07:38 MrCooper: mareko robclark: FWIW, the job output showed that neither -Wno-inconsistent-missing-override nor -Wno-error=inconsistent-missing-override was passed on the compiler command line
07:39 MrCooper: (the latter would be preferred in CI)
16:31 soreau: a question about shader cache: if a program compiles the same shader multiple times, will it get a cached version or blindly compile a new program each time?
16:34 mattst88: it'll get a cached version. that's what the cache does :)
16:35 soreau: ok, thanks
17:47 karolherbst: anyone want to ack/rb freedreno traces change? https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30574/diffs?commit_id=13216ddc09446fa6177f6ced9038e617d8de2343 and https://mesa.pages.freedesktop.org/-/mesa/-/jobs/62159627/artifacts/results/summary/results/trace@freedreno-a618@unvanquished@unvanquished-lowest.trace.html
17:48 karolherbst: also the visual differ seems not functional
18:43 daniels: karolherbst: it's a misrender
18:43 daniels: see the overlay text on the bottom left
18:43 daniels: 'press t to chat' has lost a letter
18:43 karolherbst: ohh...
18:43 karolherbst: but I don't see the difference thing..
18:43 karolherbst: not sure what's up, but the JS code doesn't work
18:43 daniels: yeah, fixing that
18:46 daniels: https://gitlab.freedesktop.org/mesa/piglit/-/merge_requests/940
18:48 karolherbst: yeah, that fixes it, thanks :)
18:49 daniels: np
18:50 daniels: anyway, if unvanquished is unreliable on a618 - i.e. it's definitely not your printf causing it - then we should just skip it
18:50 karolherbst: yeah.. I've seen that error a long time ago as well, I just forgot about it
18:50 karolherbst: but it was also the text being wrong
19:55 robclark: daniels: I did look at the ci stats graphana but couldn't other instances of a618-traces (or a630-traces) flaking.. but that was defn some sort of flake.
20:16 karolherbst: robclark: I'm sure that I've seen that exact flake a long time ago
20:20 robclark: oh, I guess I wasn't looking that far back.. https://ci-stats-grafana.freedesktop.org/d/Ae_TLIwVk/mesa-ci-quality-false-positives?orgId=1&from=now-30d&to=now
20:25 karolherbst: if I'd had to guess, it happened to me like 1-2 years ago :D
20:50 sergi: Robclark: following the panel you linked, you may know (bitaybe useful for others here), in flakes tab, there is a ''Flake unit tests - counter'' panel to filter by the tests one would be looking for. I think with the ''interval'' more info about frequency of the flake could be found.
20:52 robclark: yeah.. saw that.. although it does not show which hardware/runners the test is flakey on
20:58 sergi: Agh! We need to link this info as well