FR: Print a hint of how to resolve conflicted branches or conflicted change ids

[ilyagr]

When a user has a conflicted branch, they usually run into a message like the following:

Error: Revset "deneb-produce-blockv3" resolved to more than one revision
Hint: The revset "deneb-produce-blockv3" resolved to these revisions:
179b3c05123f practical review :)
c79ed56ed347 config/params: Require placeholderFields to be sorted and distinct (#12834)

When a user has a conflicted change id, they run into a very similar message.

It would be nice if we did one or both of the following:

1. jj informed the user when such conflicts are created (this could also apply to file conflicts)
2. jj could automatically determine that it's in one of these situations when resolving a revset.

In either case, jj could then the appropriate advice (jj branch set for conflicted branches, jj abandon for conflicted change ids).

Related Discord discussions: 1, 2.

[james-prysm]

having some of these tips would be great!

[wmrmrx]

Hi, could this issue be assigned to @isinyaaa and me (@wmrmrx)? We'll try looking at it!

[martinvonz]

I could only assign it to you, @wmrmrx, probably because @isinyaaa hasn't said anything on this issue yet. @isinyaaa, can you add a comment so we can see if GitHub lets me assign it to you too after that?

[isinyaaa]

Here you go :) Thanks!

[martinvonz]

It worked :)

[wmrmrx]

Hi, checking if the revset is a branch would be the way to go here?
https://github.com/martinvonz/jj/blob/e3c85d6ecc20ee4240cdf3ed80107ce7dc268abe/cli/src/cli_util.rs#L1018C4-L1018C4

if revset_is_branch {
    hint = ... // hint to use jj branch set/jj abandon
} else {
    hint = normal_hint
}

[martinvonz]

Yes, that sounds right to me. You can inspect the revset_expression to see if it mentions any conflicted branches. It's probably good enough to check only if the expression is exactly RevsetExpression::CommitRef(RevsetCommitRef::Symbol(name)), where name is a conflicted branch.

[PhilipMetzger]

I'm going to close this issue as it has been adressed by #2244