Re: [PATCH] docs: submitting-patches: Require base tree information

[Laurent Pinchart]

On Mon, Oct 26, 2026 at 08:51:35PM +0000, Mark Brown wrote:
> On Mon, Oct 26, 2026 at 10:19:05PM +0200, Laurent Pinchart wrote:
> > On Mon, Oct 26, 2026 at 04:37:28PM +0100, Geert Uytterhoeven wrote:
> > > On Mon, 26 Oct 2026 at 15:53, Laurent Pinchart wrote:
> > > > 
> > > > Hence the prerequisite-patch-id and base tree information. Or do you
> > > > mean your base is not public at all, neither in the form of a branch
> > > > pushed to a public tree, nor as prerequisite patches sent to a public
> > > > list ?
> > > 
> > > Correct.
> > 
> > How am I supposed to review or apply something that depends on work that
> > is not public ?
> 
> It depends a lot on what the base commit is doing, if it's directly
> relevant to the patch it's an issue but if it's unrelated work for
> platform enablement that's needed for testing the changes then it
> isn't so relevant.

Indeed. I was thinking about the case where the dependencies are
relevant, but there can of course be patches at the bottom of the branch
that are entirely unrelated.

> This isn't insurmountable, you can cherry pick the
> changes over to a clean branch or something, but there's new failure
> modes introduced by being explicit about the commit (the number of times
> I've had to unwind a series apply in b4 because it paid attention to an
> unsuitable sender specified base...).

I've seen that too. So far it has bothered me less than not having base
information. With an unsuitable base I can apply the patches ignoring
the base, while with missing base I have to ask the submitter. The
latter interrupts review. Your mileage may vary though, depending on how
often you hit either case.

-- 
Regards,

Laurent Pinchart