Asbjørn Ulsberg
Asbjørn Ulsberg
> The way I see it, when there is a merge with same named branch it's reasonable to assume all commits are counted rather than only one of the branches....
> When you merge the 'release/1.2.3' branch the version of the current branch is bumped according to 1.2.3 Hm. It may be that `MergeMessageVersionStrategy` somehow deals with this, but we...
Did #2280 not fix this?
[Bug] gitversion seems to pick the wrong base version, when using git flow while tagging on develop.
Are you able to reproduce this in a test and submit it as a pull request?
Can you please try [the command line](https://gitversion.net/docs/usage/cli/) instead? If it too fails, it may be easier to figure out why.
Yeah, looks like GitVersion itself is working fine. If there's something wrong with `msbuild.yml`, that's out of scope for GitVersion, sorry.
@ucapbba, for GitVersion to execute successfully, it needs the full repository history. By default, `actions/checkout` does a shallow clone. Add the following to the workflow and it should at least...
Are you able to reproduce this in a `RepositoryFixture` test?
Interesting. But perhaps instead of replacing YamlDotNet, we can attempt the fast, naïve parser first, and if it fails, we can fallback to YamlDotNet?
> The config is YAML because it's intended for humans, which is perfectly reasonable. And requires some compatibility guarantees, this this has to stay. Agreed. > The cache however, is...