@Enrico Tassi : the snap daily build failed with an error which doesn't tell me anything. It seems to be some network issue but restarting the job didn't help. Can you please have a look at (https://github.com/coq/platform/runs/3608063515?check_suite_focus=true)?
it really looks temporary
sudo snap refresh lxd
should contact the store and download the most recent version of lxd
i can download that package, so I suspect a network failure from that VM
here it is, a known problem not fixed yet: https://forum.snapcraft.io/t/unexpectedly-empty-response-from-the-server/23393/14
in the meanwhile... retry...
The snap think is fairly persistent and I would say as is it fails more frequently than it succeeds. Should we just put in a retry loop as suggested in the post you shared?
I'm fine with that, but I'm not sure how to do it (I guess we should fork the action, but I did not try before). Also "refresh lxd" is really optional, IMO, we could just add ... || true
So essentially LXD (a new kind of container VM as far as I understand) is already there and the command just updates it to its current version?
@Enrico Tassi : I guess the command in question is part of "snapcore/action-build@v1". What would be the easiest way to change it?
I guess the line in question is this one: (https://github.com/snapcore/action-build/blob/61deecb7b83eb68caaa8b337a854c5255ba05676/src/tools.ts#L52)
So I clone this repo and redirect to the clone? Or create an issue there?
FYI: (https://github.com/snapcore/action-build/issues/23)
My understanding is that you clone the repo, you make a commit, you make a tag > 1.0.8, you change the platform CI script to point to your fork.
You don't even need a tag. You can point to a specific commit or the name of a branch.
They fixed it upstream and it seems to be working.
Last updated: Jun 03 2023 at 05:01 UTC