spack
plumed 2.9.1
On a second look, there's one issue in pipelines: https://gitlab.spack.io/spack/spack/-/jobs/12339802
Unlikely related to this MR. Is that configuration working reliably for other PRs?
Let's try to re-run, but in general yes - that configuration was building reliably before.
![spackbot-app[bot] avatar](https://avatars.githubusercontent.com/in/123749?v=4 "Published by a pub.dev verified publisher"){kind=small}
configure: error: C++ compiler cannot create executables
seems unrelated, but we still need to fix this!
@alalazo https://gitlab.spack.io/spack/spack/-/jobs/12342276#L112 looks like some preliminary download failed, so I guess that's why the compiler wasn't functional. Any ideas?
@mabraham That seems to be retried and ok 2 lines below. The real error seems https://gitlab.spack.io/spack/spack/-/jobs/12342276#L486
@mabraham That seems to be retried and ok 2 lines below. The real error seems https://gitlab.spack.io/spack/spack/-/jobs/12342276#L486
Indeed, but that makes no sense if all the change is doing is upgrading a point version of plumed. I'm guessing the failing pipeline is broken and this is just the first (gromacs?) PR to show it. I've opened https://github.com/spack/spack/pull/46050 to try to test that idea.
So #46050 has net no changes, and passes all CI, so maybe plumed 2.9.1 somehow has broken exactly this configuration. Weird. I don't see anything in their release notes https://www.plumed.org/doc-v2.9/user-doc/html/_c_h_a_n_g_e_s-2-9.html (or for the 2.8.x series)
@GiovanniBussi any suggestions?
Hi, where can I see the failure? Is this it this? It would be quite puzzling. Can you confirm that 2.9.0 can build and 2.9.1 cannot build with everything else being the same?
@eugeneswalker Do you know if there was any change to those runners recently? I have seen failures on cray-rhel in other PRs too, which seem unrelated to the PR itself.
@GiovanniBussi any suggestions?
Hi, where can I see the failure? Is this it this? It would be quite puzzling. Can you confirm that 2.9.0 can build and 2.9.1 cannot build with everything else being the same?
https://gitlab.spack.io/spack/spack/-/jobs/12342276#L486 has the full details. I tried at https://github.com/spack/spack/pull/46050 to get it to rebuild 2.9.0 but even though that CI succeeded, it didn't build any new artefacts. So my guess is still that that image is broken somehow.
I've now made an unrelated real PR to the GROMACS recipe (https://github.com/spack/spack/pull/46065), which might uncover whether that build image is working for plumed 2.9.0
I've now made an unrelated real PR to the GROMACS recipe (#46065), which might uncover whether that build image is working for plumed 2.9.0
Nope, that didn't rebuild the failing configuration either.
GROMACS 2024.4 is expected to release later this week, at which point we can make a spack PR that will definitely trigger all configurations, and thereby learn something here.
GROMACS 2024.4 is expected to release later this week, at which point we can make a spack PR that will definitely trigger all configurations, and thereby learn something here.
Let's see what https://github.com/spack/spack/pull/46117 teaches us
Indirectly we have learned that the rhel-cray configuration is indeed broken (#46111) so if we update this PR then CI should pass because failure there is tolerated.
configure: error: C++ compiler cannot create executablesseems unrelated, but we still need to fix this!
Hello, I am trying to build [email protected] +cuda + plumed with openmpi and [email protected]. But the configuration step of plumed 2.9.2 got configure: error: C++ compiler cannot create executables. I saw your post and wonder how to fix this issue. Thanks!
