Archive for March, 2007
a correction to CLOBBERin’ time March 28th, 2007
the correct path for the clobber files is mozilla/tools/tinderbox-configs/firefox/$platform/CLOBBER. Not mozilla/testing/… as I posted earlier.
Watch this space for further updates!
Posted in Build | Comments (0)
It’s CLOBBERin’ time! March 28th, 2007
That’s right, gentle reader, the unit test boxes (qm-rhel02, qm-winxp01 and qm-xserve01) are now supporting Tinderbox’ clobber support. Updating the contents of (or just touching) mozilla/tools/tinderbox-configs/firefox/$platform/CLOBBER (where $platform is in [macosx, linux, win32]) will cause mozilla/objdir to be deleted. Ideally, you should include a short note in the CLOBBER file about why you’re clobbering. This will show up in the unit test boxes’ logs. Also, your checkin comment to CVS should clearly-state that you’re doing a clobber (and ideally your bug number). As it states in the instructions, try to coordinate in irc:#developers so as not to stack multiple clobbers.
But wait, there’s more! The Windows unit test box, qm-winxp01 gets special treatment. A clobber request will cause that machine to kill a running firefox process. So if, say, a mochitest was frozen from a previous run, clobbering the box should kill that process and hopefully builds and unit tests can proceed as usual.
Sorry for the burning boxes these past couple of days. In the future I’ll go back to my previous development plan of working locally and rolling changes out to the farm when fully-tested on my dev boxes. Who knew killing a process on Windows (or, heck, just deleting a directory with Python) could be such a pain?
Posted in Build | Comments (3)
qm-winxp01 going down for upgrades and maintenance March 27th, 2007
I’m taking qm-winxp01 offline for some TLC. On the menu:
- mozillabuild
- buildbot, twisted, python upgrades
Hopefully we’ll have it back up in a couple of hours.
Posted in Build | Comments (0)
(left) Justified and Ancient March 26th, 2007
(with apologies to the KLF…)
I saw this note on Schrep’s blog this afternoon and remembered, that last week was actually kind of a milestone for me. With dbaron’s help. I made my way onto a 4-digit bug – the venerable #3247 - counters in css-generated content.
David Baron wrote the original tests (which you can see at http://dbaron.org/css2.1/tests/imported-20050702/ if you’re so-inclined). I wrote a little python to convert the tests into something that could be run with reftest a few months ago. I was looking for something to use with it and they seemed like a worthy candidate.
Fast-forward to last week, and TimR asked me about those tests I’d converted. So, I dug ‘em out, popped them on my people account and forwarded a note to David. After some cleanup (there were a couple of errors in there that my conversion conveniently included for us), they are now living happily at in http://lxr.mozilla.org/seamonkey/source/layout/reftests/counters/. You can see their results running on every checkin on the Firefox Tinderbox page.
PS, kudos to Kray2, Nickolay, Callek and others for the “Creating Reftest-based unit tests” page on MDC.
Technorati Tags: Code, Firefox, Quality, Testing
Posted in Code, Firefox, Quality, Testing | Comments (2)
What’s wrong with this picture? March 21st, 2007
Yesterday, while doing some updates to the machines in the test farm, I noticed something a little funny on qm-winxp01:
The pagefile allocation had ballooned to 1.27GB, while the physical memory on the machine only had 574MB out of 2GB left to it. That was some serious fragmentation. I took a snapshot of the process list (of all users) while I was in there and saw nothing using more than a couple dozen MB of memory. This is after an uptime of at least a couple weeks, so we’ll need to keep an eye on that machine, rebooting it as necessary.
In other related news, in an attempt to track down a spurious bug on Mac OS X, under the guidance of Robert Sayre, I’ve set qm-xserve01 to debug mode and am now cycling builds every 20 minutes. This is generating a lot of console chatter, but hopefully we’ll see the error condition again and that can provide some ammunition to get rid of the problem. Reftest numbers have returned to normal after some initial confusion by the summary generator.
Posted in Build, Testing | Comments (1)
Buildbot Testfarm going down for updates March 20th, 2007
Sorry for the short notice. I’m taking the ‘bots offline for a few minutes to make a change to the reftest summary reported on Tinderbox — now it will show number of tests passing / tests failed / tests known to fail, similar to mochitests’ numbers. Tests known to fail have been removed from the “tests passing” column since they’re more akin to a todo.
Additionally, I’m applying some updates to the machines for better performance which will take a few moments. If all goes well, the machinery should be back up before you know it.
Posted in Build, Quality, Testing | Comments (0)