Re: [PATCH v3 3/5] rust: io: use the `bitfield!` macro in `register!`

[Yury Norov]

On Fri, May 01, 2026 at 11:55:03PM +0200, Danilo Krummrich wrote:
> On Fri May 1, 2026 at 10:41 PM CEST, Yury Norov wrote:
> > On Fri, May 01, 2026 at 08:19:31PM +0200, Danilo Krummrich wrote:
> >> On Fri May 1, 2026 at 7:47 PM CEST, Yury Norov wrote:
> >> > If it comes to another round, maybe split switching to a bitfields and
> >> > getting rid of the bitfield_core? For maintainability reasons.
> >> 
> >> I'm happy to take it as a single patch -- the deleted rules are dead code the
> >> moment the switch happens, and the code is local to the same macro, so the
> >> intermediate state doesn't seem to add any value, or structure in terms of
> >> touching different components etc.
> >> 
> >> Did you have anything specific in mind when you mention maintainability reasons?
> >
> > I only mean that if there will be weird regression found in the new
> > implementation, it's better to be able to revert just a single patch
> > to restore the original code with no side effects.
> 
> Indeed, but by splitting it up that wouldn't be possible anymore, no?
> 
> Am I missing something?

The thing is that you most likely will not be able to revert the patch
cleanly - because the existing code evolves. The non-existing code
doesn't evolve. So unless it'is moved to another file, or something
like that, the revert is simpler.