Make use of ccache optional

Fix for ACTS-446

Hey there! I actually had an MR in flight at !410 (merged)

Thanks, I should have checked. It meets my needs.

Can I use this MR to debug some of my "new dev" issues? What does this mean?

 This job is stuck, because you don't have any active runners online with any of these tags assigned to them: acts cvmfs

?

Hi @smh, we were granted our own gitlab CI runners for this project, which have the additional "acts" tag. Since this is your repository, you don't have access to these runners. You could remove the requirement of these tags from the CI scripts, such that it runs again on the default runners provided by CERN.

@smh @rlangenb You should be able to push the feature branch to our origin repo, and then our private runners will execute the CI jobs.

Hi @pagessin

OK, this is somewhat surprising, at least to me. I can't see any difference between my MR:

Request to merge smh:disable_ccache into master

and

Request to merge pagessin:ccache-cmake-option into master 

? Or am I still missing something. I did

git push --set-upstream upstream disable_ccache

and now I can see the branch in the list of acts-core branches (can of course be deleted).

Probably this should be clarified on the contribution guide.

Hey @smh,

You're right! I have the acts runners enabled for my fork, for exactly this reason. I'm not sure, but you should be able to do that as well under Setting > CI/CD of your repo.

I will update the contribution guide to include this.

OK, it's not clear to me how I enable the ACTS runners: all I see is

I see. I'm not sure exactly how Gitlab decides who's allowed to enable the specific runners we use. We didn't really consider this use-case when setting things up as they are right now, so sorry for the confusion.

I'm working on an MR that will allow the jobs to run on the shared runners again, and still use all cores if they run on the ACTS specific runners.

mentioned in merge request !422 (merged)

closed

reopened

@smh to see if the runners work now, you'd have to rebase your branch on top of origin/master.

added 1 commit

• 032fe112 - Test commit

Compare with previous version

added 9 commits

• 87d11d59 - PlaneSurface::clone() returned a copy of the object without respect of the shift.
• a3087b57 - attempt to make setup scripts more robust
• 004c862e - reword getting started to remove unclear instructions
• 347293e6 - add quick start snippet to compile acts
• 41120f7c - make code snippets be in line with recommendation
• 00828d9b - Perform the resolve() check also for sLayer. This is an additional check if…
• 6082db96 - Make use of ccache optional
• c326533d - Test commit
• 77951c71 - Merge branch 'disable_ccache' of https://gitlab.cern.ch:8443/smh/acts-core into disable_ccache

Compare with previous version

Hmmm, rebased but it still gets stuck.

Did you pull from origin/master before rebasing?

origin/master should be at 477b30e5

added 12 commits

• d61648cf - PlaneSurface::clone() returned a copy of the object without respect of the shift.
• e5c558da - attempt to make setup scripts more robust
• 1c7a2a6e - reword getting started to remove unclear instructions
• c8bd1203 - add quick start snippet to compile acts
• 9d13f57e - make code snippets be in line with recommendation
• 2a5aa22c - Perform the resolve() check also for sLayer. This is an additional check if…
• d902a27b - Make use of ccache optional
• a1a44ec3 - Test commit
• 25ee953c - reword getting started to remove unclear instructions
• 329a9c01 - add quick start snippet to compile acts
• 5ec2329f - make code snippets be in line with recommendation
• 8e24fa3f - Merge branch 'disable_ccache' of https://gitlab.cern.ch:8443/smh/acts-core into disable_ccache

Compare with previous version

added 6 commits

• 8e24fa3f...477b30e5 - 5 commits from branch acts:master
• 7dd5960a - Fix merge conflict

Compare with previous version

OK, turns out I'm not great at rebasing (but of course it doesn't matter, since this branch won't be committed). Looks like it's working now, can close once the pipeline finishes.

For the future, I guess access to the ACTS runners is done based on username?

closed