daniels changed the topic of #freedesktop to: https://www.freedesktop.org infrastructure and online services || for questions about freedesktop.org projects, please see each project's contact || for discussions about specifications, please use https://gitlab.freedesktop.org/xdg or xdg@lists.freedesktop.org
pkira has quit [Ping timeout: 480 seconds]
karolherbst has joined #freedesktop
columbarius has joined #freedesktop
co1umbarius has quit [Ping timeout: 480 seconds]
<bwidawsk>
What is the magic to make a branch that passed CI not run again for the merge request?
<mupuf>
bwidawsk: git push -o ci.skip
ximion has quit [Quit: Detached from the Matrix]
cptaffe has quit [Remote host closed the connection]
cptaffe has joined #freedesktop
<bwidawsk>
mupuf: I feel like there should be a way in gitlab.yml
<mupuf>
bwidawsk: oh, of course
<mupuf>
but what you want is probably just not to run ci on branches, just in MRs
<bwidawsk>
my preference for now is all branches run CI, but if you create an MR, it won't run again
cptaffe has quit [Remote host closed the connection]
a-l-e has joined #freedesktop
a-l-e has quit [Ping timeout: 480 seconds]
pzanoni has quit [Read error: Connection reset by peer]
pzanoni` has joined #freedesktop
gchini has quit [Quit: Leaving]
gchini has joined #freedesktop
bmodem has joined #freedesktop
cptaffe has joined #freedesktop
pzanoni` has left #freedesktop [#freedesktop]
pzanoni has joined #freedesktop
cptaffe has quit [Remote host closed the connection]
swatish21 has joined #freedesktop
rsripada has quit [Read error: Connection reset by peer]
swatish2 has quit [Write error: connection closed]
rsripada has joined #freedesktop
aswar002_ has joined #freedesktop
random_james_away has joined #freedesktop
random_james has quit [Read error: Connection reset by peer]
aswar002 has quit [Write error: connection closed]
kxkamil2 has quit []
cptaffe has joined #freedesktop
cptaffe has quit [Remote host closed the connection]
cptaffe has joined #freedesktop
kxkamil has joined #freedesktop
systwi has joined #freedesktop
bmodem has quit [Ping timeout: 480 seconds]
bmodem has joined #freedesktop
systwi_ has quit [Ping timeout: 480 seconds]
bmodem has quit [Excess Flood]
bmodem has joined #freedesktop
ximion has joined #freedesktop
DavidHeidelberg has quit [Remote host closed the connection]
DavidHeidelberg has joined #freedesktop
tzimmermann has joined #freedesktop
ximion has quit [Quit: Detached from the Matrix]
pkira has joined #freedesktop
Venemo has joined #freedesktop
glehmann has joined #freedesktop
<MrCooper>
bwidawsk mupuf: it's possible in principle to set up the CI rules such that there's only a single pipeline, not separate ones for the branch & MR; though note that one consequence of this is the pipeline cannot rely on MR-related stuff, since there isn't always an MR (yet) when the pipeline runs
<mupuf>
MrCooper: oh, I'll need to dig in the doc then
mvlad has joined #freedesktop
thaller has joined #freedesktop
thaller has quit [Ping timeout: 480 seconds]
thaller has joined #freedesktop
AbleBacon has quit [Read error: Connection reset by peer]
<eric_engestrom>
bwidawsk: I don't quite understand when you want the pipeline to run or not, but you can write all the complex `workflow:rules:` that you want with `&&` and `||` (but you need to quote the entire value of the `if:` if you use the latter) and it will create a pipeline if any `if` matches and not otherwise
thaller has quit [Read error: No route to host]
<eric_engestrom>
as for your alpine issue I can't help you :(
<MrCooper>
bwidawsk mupuf: also, this means that any rules based on changed files will match only for the files changed in the last push, so in some cases the last pipeline might be green and allow an MR to be merged, even though previous pipelines failed in jobs which didn't run in the last pipeline
<emersion>
> Your comment could not be updated because your comment has been recognized as spam. please, change the content to proceed..
<emersion>
and no way to get past this?
Haaninjo has joined #freedesktop
<mupuf>
emersion: yeah, so annoying. Need to look into it
<mupuf>
eric_engestrom: yeah, but how can we tell the merge request to use this pipeline to gate merging?
<MrCooper>
if there's only one pipeline, it'll have to use that?
<eric_engestrom>
yeah gitlab will gate the merge (if configured to do that) on whatever pipeline it finds for the commit that the MR is on
<eric_engestrom>
if there's a merge pipeline it will pick this one, otherwise any other
<eric_engestrom>
well maybe not "any" other, but I'm not sure the rule for that
<eric_engestrom>
but at least a branch pipeline will count
bmodem has quit [Ping timeout: 480 seconds]
Haaninjo has quit [Quit: Ex-Chat]
<eric_engestrom>
anholt: if I have a deqp-*.toml that include a piglit run with `profile = "quick_shader"`, is there a way to override that, with something like `variables: PIGLIT_PROFILES: all` (which I _think_ only works with `piglit-runner.sh` and not with `deqp-runner.sh`) ?
<eric_engestrom>
I'm trying to have a "full" variant of the job, but I don't want to have to copy/paste all the deqp-*.toml just to change that one line
GNUmoon2 has quit [Remote host closed the connection]
<psykose>
you need RUSTFLAGS="-C target-feature=-crt-static" in env
<bwidawsk>
psykose: sweet, thanks I'll try it
<psykose>
you're using the rustup toolchains with the *-linux-musl triple so it defaults to ""static linking"" to be helpful, so that's not found because there's no libgbm.a or libinput.a
<bwidawsk>
yeah, oddly I couldn't get install-rust.sh to work
<psykose>
the equivalent rust toolchain in alpine has it patched to be the other way around and default to non-static like *-linux-gnu, but only distros do that
<bwidawsk>
so I had to use rustup, which picks musl
<psykose>
yeah that part is normal, it's just the default for the triple
<psykose>
it breaks more than that even if there was a .a
blatant has quit [Quit: WeeChat 4.0.5]
<psykose>
but yeah should work with that passed
<psykose>
can't believe i've had to see this same issue for like the past 3 years
<bwidawsk>
psykose: so the explicit static packages I picked can go away with those flags, right?
<bwidawsk>
I guess that's just libxkbcommon for now
<bwidawsk>
oh, and zlib
<psykose>
yep
<psykose>
ah and yeah the reason install-rust doesn't work is because it installs *-linux-gnu, but alpine is linux-musl; i.e. you're installing a cross compiler to -linux-gnu and then that needs a linux-gnu targetting C toolchain as well, as well as linux-gnu versions of all the dependencies you're linking, etc
<psykose>
which is not the case by what you're getting since alpine is not glibc
<bwidawsk>
psykose: thank you for the explanation
<psykose>
the fact you got a link error at all means they probably actually fixed the issue these days now that i think about it (offtopic)
<psykose>
before the latest rust version, it would silently just link the crt static, but add -Wl,-Bdynamic of any .so it needed anyway for the native libs instead of failing to find a .a
<psykose>
this gave you a broken binary that was not actually static and just segfaulted because it didn't really make sense
alanc has quit [Remote host closed the connection]
alanc has joined #freedesktop
pq has joined #freedesktop
thaller has quit [Ping timeout: 480 seconds]
jarthur has joined #freedesktop
thaller has joined #freedesktop
<anholt>
eric_engestrom: nope
<anholt>
but, also, don't use "all" ever, it's just slow for no meaninful coverage gain. Use quick_gl. And at that point I'm not sure why you'd have a separate run doing only the shader tests subset.
swatish2 has joined #freedesktop
<bwidawsk>
psykose: hmm, that didn't seem to work
<psykose>
same failure? where did you put the env?
<psykose>
i don't know how the rest of the build works but you have to check if it's set at the place `cargo` gets called in the end and not overriden with something else etc
<bwidawsk>
I can do that
<bwidawsk>
psykose: ah, I think I see the problem. I need that env for cargo invocations after initial image creation
i509vcb has quit [Read error: Connection reset by peer]
cptaffe has joined #freedesktop
i509vcb has joined #freedesktop
co1umbarius has joined #freedesktop
columbarius has quit [Ping timeout: 480 seconds]
a-l-e has joined #freedesktop
psukys has joined #freedesktop
<bwidawsk>
psykose: hmm, now it just complains about libxkbcommon? Maybe I do need xkbcommon-static for that package?
<psykose>
i don't know if you don't post the error, but as an accurate guess you probably need to add libxkbcommon-dev because you didn't and it doesn't exist, so the .so is missing
flom_84 has quit [Remote host closed the connection]
sima has quit [Ping timeout: 480 seconds]
thaller has quit [Ping timeout: 480 seconds]
todi has quit [Ping timeout: 480 seconds]
swatish2 has quit [Ping timeout: 480 seconds]
cptaffe has quit [Remote host closed the connection]
<i509vcb>
Is it known that when I log in with github sso on gitlab I need to press the remember me button on the regular email and password sign in in order to stay logged in for a longer period of time?
<i509vcb>
And the remember me on sso sign in effectively does nothing
Haaninjo has joined #freedesktop
cptaffe has joined #freedesktop
dcunit3d has quit [Remote host closed the connection]
cptaffe has quit [Remote host closed the connection]
psukys has quit [Ping timeout: 480 seconds]
cptaffe has joined #freedesktop
cptaffe has quit [Remote host closed the connection]
<glehmann>
is mesa CI working as expected? 54min to merge a MR doesn't sound right
cptaffe has joined #freedesktop
<airlied>
glehmann: was it queued or running for 54 mins?
<bwidawsk>
Tells you at the top `52 minutes 22 seconds, queued for 11 seconds`
a-l-e has quit [Ping timeout: 480 seconds]
<airlied>
but yeah if you hit the full gamut of hw devices which you did, then that seems to be about where we are at
<airlied>
once you change nir you can often hit the 1hr timeout, and yes this seems to be acceptable behaviour until we can get some priorities or someat
<glehmann>
being this close timeout shouldn't be acceptable
<daniels>
we’re trying to reduce it, but this is also the absolute peak time
mvlad has quit [Remote host closed the connection]
<airlied>
peak is pretty much median if you do a lot of core changes unfortunately