We have deleted the branch that hosted the v5 version of our drivers and we don't intend to push a new version in the v5 path, so retracting isn't feasible.

Why not push a new version in order to publish the retract directive and deprecation notice?

Thanks for the quick reply, @bcmills . As mentioned, the branch from which we released this v5 version doesn't exist anymore and we aren't planning on releasing any more versions in the v5 path.

Here's the go.mod for our v4 path: https://github.com/fauna/faunadb-go/blob/v4/go.mod - my understanding is that if I updated the file temporarily to point to the v5 path and published a retraction of 5.0.0-beta, that would necessarily push a new v5 version, correct? And my guess is that I can't simply add the retract directive to the current go.mod (that points at v4) to retract a v5 version.

Appreciate any help you can provide; if I've made a wrong assumption (I'm not a golang expert), let me know. Thanks!

As mentioned, the branch from which we released this v5 version doesn't exist anymore and we aren't planning on releasing any more versions in the v5 path.

You don't need to recreate the whole branch to publish a v5 version. (All it takes is a single commit with a version tag, and if it has an explicit tag it doesn't even need to be reachable from any branch that remains.)

my understanding is that if I updated the file temporarily to point to the v5 path and published a retraction of 5.0.0-beta, that would necessarily push a new v5 version, correct?

Yes, you would need to publish a new v5 version (like v5.0.0-deprecated) and have its retract directive include that version as well (so that v5.0.0-deprecated retracts itself along with all previous versions of its path).

my understanding is that if I updated the file temporarily to point to the v5 path and published a retraction of 5.0.0-beta, that would necessarily push a new v5 version, correct?

Yes, you would need to publish a new v5 version (like v5.0.0-deprecated) and have its retract directive include that version as well (so that v5.0.0-deprecated retracts itself along with all previous versions of its path).

Ah, I think this was my misunderstanding; I presumed that I couldn't retract the version I was also publishing. I'll go that route. Thanks!