uh I give up..... (/me really needs to go to revise C) | 13:02 |
pitti | jibel: hm, is it still creating new jobs properly? I fixed pango1.0's XS-Testsuite header, and it's in raring already, but not in jenkins | 13:03 |
dkessel | pitti: sure, do that. i'll be back after lunch :) | 13:04 |
pitti | dkessel: nah, I'm doing it while merging; just a FYI | 13:04 |
jibel | pitti, 1.30.1-1ubuntu1? let me check | 13:04 |
pitti | jibel: oui | 13:04 |
jibel | sometimes lillypilly is lagging a bit | 13:04 |
smartboyhw | pitti, I give up... | 13:06 |
smartboyhw | I really like manual testing more:P | 13:06 |
mvo | pitti: well, I need to run it manually, but that I can fix easily so far two test failures | 13:06 |
mvo | (and what looks like a hang :/ | 13:07 |
mvo | (one releated to not having network it seems) | 13:08 |
jibel | pitti, the mirror on lillypilly is a bit behind and only knows about 1.30.1-1 | 13:15 |
pitti | jibel: ah ok, so it's just that; thanks for checking! | 13:15 |
jibel | dkessel, atk1.0 2.7.2-0ubuntu2 \o/ | 13:16 |
jibel | congrats | 13:16 |
zyga | hey mvo, pitti :) | 13:24 |
pitti | hey zyga, how are you? | 13:24 |
zyga | pitti: fine, enjoying my free time :) | 13:25 |
zyga | lots of holiday added up | 13:25 |
pitti | oh, nice; enjoy! | 13:25 |
dkessel | jibel: yay, my first code contribution to OSS :) | 13:39 |
dkessel | so... next package ... who will it be... | 13:39 |
pitti | dkessel: poppler? :-) | 13:41 |
pitti | dkessel: what I suggested to smartboyhw a while ago, with poppler-document-new-from-file() and checking some properties | 13:41 |
dkessel | pitti: ok, i can give it a try | 13:44 |
jibel | pitti, dholbach or anyone could you review https://code.launchpad.net/~jibel/apt-clone/adt-fixes/+merge/139702 | 13:45 |
pitti | looking | 13:49 |
pitti | jibel: did you already test it in run-adt-tes? | 13:50 |
jibel | pitti, I did | 13:50 |
jibel | with --user=ubuntu | 13:50 |
jibel | it fails as root | 13:50 |
pitti | replied | 13:51 |
jibel | pitti, thanks, added a comment in r19 | 13:55 |
jamespage | o/ | 13:56 |
pitti | jibel: ah, perhaps we can also fix the tests to write to stdout? | 13:56 |
pitti | jibel: I usually use | 13:57 |
pitti | unittest.main(testRunner=unittest.TextTestRunner(stream=sys.stdout, verbosity=2)) | 13:57 |
jibel | pitti, this is better indeed. fixing | 13:57 |
pitti | ah, there goes a new green "pango1.0" dot | 13:59 |
jibel | hey jamespage | 14:03 |
jamespage | hey jibel | 14:03 |
jibel | I noticed, dovecot, cyrus, keystone for example have dep8 tests but no testsuite header | 14:03 |
jibel | jamespage, maybe it's worth having a look | 14:04 |
jamespage | jibel, those sounds like some easy wins! | 14:04 |
jamespage | preparing my test bed now! | 14:04 |
pitti | oh, ubuntu-drivers-common failure detects an actual regression in bcmwl-kernel-source | 14:04 |
jibel | pitti, if I set stream=sys.stdout I still have the following warning on stderr | 14:06 |
jibel | WARNING:root:can't add notification-daemon (pkg notification-daemon not marked upgrade) | 14:06 |
jibel | WARNING:root:can't add notify-osd (pkg notify-osd not marked upgrade) | 14:06 |
pitti | jibel: ah; no idea about them; so perhaps leave your >&2 in place for now? | 14:07 |
jibel | pitti, yes, I didn't find where they come from | 14:07 |
smartboyhw | God's sake my internet bandwidth is REALLY SLOW today | 14:08 |
* smartboyhw reminds himself to do all the cadence testing tmr | 14:08 | |
jibel | another easy target in main: html2text | 14:09 |
dholbach | hum | 14:17 |
dholbach | https://jenkins.qa.ubuntu.com/search/?q=raring-adt seems broken | 14:17 |
dholbach | it just shows 21 test-cases | 14:17 |
dholbach | https://jenkins.qa.ubuntu.com/view/Raring/view/AutoPkgTest/ has something like 60 | 14:17 |
* pitti adds another green gdk-pixbuf bullet | 14:17 | |
* dholbach hugs pitti | 14:17 | |
jibel | dholbach, 65 | 14:18 |
jamespage | jibel, pitti - is there a way that I can hack the testbed image to use a proxy? | 14:19 |
dholbach | jibel, ah yes | 14:19 |
pitti | jamespage: you can login with "run-adt-test -sl", then set up the proxy, then run "sudo ./run-adt package" | 14:19 |
pitti | jamespage: (for testing) | 14:20 |
pitti | jamespage: if your test itself wants to set up a proxy, it's got full root privs | 14:20 |
pitti | jamespage: (if you set Restrictions: needs-root) | 14:20 |
jamespage | pitti, no - I just sit on the end of a sucky internet connect and have a relatively well populated local squid-deb-proxy | 14:20 |
pitti | jamespage: oh, I see | 14:20 |
pitti | so prepare-testbed should grow a --mirror option | 14:21 |
pitti | jamespage: ah, wait, proxy not mirror | 14:21 |
jibel | jamespage, in prepare-testbed you can add an apt-proxy directive to the cloud-config script | 14:22 |
pitti | jamespage: so running "http_proxy=... prepare-testbed ..." doesn't work? | 14:22 |
jibel | or apt_proxy whatever it is called | 14:22 |
pitti | jibel: I'm a bit confused about http://10.98.0.1:8080/view/Raring/view/AutoPkgTest/job/raring-adt-setup-testbed/52/ | 14:24 |
pitti | jibel: the job is red, but all 8 subtests are green | 14:25 |
jibel | pitti, the provisioning failed on alderamin/i386 because a port was already in use, I forced a rebuilt of the failing image instead of everything | 14:27 |
jibel | in this case the status is not propagated to the master job | 14:27 |
pitti | jibel: so that'll just fix itself tomorrow then | 14:29 |
* jamespage recreates his testbed | 14:32 | |
jamespage | jibel, pitti: thanks for that tip - hopefully that will make things a bit quicker :-) | 14:39 |
jibel | pitti, a trivial one https://code.launchpad.net/~jibel/mawk/xs-testsuite tested with adt-run-test | 14:58 |
jibel | I don't know why I cannot propose a merge | 14:59 |
pitti | ok, u-d-common and bcmwl should be sorted out now | 15:03 |
pitti | jibel: looking | 15:06 |
pitti | jibel: peut-être que c'est un upstream branch | 15:07 |
pitti | jibel: is it against an UDD packaging branch? | 15:07 |
pitti | jibel: can you please refer to debian bug 692662 in the changelog? and then upload | 15:08 |
ubot5 | Debian bug 692662 in src:mawk "mawk: missing (XS-)Testsuite: autopkgtest header in source stanza" [Wishlist,Open] http://bugs.debian.org/692662 | 15:08 |
* dkessel run-adt-tests his poppler test | 15:09 | |
pitti | jibel: oh, need sponsoring? je peux le faire | 15:09 |
pitti | dkessel: *daumendrueck* | 15:09 |
dkessel | pitti: "no such file or directory: '.debian/control' ....... again | 15:10 |
pitti | dkessel: but ".debian" looks suspicious indeed | 15:11 |
pitti | it should be ./debian or debian but not .debian | 15:11 |
pitti | dkessel: how do you run this exactly? | 15:11 |
dkessel | oh... i typed. it ... spelling error on my side | 15:11 |
pitti | run-adt-test -sb lp:~d-kessel/ubuntu/raring/atk1.0/autopkgtest atk1.0 | 15:11 |
pitti | that's what I ran | 15:11 |
pitti | you probably need a -r raring | 15:12 |
dkessel | ./bin/run-adt-test -r raring -b lp:~dkessel/ubuntu/raring/poppler/autopkgtest | 15:12 |
dkessel | i'll paste the error log | 15:12 |
dkessel | ah... d-kessel maybe? | 15:13 |
jibel | pitti, I added a ref to debian bug 692662, and it should probably go directly there instead of ubuntu | 15:13 |
ubot5 | Debian bug 692662 in src:mawk "mawk: missing (XS-)Testsuite: autopkgtest header in source stanza" [Wishlist,Open] http://bugs.debian.org/692662 | 15:13 |
pitti | jibel: fine to upload IMHO; it's very likely to get fixed with the next debian upload anyway, so we can sync it back | 15:13 |
* pitti looks at samba4 | 15:14 | |
pitti | jibel: ah, did you upload? I thought you weren't core-dev | 15:16 |
jibel | pitti, no I didn't | 15:16 |
jibel | and I am not a core dev | 15:16 |
pitti | J'ai voir ton changement à pad | 15:16 |
jibel | ah sorry | 15:17 |
pitti | j'ai vu | 15:17 |
jibel | reverting | 15:17 |
pitti | jibel: I can upload it | 15:17 |
jibel | I also readded apt-clone because it is not uploaded | 15:17 |
pitti | jibel: you can push those to lp:~jibel/ubuntu/raring/pkgname/autopkgtest (i. e. to a distro branch) | 15:18 |
pitti | then you can MP | 15:18 |
pitti | (but don't worry about this for that one) | 15:18 |
pitti | jibel: done both | 15:20 |
dkessel | hm... how could i have stopped run-adt-test from building the package? | 15:21 |
dkessel | poppler takes a while to build :/ | 15:21 |
jibel | pitti, danke schön! | 15:21 |
pitti | dkessel: specify "poppler" as an additional argument at the end | 15:22 |
pitti | dkessel: it's the difference between "-b branch" and "-b branch pkgname" | 15:22 |
pitti | ah, samba4 est très facile | 15:26 |
pitti | dkessel: want me to run a branch for you? | 15:26 |
dkessel | pitti: i had a missing dependency, i am now trying to run my branch again... | 15:45 |
dkessel | pitti: i guess my test just passed | 15:52 |
pitti | \o/ | 15:52 |
pitti | dkessel: how did you get around the failure? | 15:52 |
dkessel | i can't really tell... | 15:53 |
dkessel | pitti: want to review https://code.launchpad.net/~d-kessel/ubuntu/raring/poppler/autopkgtest ? | 15:54 |
pitti | dkessel: running | 15:55 |
dkessel | pitti: i used -Wall and -Werrors this time :) | 15:55 |
zyga | balloons: hey | 15:55 |
zyga | balloons: would you mind if the hw cert team used this channel for all public discussions? | 15:56 |
pitti | dkessel: I think you should replace libpoppler-dev, libglib2.0-dev with libpoppler-glib-dev | 15:56 |
dkessel | ok | 15:56 |
pitti | dkessel: actually no, you don't use poppler-glib, but just libpoppler-dev | 15:56 |
pitti | dkessel: so drop the glib test dep, as you don't use it | 15:56 |
pitti | dkessel: and your pkg-config calls libs poppler-glib which won't be there without libpoppler-glib-dev | 15:57 |
dkessel | pitti: *confused* | 15:58 |
pitti | dkessel: so, your code uses the poppler API; your pkg-config call wants poppler-glib; and your test dependency pull in poppler and glib | 15:58 |
pitti | make up your mind :) | 15:58 |
pitti | dkessel: in fact, if you want to write two tests, one for poppler and one for poppler-glib (with respective pkg-config and dependencies), that's even better | 15:59 |
pitti | oh hm, libpoppler-dev doesn't have header files | 16:00 |
pitti | thse are in libpoppler-private-dev | 16:00 |
dkessel | yup | 16:01 |
dkessel | pitti: that's why i went and searched for poppler.h ;) | 16:01 |
pitti | dkessel: so I'm a bit confused what pulls in libpoppler-glib-dev | 16:01 |
dkessel | that is where i found poppler.h | 16:02 |
pitti | yes, but why does it get installed? | 16:02 |
pitti | ooh | 16:02 |
pitti | http://bazaar.launchpad.net/~d-kessel/ubuntu/raring/poppler/autopkgtest/revision/125 | 16:02 |
pitti | I was only looking at r124 | 16:02 |
pitti | dkessel: so yes, libpoppler-dev and libglib2.0-dev shoudl go | 16:03 |
pitti | but I can do that during merge | 16:03 |
pitti | dkessel: thanks! I'll upload this | 16:03 |
dkessel | pitti: you're welcome. I am still not sure if my work saved anyone any time though ;) | 16:04 |
pitti | oh, it does | 16:04 |
pitti | and also, I hope you had some fun with it and you learned something | 16:04 |
pitti | +# Author: Martin Pitt <martin.pitt@ubuntu.com> | 16:05 |
pitti | dkessel: ^ you might want to fix that :) | 16:05 |
dkessel | sure, I learned some stuff. And I think "two packages" was my goal for today :) | 16:06 |
dkessel | ooh :) | 16:06 |
dkessel | I will :D | 16:06 |
pitti | then that was a good day, I say :) | 16:06 |
pitti | dkessel: https://jenkins.qa.ubuntu.com/view/Raring/view/AutoPkgTest/job/raring-adt-atk1.0/ | 16:07 |
balloons | zyga, no of course not :-) | 16:07 |
balloons | it would be a good thing for people to see and know you | 16:07 |
zyga | roadmr: it's set then | 16:08 |
zyga | ara: o/ | 16:08 |
zyga | brendand: :) | 16:08 |
roadmr | zyga: awesome! | 16:08 |
zyga | :) | 16:08 |
dkessel | pitti: juhu, grüne lampen :) | 16:08 |
zyga | one step closer to building a community | 16:08 |
dkessel | pitti: author is fixed | 16:08 |
dkessel | is this wiki page updated after the hackfest? https://wiki.ubuntu.com/QATeam/RequiredTests | 16:09 |
pitti | dkessel: danke | 16:09 |
dkessel | i might do some more tests on my christmas holidays... | 16:09 |
jtaylor | does the jenkins instance only build from the archive packages or can one also let it use branches to avoid uploading a package for e.g. a faulty test? | 16:10 |
jibel | dkessel, not completely up to date, I'll update it with today's effort | 16:10 |
pitti | jtaylor: only from the archive | 16:10 |
dkessel | oh and i have another question... are there any packages which require tests written in java or c#? | 16:11 |
jtaylor | dkessel: don'T limit yourself to the list, there are plenty packages in universe that need love too :) | 16:12 |
pitti | in general it doesn't matter which language the tests are written in | 16:12 |
pitti | but scripts are considerably easier of course as they don't have huge dependencies and don't need to be compiled | 16:12 |
dkessel | i understand | 16:13 |
jtaylor | c# has an interpreter | 16:13 |
jtaylor | which should make writting tests a bit easier | 16:14 |
jtaylor | I think it even works with hashbangs | 16:14 |
dkessel | jtaylor: yeah, but I don't think I could pick one package myself that may be important to anyone... | 16:15 |
dkessel | I guess I will just check by here when I want to do something more :) | 16:18 |
dkessel | alright, see you guys. It has been an interesting event. | 16:19 |
balloons | dkessel, thanks for hanging out with us! | 16:19 |
jibel | more green dots with ubuntu-drivers-common | 16:39 |
pitti | \o/ | 16:41 |
=== josepht_ is now known as josepht | ||
pitti | unfortuantely samba4 fails to build in a weird way, I'm fighting with that now | 16:41 |
jibel | I'm on pbuilder and fighting with signatures :) | 16:48 |
jibel | 68 tests, 2 to go to 70 | 16:50 |
pitti | dholbach, jibel: FYI, I need to leave in about 30 mins; in case you want to give an intro again, http://people.canonical.com/~pitti/tmp/intro.txt was what I had prepared | 16:51 |
dholbach | pitti, awesome - have a great rest of your day! | 16:52 |
jibel | pitti, great, thanks for your help and enjoy the evening! | 16:52 |
pitti | thanks! | 16:53 |
=== salem_ is now known as _salem | ||
=== _salem is now known as salem_ | ||
=== salem_ is now known as _salem | ||
pitti | good night everyone! | 17:25 |
jibel | good night pinky | 17:27 |
jibel | pitti | 17:27 |
pitti | uploaded a fix for the samba4 FTBFS, so hopefully that should go green, too | 17:27 |
jibel | great way to end the day | 17:27 |
dholbach | pitti, jibel: http://daniel.holba.ch/autopkgtests/ (I finally picked the right place to get the data about the number of raring test cases from :-)) | 17:33 |
dholbach | getting closer to 70 :) | 17:33 |
=== _salem is now known as salem_ | ||
phillw | balloons: are you about? | 19:27 |
balloons | phillw, indeed | 19:27 |
phillw | I've got a bug with installer, but a weird one that seems to affect only KVM running raring installs... as KVM is okay with p & Q I'm struggling to work out what to file it against? | 19:28 |
balloons | you can cite ubiquity if you wish, but include kvm too | 19:29 |
balloons | did you try the new version in the ppa? | 19:29 |
balloons | not sure it landed in raring yet | 19:29 |
balloons | I didn't check | 19:29 |
balloons | kvm I mean | 19:29 |
phillw | I'm running KVM on qua | 19:29 |
phillw | quanatal | 19:29 |
balloons | ahh | 19:29 |
balloons | then you haven't got the new stuff | 19:29 |
phillw | KVM in Q happily installs L, P & Q as VM's... on when trying to install R do I get the video corruption | 19:30 |
phillw | s/on/only/ | 19:30 |
balloons | ohh | 19:31 |
balloons | sorry I misunderstood your question | 19:31 |
balloons | yes, blame ubiquity.. haha | 19:31 |
balloons | have a look first though | 19:32 |
phillw | I do recall an issue with video corruption a while back. but thought it had been resolved. | 19:32 |
balloons | there is the blank screen bug already filed | 19:32 |
phillw | this is not blank... just corrupt. | 19:32 |
phillw | I'll re-assign it to ubiquity, I had it against lubuntu-artwork, but it also affects xubuntu - so deffo not lubuntu fault :) | 19:33 |
phillw | balloons: hmm, not a ubiquity bug :/ | 19:40 |
phillw | video corruption still exists after installing via alternate. | 19:40 |
balloons | hmm corruption when? | 19:40 |
phillw | look like an X bug? | 19:40 |
balloons | yes.. video corruption is not the installer | 19:40 |
balloons | well, we can pinpoint it more perhaps | 19:41 |
balloons | i guess xorg is fine | 19:41 |
phillw | Very odd bug... the application windows work fine... | 19:41 |
balloons | it might be plymouth | 19:41 |
phillw | background is corrupt, and when installing from desktop the world-map where you choose where you are is corrupt. | 19:42 |
=== salem_ is now known as _salem | ||
balloons | ok, so might be a compiz / ubiquity bug | 19:47 |
balloons | does it do that no matter how you install it? | 19:47 |
balloons | aka, install now, or boot into desktop, then launch installeR/ | 19:47 |
balloons | they are different! | 19:48 |
phillw | balloons: I battled through a desktop installation and get the corruption in the end product. alternate is less painful to install, but the resultant system is the same. | 19:50 |
phillw | bug 1088692 has now got some screen shots | 19:51 |
ubot5 | Launchpad bug 1088692 in ubiquity (Ubuntu) "Screen corrupt when installing 13.04 using KVM" [Undecided,New] https://launchpad.net/bugs/1088692 | 19:51 |
balloons | kk | 19:51 |
balloons | hello letozaf_ | 19:51 |
letozaf_ | hello :) | 19:51 |
letozaf_ | what are you guys doing ? | 19:51 |
phillw | letozaf_: bug hunting and there is an automated hack-fest going on as well | 19:52 |
letozaf_ | I guess I am late for the hack-fest | 19:55 |
phillw | letozaf_: shouldn't be, I believe that jibel is covering now :) | 19:57 |
phillw | balloons: if you can have a look & a think. I'll be be back in 1 hour. Thanks. | 19:59 |
balloons | letozaf_, yes, also working on docs | 20:00 |
* SergioMeneses im back | 20:00 | |
balloons | have you seen the ubuntu.com/community page? | 20:00 |
SergioMeneses | hey hey guys! | 20:00 |
balloons | hello SergioMeneses, I owe you an email still | 20:00 |
letozaf_ | I'm looking at in now | 20:01 |
SergioMeneses | balloons, perfect! Im working on lococouncil things but when I have time enough I see it :D | 20:01 |
balloons | well, it's missing a quality page.. and as part of a drive to get all the pages updated, we have a pad: http://pad.ubuntu.com/communitywebsite-contribute-quality | 20:01 |
SergioMeneses | I was reading dpm's email about the new documents, it sounds really nice | 20:01 |
balloons | full details are on the wiki page: https://wiki.ubuntu.com/CommunityWebsite | 20:02 |
SergioMeneses | balloons, Im working on another http://pad.ubuntu.com/communitywebsite-help-meeting-locos | 20:02 |
balloons | I've started putting some stuff on there, but I'd like to get some of you folks a pass a editing it too | 20:02 |
balloons | SergioMeneses, excellent! | 20:02 |
balloons | phillw, when you get back you should have a look as well | 20:02 |
balloons | :-p | 20:02 |
* letozaf_ is reading | 20:03 | |
SergioMeneses | omg balloons said my name in his email, ty | 20:04 |
balloons | SergioMeneses, yw | 20:05 |
Noskcaj | hello, somehow my ibook is capable of being a computer so i am using it. unfortunatly i can do very little else as it has 376mb of ram | 20:15 |
letozaf_ | balloons: Well what can I do to help ? | 20:15 |
balloons | Noskcaj, nice | 20:15 |
balloons | old machines are fun sometimes | 20:15 |
balloons | letozaf_, well, feel free to edit the pad or just comment on things in it | 20:16 |
Noskcaj | balloons, i dont believe you. i will joiin the pad | 20:16 |
balloons | It's all my writing, and I haven't really attempted to edit it down yet | 20:16 |
balloons | There's a secret inside the document.. Since it wanted links in the resources section, I thought I would make a new wiki page that could stay static and contain things of interest | 20:18 |
vkottilil | hi, some questions on autopkgtest, this is my first time | 20:20 |
balloons | vkottilil, sure | 20:20 |
vkottilil | where should I create debian/tests/control ? | 20:21 |
Noskcaj | how do i scroll in the pad? it just opens and closes the chat bar | 20:21 |
jtaylor | vkottilil: in the debian package source | 20:22 |
jtaylor | vkottilil: obtained e.g. with apt-get source package | 20:22 |
balloons | Noskcaj, hmm.. my screen is too big, no scrollbar | 20:22 |
balloons | let's see | 20:22 |
balloons | ugh, yea, I can't use the scrollbars | 20:22 |
balloons | I can scroll with my wheel or arrow pad | 20:22 |
vkottilil | jtaylor: thanks, i just completed ./bin/prepare-testbed -r raring amd64 | 20:22 |
Noskcaj | i found a workaround, select the text and pull down | 20:23 |
balloons | yea, similar idea | 20:23 |
balloons | pg up / pg down | 20:23 |
vkottilil | next, I want to run the sample test scripts already posted on the wiki for glib2.0 | 20:24 |
balloons | vkottilil, ok, have you installed autopkgtest? | 20:25 |
balloons | sudo apt-get install autopkgtest | 20:26 |
vkottilil | balloons, I thought prepare-testbed would do that? | 20:26 |
balloons | ahh | 20:26 |
balloons | if your using the lp branch, it's much nicer | 20:27 |
balloons | prepare testbed will setup a vm to run tests | 20:27 |
balloons | once it's setup, you can run the test with run-adt-test | 20:27 |
vkottilil | balloons: I am using lp branch - bzr branch lp:auto-package-testing | 20:28 |
balloons | ok, are you following this? http://developer.ubuntu.com/packaging/html/auto-pkg-test.html | 20:28 |
balloons | if so, tell me where your getting stuck ;-) | 20:28 |
vkottilil | yes | 20:28 |
vkottilil | i finished ./bin/prepare-testbed -r raring amd64, next want to run the sample script already posted for glib2.0 | 20:30 |
balloons | ok, so run this: | 20:31 |
balloons | ./bin/run-adt-test -r quantal -a amd64 libglib2.0-0 | 20:32 |
balloons | I *believe* that should work for you | 20:32 |
jtaylor | you probably want raring | 20:32 |
balloons | bah | 20:32 |
balloons | ./bin/run-adt-test -r raring -a amd64 libglib2.0-0 | 20:32 |
balloons | did we loose you vkottilil_ ? | 20:33 |
vkottilil_ | balloons:sorry, i lost the connection | 20:33 |
balloons | ./bin/run-adt-test -r raring -a amd64 libglib2.0-0 | 20:34 |
balloons | I believe that should work for you | 20:34 |
vkottilil_ | where should I put the script for glib2.0 to run it? | 20:35 |
jtaylor | vkottilil_: the run-adt way would be to branch glib2.0, push your changes to your local lp space and use the -b option | 20:37 |
jtaylor | vkottilil_: bzr branch lp:ubuntu/glib2.0; <edit, commit>; bzr push lp:~username/+junk/branchname; run-adt-test ... -b lp:~username/+junk/branchname glib2.0 | 20:38 |
vkottilil_ | ok, I am not changing any thing right now, just wanted to do the whole workflow to learn it, before making changes | 20:38 |
jtaylor | you can run the script locally for development of the script | 20:39 |
vkottilil_ | ok - thats what i was looking at | 20:39 |
jtaylor | there is adt-run as mentioned in the wiki | 20:40 |
jtaylor | also sadtrunner.py by jwilk | 20:40 |
jtaylor | I use this awful thing for chroot tests http://paste.ubuntu.com/1430489/ | 20:40 |
jtaylor | https://bitbucket.org/jwilk/debian-misc/src/tip/sadt | 20:40 |
jtaylor | its all just bikesheeding how to run the scripts, the script itself is what is important | 20:41 |
vkottilil_ | ok, great, I will look at these as a starting point; | 20:41 |
vkottilil_ | once I can run the already posted scripts, I can add more calls/apis from a pkg and try to run them | 20:42 |
=== toddyhb is now known as toddy | ||
vkottilil_ | thank you jtaylor, balloons. | 20:45 |
balloons | ty vkottilil_ .. best of luck.. | 20:48 |
vkottilil_ | sure, thanks | 20:49 |
=== micahg_ is now known as micahg | ||
phillw | balloons: I'm back :) Don't worry about links and notes from todays hack-fest... I have it all logged to review once I've sorted out why I cannot test Raring with KVM.... | 21:08 |
balloons | phillw, no worries.. come join the fun on the pad: | 21:09 |
balloons | http://pad.ubuntu.com/communitywebsite-contribute-quality | 21:09 |
balloons | if you haven't seen it, feedback and editing please | 21:09 |
jtaylor | can tests be symlinks to other tests? | 21:11 |
phillw | balloons: I'm there :) | 21:13 |
jtaylor | e.g. to reuse tests for alternatives dependencies | 21:13 |
balloons | :-) | 21:15 |
phillw | balloons: I can confirm that CentOS is also not affected by bug 1086974 Have you any further thoughts on what I should log it against? | 22:02 |
ubot5 | Launchpad bug 1086974 in libguestfs (Ubuntu) "libguestfs: error: cannot find any suitable libguestfs supermin" [Undecided,Fix released] https://launchpad.net/bugs/1086974 | 22:02 |
balloons | ? | 22:03 |
balloons | wrong bug phillw ? | 22:03 |
phillw | oops.. that is my other one! | 22:04 |
phillw | bug 1088692 | 22:05 |
ubot5 | Launchpad bug 1088692 in ubiquity (Ubuntu) "Screen corrupt when installing 13.04 using KVM" [Undecided,New] https://launchpad.net/bugs/1088692 | 22:05 |
SergioMeneses | balloons, phillw you have done a great job http://pad.ubuntu.com/communitywebsite-contribute-quality | 22:08 |
balloons | phillw, lol | 22:08 |
balloons | ok, so on this bug, 1088692 | 22:09 |
balloons | https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/1088692 | 22:09 |
ubot5 | Launchpad bug 1088692 in ubiquity (Ubuntu) "Screen corrupt when installing 13.04 using KVM" [Undecided,New] | 22:09 |
balloons | your on raring, what video and drivers? | 22:09 |
balloons | ohh you see the same corruption on xubuntu? | 22:09 |
balloons | cent os isn't the same kernel | 22:10 |
balloons | does quantal have the same issue? | 22:10 |
phillw | balloons: yes, that is why I removed it being lubuntu only. | 22:10 |
balloons | and vice versa.. does running kvm on quantal (installing raring) work? | 22:10 |
phillw | L, O & Q run perfectly. Only R is affected. | 22:11 |
phillw | and 'P' | 22:11 |
balloons | yes.. but that's when the host os is raring | 22:13 |
phillw | balloons: the host is Quantal. | 22:13 |
balloons | what about when the host os is quantal or precise? | 22:13 |
* balloons rattles head | 22:13 | |
balloons | ok, try host os raring, installing raring | 22:13 |
phillw | I have not installed raring, as the idea of the lesson I'm to hold is that you can test raring on a machine running Quantal? | 22:14 |
balloons | yes.. good stuff | 22:14 |
balloons | anyways, I guess that one is on me to try then | 22:15 |
phillw | I will install a raring 'real' area, but it does fly in the face of using VM's to test if you cannot! | 22:15 |
phillw | There is deffinately something odd with raring. | 22:16 |
phillw | I can get raring to install with the quantal version (from synaptic) of virtualbox. | 22:19 |
balloons | well, it's just to cover all the bases | 22:19 |
balloons | I know kvm had some stuff done to it | 22:19 |
balloons | so my guess is the new package works fine | 22:19 |
phillw | balloons: it has had a lot done... mainly due to ubuntu realising that KVM is at the heart of cloud computing :D | 22:19 |
balloons | anyways, we'll know it a moment | 22:20 |
phillw | which as Canonical are a platinum partner is "slightly" embaressing :) | 22:20 |
balloons | new cd is almost synced | 22:20 |
balloons | ok, kvm'd with raring | 22:26 |
balloons | install seems fone | 22:27 |
balloons | ohh.. | 22:27 |
balloons | there it is | 22:27 |
balloons | :-p | 22:27 |
phillw | balloons: there 'what' is? | 22:28 |
balloons | the bug | 22:28 |
balloons | i'll be adding some screenshots | 22:28 |
phillw | thank heavens for that! I even re-installed my entire system as I'd been 'playing' with KVM owing to the bug I pasted up earlier. | 22:29 |
phillw | so, a click on 'affects me' and a note stating it is present in raring would really help :D | 22:29 |
balloons | already doing my friend | 22:30 |
phillw | now, do we have any idea of what to report it against? | 22:30 |
balloons | I hate to ping xnox on this, but he will likely be able to tell us more | 22:33 |
phillw | For all I know, it could be clash with gtk2 & gtk3... it's certainly like nothing I've seen before. | 22:35 |
=== fader_` is now known as fader_ | ||
phillw | balloons: ssh -X with a speed of 76Kb/s is painful... but I think a worthwhile exercise to try raring on a full centos version of KVM | 23:21 |
balloons | haha | 23:21 |
balloons | full speed ahead | 23:22 |
phillw | It's on my dedicated server.... it would be much faster for you to do it! | 23:22 |
phillw | it has 100 Mb/s backbone.... it's like wading through treacle :( | 23:23 |
phillw | balloons: you're too young to remember bulletin boards when we were running at 300 b/s | 23:27 |
balloons | 2400 baud was the first for me | 23:28 |
balloons | I used 14.4k/28.8k for awhile, then 56k for a long time | 23:29 |
phillw | ooh, we had 300 or, if lucky the 75/1200 was just being supported. But... happy days. In those days people wrote really 'tight' code. | 23:29 |
phillw | here in the lovely countryside I get not much more than 56K .... too many miles of copper to the exchange. | 23:30 |
phillw | I must have a session some time with you so you can set up stuff on the server. It's doing a really good job for teams. | 23:32 |
balloons | I too have find the niceities of a always on, high speed server | 23:35 |
balloons | I don't have fast or slow connection.. honestly 70k isn't slow.. but by today's standards it definitely is | 23:35 |
phillw | balloons: for any things using 'X' it is painful, for downloading iso's..... well it is :'( | 23:36 |
phillw | It would be faster for me to get on a couple of buses, travel to my Dad's works, download it and catch the bus back. | 23:37 |
jtaylor | hurray finally also managed to get a test done | 23:39 |
phillw | jtaylor: congratulations! You see, they are not *that* scary :) | 23:40 |
jtaylor | phillw: that is not scary? http://paste.ubuntu.com/1430814/ ;) | 23:41 |
jtaylor | (its not the one I forwarded right now, thats a lot nicer) | 23:41 |
jtaylor | bug 695881 is what I did today | 23:42 |
ubot5 | Launchpad bug 414976 in Mudlet "duplicate for #695881 STR-TAB should change Tabs" [Wishlist,Fix released] https://launchpad.net/bugs/414976 | 23:42 |
jtaylor | debian bug 695881 | 23:42 |
ubot5 | Debian bug 695881 in python-numpy "python-numpy: add autopkgtests" [Wishlist,Open] http://bugs.debian.org/695881 | 23:42 |
phillw | jtaylor: when you are writing stuff, it seems scary. I am a fiend for making 2 comment lines per line of code... but that's so when I look at the code 2 years later, it tells me what on earth was in my mind when I wrote it :P | 23:43 |
jtaylor | phillw: yeah, that particular one is still in development, I'll comment it before the upload | 23:43 |
jtaylor | the symlink crazyness is due to the package not installing the tests and python relative imports preventing just copying the tests from the source | 23:44 |
phillw | jtaylor: your older self will thank you for doing so.... on that matter, you'll just have to trust me :) | 23:44 |
phillw | I look back at some of the 'one off' sections of code I've written and really could not follow it without the comments. | 23:46 |
jtaylor | the wiki could probably use a python dep8 test example | 23:48 |
jtaylor | very many python packages can be tested the same way | 23:48 |
jtaylor | e.g.: https://github.com/neurodebian/pandas/blob/debian/debian/tests/unittests | 23:50 |
jtaylor | just I'd now add pys=${pys:-$(...)} so it can be reused in debian/rules too | 23:50 |
phillw | jtaylor: I'm not a python person, I'm PHP, but as python / perl / php can all achieve similar results when 'chatting' to MySQL it is fortunate that they all start in 'P' for LAMP. | 23:51 |
phillw | jtaylor: but, as for commenting your code?... http://forum.phillw.net/viewtopic.php?f=5&t=80 | 23:52 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!