Yuya Nishihara

@ilyagr any thoughts on whether `annotate`/`blame`/`credit` should be top-level or subcommand of `file`? I agree with the following rule, but `annotate` is file oriented, and can be considered a single...

It might be simpler to just rely on patch id (or change id if the backend or forge natively supports it.) Computing hash of diffs wouldn't be expensive compared to...

Looks like a bug. `visit_directory()` first tests the directory "test/", and it doesn't match the negative pattern _literally_, the whole directory is ignored. #2499 might fix the issue, but I...

> Maybe a set of functions on `String`? (the naming is discussable) Perhaps, it will be a global function similar to `fill(width, content)`. `author` isn't of `String` type for example,...

Incidentally I've implemented a variant of 2. Shell completion is another story. iirc, clap doesn't support custom completion logic yet. ```sh % jj log -Tbuiltin Error: Failed to parse template:...

> Is it a valuable addition if `jj` would list templates in the error? > > ```shell > % jj log -T > error: a value is required for '--template...

> 5. `echo "testfile" > .gitignore` > 6. `jj st` After the step 5, you'll need to run `jj untrack testfile`. Perhaps, `jj status` can show hint for the added...

> I think we should clearly label one of these orderings as "reverse topological order" to avoid confusion. +1 In DVCS context, I expect (unqualified) "topological order" is reversed, but...

> So just to be clear, you're saying that "topological order" should put the children before the parents, where the graph has edges from the children to the parents. My...

Perhaps, things are confusing because roots and heads are defined differently?