[mCOLlSVIxp6c]

I predict that under the proposed minimum version selection system some package will decide it's important and flawless enough that users should always depend on the most recent version. The package will recommend that users depend on version v1.0.0 but the first release will be v1.100.0. Subsequent releases will decrement the minor version to make sure newer releases are selected.

[overcyn]

User's wouldn't be able to depend on version v1.0.0 if it doesn't exist. From what I understand the tool isn't downloading the list of all versions from GitHub and selecting the minimum. Its searching for the maximum version number (i.e. minimal required version), from your go.mod file and all your dependencies' go.mod file.

[mCOLlSVIxp6c]

I think you're correct. So my nasty little hack wouldn't work. That's probably a good thing.

[readams]

It certainly complicates the picture. I think the only way to make this workable is if you adopt a convention that you specify your minimum dependency if you're a library, but the current version if you're an executable. Of course, this could just be codified in the tool, which would be a big improvement. It also makes testing a lot harder though since your library ends up tested against a different version than what your users are using.

Ugh.

[skybrian]

Interesting loophole, but it requires the community to go along with it. If it turns out people like the minimum version policy and the package is actually important, I would expect someone to fork the repo and do things the normal way?