All times shown according to UTC.

Time Nick Message
02:22 CarlFK joined #miro-hackers
07:51 bendk joined #miro-hackers
07:51 uwe_mobile joined #miro-hackers
07:51 ChanServ joined #miro-hackers
07:51 zanoi joined #miro-hackers
07:51 mick_laptop joined #miro-hackers
07:51 paroneayea joined #miro-hackers
07:51 melinath joined #miro-hackers
07:51 z3p joined #miro-hackers
07:51 mattjohnson joined #miro-hackers
07:51 uwe__ joined #miro-hackers
07:51 jeremyb joined #miro-hackers
07:51 glee joined #miro-hackers
07:51 Jarrhead joined #miro-hackers
07:52 jamessan joined #miro-hackers
07:52 maggie_s joined #miro-hackers
07:54 Odysimus joined #miro-hackers
09:52 arpu joined #miro-hackers
10:55 janet joined #miro-hackers
11:16 janet left #miro-hackers
11:18 janet joined #miro-hackers
11:21 willkg joined #miro-hackers
11:27 willkg hi!
12:06 janet heywillkg - can you point me to an example of how your linux build unittests were detecting the failures - that you then would send out in the email?
12:10 willkg janet: are you asking to see the script that checks that things were ok and sends the email?  i can't tell if that's what you're looking for or not.
12:10 janet yes
12:11 willkg it's a (poorly written) shell script that runs on pcf8 and is in /home/wguaraldi/bin/nightlies_check.sh .
12:11 janet thx
12:11 willkg since it's on pcf8, it's on the same machine as the nightly builds, so it just checks for the existence of a bunch of files and whines if they're not there.
12:12 janet oh - so you were checking after you uploaded the build and results
12:12 willkg for the unit tests file, it also checks for a line that starts with "RuntimeError" or "FAILURE".
12:12 sort of.  it's "after" in the sense that the cron job that kicks off the nightlies_check.sh script is set to kick off after the nightly builds.  but that's all set manually.
12:13 so if someone sets one of the build servers later, then it's possible the nightlies check script is kicking off before hand.
12:13 janet i see
12:13 I was messing around with my future os x build box -
12:13 * willkg nods.
12:13 janet and seeing if I can ssh in to run the scripts
12:14 and then I modified the nightly on - to just run the unittests
12:14 willkg the reason i did it this way is that the script runs on a machine that's not doing builds.  so if one of the build machines falls off the network, has some other issue, or whatever, then those issues don't affect checking whether things went ok.
12:14 janet mainly because we could use jenkins to do the ssh part
12:14 that makes total sense
12:14 but I couldn't see where you were checking the failure parts
12:14 willkg i think having jenkins do all the coordination work instead of kicking things off with cron is a good idea.
12:15 janet but then jenkins could do that check and then report the pass /fail
12:15 * willkg nods.
12:15 janet and it can whine to the irc - which I love
12:15 willkg yeah--that'd be super.
12:15 plus if you ever move to continuous integration, you could have it whine just after someone checks in bad stuff.
12:15 janet and then it could build and run unittests - maybe w/o upload - and then check the results
12:15 and whine
12:16 on checkin
12:16 * willkg nods.
12:17 willkg also, it's kind of funny that all my credentials are intact.
12:17 janet oh - and I can execute the sikuli tests remotely
12:18 which means jenkins could too
12:18 i guess rob should disallow you
12:19 I can't see into your dir on pcf8
12:20 oh - but I was able to scp it
12:22 I had another question for you willkg
12:23 if I get you margarita's mailing address  - would you be able to send her the mac mini ?
12:24 willkg janet: so you have the nightly script?
12:25 janet: and, yes, if you give me margarita's mailing address, i can pack up and send her the mac mini.
12:50 CarlFK joined #miro-hackers
13:21 janet thanks willkg I'll let you know when I catch up with her
13:21 * willkg nods.
13:22 willkg janet: also, i <3 your behavior design documents.
13:23 janet ok - thank you
13:23 I hope they will make everyone's lives easier
13:24 willkg janet: also, as a side note, bugzilla is great for discussion about an issue, but it's not a good tool for maintaining a consensus.  so it's not a good substitute for a design document.
13:25 janet exactly
13:25 but if you have a good doc - you can then annotate it with specific bugs for each features so you know you've got them covered
13:25 * willkg nods.
13:25 willkg it's super for that.  bugs are good action items for discrete actions.
13:25 janet bugz gets too hard to read once we've changed our minds 6 times
13:28 willkg my boss recently created a "tracker" bug which blocks on a bunch of other bugs.  so then you see something like this: https://bugzilla.mozilla.org/s[…]ree.cgi?id=687677
13:29 so one way to do a new feature would be to work out the design document, then link it to a new tracker bug which is blocked by all the individual discrete action bugs.
13:29 then you get that nice dependency tree that's easy to look at and see where you're at.
13:30 also, someone threw together bugzillajs: https://addons.mozilla.org/en-[…]addon/bugzillajs/
13:30 that only works on the mozilla bugzilla, but it sure does make it more functional and pretty.
13:42 janet willkg:  - we do metatickets sometimes - but the problem is they end up with about 1000 parts - that really belong in different bugs
13:42 the inline images are really cool
13:43 a version supporting all Bugzilla installations will be out soon.
13:46 willkg if you look at that tracker bug (the one at the top), all the parts are defined in separate bugs.  it was created _after_ the planning work was done.
13:46 pretty sure we never did it that way on miro bugs.
13:46 at a minimum, i can't remember doing anything like that.
13:48 janet no - not really - it makes sense
13:57 maggie_s joined #miro-hackers
14:20 z3p willkg: janet: we've done that for device tickets, I think
14:21 but certainly not frequently
14:21 ajonas joined #miro-hackers
14:52 kcw joined #miro-hackers
14:55 errado joined #miro-hackers
17:00 kcw_ joined #miro-hackers
17:07 odenthaa joined #miro-hackers
17:08 errado joined #miro-hackers
17:30 Odysimus joined #miro-hackers
17:50 kcw joined #miro-hackers
18:02 glee1 joined #miro-hackers
18:20 DGMurdockIII joined #miro-hackers
18:20 DGMurdock joined #miro-hackers
19:02 janet joined #miro-hackers
19:11 maggie_s joined #miro-hackers
19:15 janet left #miro-hackers
19:25 CarlFK joined #miro-hackers
19:32 Jarrhead joined #miro-hackers
19:41 Jarrhead joined #miro-hackers
20:04 GitHub17 joined #miro-hackers
20:04 GitHub17 [miro] kazcw pushed 1 new commit to master: https://github.com/pculture/mi[…]fbe751369454c9cd7
20:04 [miro/master] Merge pull request #19 from pculture/bz18239 - kazcw
20:04 GitHub17 left #miro-hackers
20:06 kcw joined #miro-hackers
20:06 bendk willkg: you around?
20:13 okay, I'm taking a quick break, but I think I have a complete solution to the gstreamer problems.  Using the GPL version rather than the LPGLed one
20:13 I would have figured the opposite, but the GPL version has a lot more modules (especially faad/faac)
20:19 willkg bendk: wait, really?
20:19 bendk: oh...  actually, that makes sense if the module code is gpl'd.
20:20 bendk: i just picked the lgpl stuff on a whim.  i didn't have a good reason.
20:20 kcw joined #miro-hackers
22:47 maggie_s joined #miro-hackers
23:03 arpu joined #miro-hackers

← Previous day | Index | Server Index | Channel Index | Today | Next day → | Atom Feed | Search | Google Search | Plain-Text | plain, newest first