templates: Expose copy/move path formatting and status char

[algmyr]

This allows replicating the default diff summary display in templates.

This will go well with #8183 for custom log templates. E.g. with

'diff_listing(diff)' = '''
  diff.files().map(|f|
    label(f.status(),
      separate(' ',
        f.status_char(),
        f.formatted_path(),
        if(f.source().conflict() && !f.target().conflict(),
          label('conflict_description', '(resolved conflict)'),
        )
      )
    )
  ).join("\n")
'''

'conflict_listing(commit)' = '''
  commit.conflicted_files().map(|f|
    separate(" ",
      label("conflict", f.path().display()),
      label('conflict_description', f.conflict_sides() ++ "-sided conflict"),
    )
  ).join("\n")
'''

you can create a log template looking something like

Checklist

If applicable:

• I have updated CHANGELOG.md
• I have updated the documentation (README.md, docs/, demos/)
• I have updated the config schema (cli/src/config-schema.json)
• I have added/updated tests to cover my changes

[yuja]

Thanks.

[yuja]

nit: since this only matters for copied entries, I think the function name should include "copy", e.g. display_copied_path(). Maybe we could add copied_path() -> CopiedTreeDiffEntryPath function, but I don't think we'll add functions other than copied_path().display().

Another option is to add path.display_copy_from(source.path()) function to RepoPath template type.

[algmyr]

In my head this is not a copy specific function. It only makes a difference for copies, but it has the behavior you would want in general. Still open yo name changed, but I would prefer something that doesn't suggest it's copy specific.

[yuja]

I mean path().display() vs formatted_path() matters only when the entry was copied/renamed. It may be confusing if we have two ways of formatting paths.

[algmyr]

Ideally path().display() should take the move into account so that we only have one way of doing formatting. I guess that would involve making path() return some new type.

It would be a mildly breaking change, but code relying on the old behavior can easily migrate to use (the much more explicit) target().

[yuja]

It's probably better to rebase this on top of #8123.

[Velociraptor115]

The relevant code is now in main

[yuja]

nit: can you instead update existing tests to minimize test execution time? I think there are tests for .status() in test_commit_template.rs or test_diff_command.rs.