Ability to write more detailed review comments

[hazbo]

When writing a comment on a review, you're able to pass -m followed by the message, as you'd expect. The <file> and <line> at this point are both optional however, so something like the following would be fine:

git appraise comment -m "Nice code!" 2c9bff89f0f8

Would it be unreasonable to suggest allowing the commenter to leave out the -m, much like how you'd write a proper commit message, by opening the text editor associated with git, to then write a more detailed comment?

[hazbo]

I noticed that the message is not actually required for a comment, is this the case due to there being both -lgtm and -nmw available instead of a message? If so, should it be the case that if either of those flags haven't been set, that a message is required?

[ojarjur]

That's a really good point.

Actually, I might state it more strongly than you have; posting a -nmw comment with no message is not helpful, and an -lgtm comment with no message can be posted with the "accept" subcommand, so I'm fine with the first proposal (if -m is missing, then open an editor).

I think the only reason we aren't already doing that is just that it didn't occur to me.

[hazbo]

That's weird you mention that, just earlier on I was thinking -nmw on it's own is actually really unhelpful. Maybe even a little rude!

So to be clear, are you proposing that the -lgtm (on it's own) is only going to be used with the accept subcommand? Or will it also still remain available for a comment too?

[ojarjur]

So to be clear, are you proposing that the -lgtm (on it's own) is only going to be used with the accept subcommand? Or will it also still remain available for a comment too?

Yes, that's what I'm thinking. The accept command is just a shorthand for "comment -lgtm", so I'm fine with that being the only command that allows an empty comment.

[hazbo]

Yeah that seems to make the most sense to me.