Re: [feature suggestion] fixup patch review

SeongJae Park <[email protected]> writes:

> On Wed, 15 Apr 2026 01:11:11 +0000 Roman Gushchin <[email protected]> wrote:
>
>> SeongJae Park <[email protected]> writes:
>> 
>> > Hello,
>> >
>> >
>> > When a large patch series is posted and an issue on a single patch is found, it
>> > is useful to post a fixup patch as a reply to the review, instead of posting
>> > the entire new version of the series.  It is useful for doing quick reviews
>> > without polluting mailing list and reviewer's inbox.  But it seems sashiko is
>> > not reviewing such fixup.  Would it be possible to make sashiko understand and
>> > review such fixup patches?
>> 
>> Hi SJ,
>> 
>> sorry for the late reply!
>
> No worries, it is never urgent!
>
>> 
>> I like the idea but realistically I'm not sure I'll have time to work on
>> it short-term. These follow-up fixes are often too context-heavy, e.g.
>> are raw untested ideas or are drop-in replacements for some patch in the
>> series, so accurately handling all edge-cases might be not exactly trivial.
>
> Thank you for kindly explaining the challenges, I totally agree that.  Having a
> sort of command-like reply rule for giving the context in easier and explicit
> way, like syzkaller [1], might make it easier, though?

Yeah, something like this might work.

Currently I'm not trying too hard to find a baseline for a patch to
apply, simple because I want to avoid a situation when Sashiko is saying
that the patch is fine [against *some* baseline], but it's not against
the tree where it goes. I guess once Sashiko will start review
linux-next and Linus's tree, we can make it more flexible.

Thanks for the idea!

R