=== swordsmanz is now known as hugbot | ||
=== hugbot is now known as swordsmanz | ||
=== swordsmanz is now known as hugbot | ||
smb | jdesfossez, Possibly those messages are not of high enough level. You could try dmesg -n <level> (level -> eg. warn|notice|info). | 08:16 |
---|---|---|
bjf | apw, the rebuild DEP-8 test. that's going to get run all the time now for all kernel packages. is that really what we want? | 16:20 |
bjf | apw, i know why that test exists, i'm just wondering why it's in the kernel. if it's to ensure that new toolchain packages properly build the kernel why isn't there a test in the toolchain packages for it rather than the kernel? | 16:21 |
apw | bjf, it is in there because it is the kernel test one wants to run when any of the packages we depend on, changes | 16:24 |
apw | bjf, unfortuanatly there is no way to know you are testing for "me" rather than a "dependancy" | 16:24 |
apw | bjf, which is what we would really want ... | 16:25 |
bjf | apw, yeah | 16:25 |
apw | bjf, i think there was a proposed way to tell, which would let one work it out | 16:25 |
bjf | apw, so it's adding multiple hrs. to the testing and when we start running these on multiple archs is going to get worse | 16:26 |
bjf | apw, i guess if it is a problem maybe someone will fix it so it only runs when we want rather than all the time | 16:28 |
apw | bjf, yeah i cannot deny it is stupid, just asking pitti ritht now if there is a way | 16:28 |
apw | bjf, is this in britney testing or in "bjf" testing btw | 16:31 |
bjf | apw, new britney testing that CI is adding | 16:32 |
apw | bjf, ok so official, not runnig somewhere where we get to help out | 16:32 |
apw | to say UNIT_TESTING_FOR_KERNEL and let that short circuit it | 16:32 |
bjf | apw, correct. | 16:32 |
apw | bjf, i am getting the "it doesn't know, and it is all very hard" but i do struggle to say why it doesn't know myself | 16:34 |
infinity | apw: DEP-8 doesn't know why a test was triggered, just that you decided to run some tests. The fix belongs one layer above that, I suspect. | 16:44 |
infinity | apw: ie: we should only run the test under certain circumstances, but the test itself can't know that, it's the framework calling it (adt-britney, whatever) that needs to be smarter. | 16:44 |
apw | infinity, well likley, though it is odd it isn't parameterised with why it is being run | 16:45 |
infinity | apw: That implies more intelligence than DEP-8 is meant to have. It should be entirely framework agnostic. | 16:45 |
infinity | apw: You can run tests by hand, you can write infra that runs all your tests in a loop, you can trigger them for $reasons, none of that matters to the tests. | 16:45 |
apw | yeah, and i'd be cool with the reason it ran being "run by hand" or "none" | 16:46 |
cyphermox | jsalisbury: hey, just curious if you had found anything about bug 1359689 | 17:01 |
ubot5 | bug 1359689 in plymouth (Ubuntu) "cryptsetup password prompt not shown" [High,Triaged] https://launchpad.net/bugs/1359689 | 17:01 |
jsalisbury | cyphermox, yes, I backported commits abd69c55d and 144ecb97c. I should post a link in the bug with a test kernel shortly. | 17:03 |
cyphermox | ok | 17:03 |
cyphermox | thanks :) | 17:03 |
jsalisbury | cyphermox, my test kernel build failed, so I just need to fix up my backports a bit. | 17:05 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!