To my knowledge there is no way to get ObjectiveGit to give me an "unborn" branch, that is one with an empty SHA, so GTMergeAnalysisUnborn is never returned.

Any idea why the test fails? Locally the (iOS) tests all succeed with the newer versions of Quick and Nimble.

My guess is that we are still running tests with Xcode 8.0 on Travis until #631 is merged. There could be some changes in Xcode and/or Nimble and Quick that cause this.

Hope the tests pass now.

About "unborn" : you can get one by deleting the ref pointed to by HEAD.

Since you can only merge a branch in HEAD, I'm wondering about the use case for that method.

The reason why I needed this is for pulling a branch other than the current one. Think git pull origin master while you are on branch feature-1.

What I meant was that since the method takes branches, there is no way to pass it the nonexistent target of head, which means GTMergeAnalysisUnborn is never passed out by itself (even though the SHA property is Nullable).

Btw, feel free to sqash merge this, I realized I merged master into it instead of rebasing ontop of it.

@pks-t @ethomson Would you guys be interested in having that capability upstream ? I'm a little wary of adding API not backed by what's in The Library...

Yeah; it seems that this is similar to giving git_merge_analysis the knowledge of a different ours. I can see how that would be useful.