Re: [PATCH 4/5] format-patch: do not respect diff.noprefix

From: Jeff King <peff@peff.net>
Date: Mar 10, 2023, 09:49 UTC
Message-ID: <ZAr9PbevIMASG5b+@coredump.intra.peff.net>
In-Reply-To: <xmqqy1o5op1i.fsf@gitster.g>

On Thu, Mar 09, 2023 at 08:41:29AM -0800, Junio C Hamano wrote:

Show 10 quoted lines

> Jeff King <peff@peff.net> writes:
> 
> >   - if a project really does have a workflow that likes prefix-less
> >     patches, and the receiver is prepared to use "-p0", then the sender
> >     now has to manually say "--no-prefix" for each format-patch
> >     invocation. That doesn't seem _too_ terrible given that the receiver
> >     has to manually say "-p0" for each git-am invocation.
> 
> It does seem very terrible if any existing projects do use the
> workflow, as their receivers need to change their workflow, though.

I think the escape hatch there is patch 5, where the sender just sets the new variable to say "no, really, I actually want to send patches without a prefix".

I had originally thought to squash them together to help explain that better, but I wasn't 100% sure we'd want format.noprefix.

> But we can declare that we do not care about such projects that do
> not honor our -p1 worldview, and I have no objection to this change
> if we can have list consensus for us to go in that direction.

Yeah, I would very much like to hear from others on the list, especially anybody who does have a "-p0" workflow.

> Colored patches, by the way, cannot be applied, so perhaps we should
> disable ui_config altogether, on the other hand?  I dunno.

Yeah, color is a bit weird there. We auto-disable it when the patch isn't going to stdout or a pager, so it's mostly a non-issue. I think more interesting cases are ones like diff.algorithm, diff.context, etc, where they don't break the diff, but we don't quite consider them vanilla enough for plumbing.

I do wonder about diff.relative, which may or may not cause confusion on the receiving end, depending on what you're trying to achieve (are you sending a patch for somebody else's git repo, or did you make a git repo yourself and want to send a diff of some subset).

Also diff.submodule, but the implications of submodule-via-format-patch are too scary for me to even contemplate. ;)

-Peff