In test run, what is a fixable pass?

[badp]

Hi, I stumbled upon git-branchless as a citation for the docs of the future jj run command documentation. Unlike jj run, git-branchless test run exists here and now, and I thank you for it. I'm very deep into this commit series I'm working on and this will help me make sure each commit is as good as can be.

One thing that I don't see explained in the docs is, what's a fixable failure?

My command looks something really goofy like:

~/.cargo/bin/git-branchless test run $(jj log --no-graph -T 'commit_id ++ "|"'; echo -n HEAD) -s worktree -j0 -x './gradlew --no-daemon --no-parallel compileTestJava'

and that's all fine and dandy and works great. I get 26 passes! but some of these are "Passed (fixable)".

If I use git-branchless test show -v to inspect some of the success runs and some of the "Passed (fixable)" runs and some of the "Passed" runs I don't see any obvious difference in stdout. So I'm a little stumped.

What, then, does "fixable" mean? (--version says I'm running "git-branchless-opts 0.10.0")

[arxanas]

Hi @badp, this is specific to the fix mode. It indicates that the command left unstaged changes in the working copy. This is more useful for formatters and linters with autofixes, which are supposed to modify the working copy.

You can ignore it if it's not relevant to you. But you might want to consider either committing or ignoring whatever unstaged changes the gradlew command is causing, generally.

The results are saved/cached, so if you were to run git test fix instead of git test run with the same options, then it would update each commit with the new working copy contents without having to re-run the command.