| === EnvyOpla is now known as Pikiora | ||
| thomi | lifeless: ping? Got a minute? | 04:25 |
|---|---|---|
| lifeless | hi | 04:40 |
| lifeless | sure | 04:40 |
| thomi | lifeless: so we're using python-junitxml to get test results published in jenkins. Everything was working smoothly, and then today we start getting 0-byte result files. | 04:44 |
| thomi | it seems that if we run just part of the test suite everything works as expected | 04:44 |
| thomi | ...but if we run the whole thing we end up with an empty file. | 04:44 |
| thomi | I've tried writing a test to reproduce the problem, but haven't had any luck so far. I was wondering if you've come across this before? | 04:45 |
| lifeless | something is exiting the process without letting normal cleanup occur | 04:45 |
| lifeless | e.g. calling _exit() | 04:45 |
| lifeless | or segfault | 04:46 |
| thomi | lifeless: hmmmm, that sounds plausible | 04:46 |
| lifeless | the way the file gets written | 04:46 |
| lifeless | is that each test accumulates data in an object | 04:46 |
| lifeless | and result.stopTestRun() triggers serialization | 04:46 |
| lifeless | stopTestRun is not getting called | 04:47 |
| lifeless | -> one of the things I listed is the cause ;) | 04:47 |
| lifeless | sadly junit files cannot be written incrementally | 04:47 |
| thomi | lifeless: awesome, thanks. I should be able to track it down now | 04:47 |
| lifeless | gl | 04:50 |
| thomi | ty | 04:52 |
| thomi | lifeless: heh, I figured it out - jenkins was aborting our test runner since it had taken longer than the configured timeout, and I didn't notice the log message :) | 05:06 |
| lifeless | :) | 05:15 |
| chilts | morning | 21:41 |
| chilts | ibeardslee: yeah, lovely holiday | 21:41 |
| chilts | really didn't want to come back to the winter :) | 21:41 |
| chilts | it was a crazy 3 weeks going all around the UK to see people, but a good half a week in Paris at the end | 21:42 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!