Re: b4 review batching operations

Mark Brown <[email protected]> Tue, 30 Jun 2026 14:05:33 +0100

Newsgroups	org.kernel.linux.tools
Message-ID	<[email protected]>

On Mon, Jun 29, 2026 at 03:01:20PM -0400, Konstantin Ryabitsev wrote:

> Yes, I should definitely think about features that would aim at high-rate
> reviewers like yourself. The current TUI model works well for a R: reviewer,
> but for an M: maintainer of a busy subsystem (or several), a lot of design
> choices would be actively slowing them down.

FWIW for me it's usually more just the general lag in the UI than the
batching thing, for my flow more usually I'm reviewing and applying as I
go.  It's if I get a large number of patches that I've already reviewed
but not applied for some reason like the merge window that it builds up.

> What I'm thinking:

> - tag multiple series at once
> - create a playbook for the order in which to apply them
> - run the playbook as a single take+thank+archive routine

That sounds reasonable, yes - TBH it'd be kind of nice to be able to tag
things permanently as a note about their state and then be able to say
"For all things with tag X do Y", the tags could also be useful for
sorting in process stuff as well as more structured notes.  My custom
scripting lets me say "This is good for applying to branch X" and then
actually do that later which was working well.

> There will be caveats and corner-cases to deal with, such as backing out if
> a series in the middle of the playbook causes too much grief, but nothing
> seems insurmountable.

I think I'd be happy with just skipping anything that causes any kind of
issue then going back and dealing with it separately (that's how my
custom stuff works).

signature.asc (application/pgp-signature, 488 B)