Cian Johnston

What should envbuilder do in this case if the repository contains uncommitted changes? Should it still fetch changes from the remote?

Alternative possibility is add -> stash -> pull -> unstash. This does run the risk of users being left to clean up a nasty merge conflict though.

Parking this for now due to some open questions: - Do we `fetch` and `checkout` versus `pull --ff-only`? The former seems safer to me. Or do we `stash` -> `pull`...

@mafredri One other scenario I could imagine is that a different git repo is cloned to the path into which we wish to clone. Do you think it's sufficient to...

> Hi, is this issue still open? I would like to give it a try. Yes, this issue has not yet been worked on!

https://github.com/coder/terraform-provider-coder/pull/287 updates the provider to add the field. A separate PR will be required in `coder/coder` to set `CODER_WORKSPACE_OWNER_LOGIN_TYPE`.

@dannykopping @matifali looks like it wasn't. I can add one quickly. Edit: https://github.com/coder/coder/pull/15499 Edit edit: the above PR has been merged to coder/coder#main, we can release the updated provider version...

Hi @aarosil , can you provide the versions of Coder and the Terraform provider for which you are seeing this behaviour?

Potentially out of scope of this issue, but also having some more ability to introspect the various stages of the terraform provider would be useful here.

I propose that we expose `coderd_provisioner_job_timings` aggregated by the following: - template_name - template_version - job_type (template_version_import, template_version_dry_run, workspace_build) - provisioner_type (terraform, echo) - stage (init, plan, graph, apply) (ref:...