Re: [PATCH v2] tests: do not use "slave branch" nomenclature

From: Paolo Bonzini <pbonzini@redhat.com>
Date: Jun 19, 2020, 14:23 UTC
Message-ID: <e7611e0f-3d62-fc92-7f35-5abcc11f2fd8@redhat.com>
In-Reply-To: <20200619130058.GA5027@danh.dev>

On 19/06/20 15:00, Đoàn Trần Công Danh wrote:

Show 5 quoted lines

> I think common terminology in Git's test is this kind of branch is side.
> In this inaccurate comparison:
> 
> 	git grep -E '(branch|checkout|switch).* side '
> 	git grep -E '(branch|checkout|switch).* feature'

Side branch is the name that git uses for "parents other than the first one in a merge commit", for example

	Verify that the tip commit of the side branch being merged is
	signed with a valid key

Feature branch is what you call branches in a workflow that does feature development in a dedicated branch instead of the master branch. In addition to the two that you point out, there are other occurrences of "feature branch". For example in t5520-push.sh:

# add a feature branch, keep-merge, that is merged into master, so the # test can try preserving the merge commit (or not) with various # --rebase flags/pull.rebase settings.

and that has some resemblance with the format-patch test. (However, t5520-push.sh doesn't call its branch "feature"

So I think both terms are acceptable. Certainly "feature branch" is used a lot by git users (and was suggested in the v1 review) even though it's not as prevalent in the source code.

Paolo