qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: How about using clang-format instead checkpatch for fixing style?


From: Yonggang Luo
Subject: Re: How about using clang-format instead checkpatch for fixing style?
Date: Fri, 9 Oct 2020 21:44:11 +0800



On Fri, Oct 9, 2020 at 9:38 PM Daniel P. Berrangé <berrange@redhat.com> wrote:
>
> On Fri, Oct 09, 2020 at 03:25:04PM +0200, Paolo Bonzini wrote:
> > On 09/10/20 15:02, 罗勇刚(Yonggang Luo) wrote:
> > > We can even using clang-format to format the whole repository.
> >
> > checkpatch does other checks than formatting.
> >
> > Reformatting the whole repository has been mentioned many times, people
> > were worried of messing up the result of "git blame".  But without doing
> > that, it's hard to get rid of checkpatch because checkpatch only looks
> > at the lines that are touched by the patch.
>
> It is a no-win situation.
>
> checkpatch.pl is code that makes people run away screaming in horror,
> because who really wants to look at Perl code that tries to parse C
> code with regexes. The fact that you can submit a patch and get
> complaints about things you didn't actually change is a poor experiance,
> especially for new contributors who will wonder what they did wrong.
>
> Certain subsystem maintainers have done bulk cleanups for pieces of
> code before, most notably culling tabs. So we have taken the pain
> a little before. I presume we'll continue to periodically clean
> code.
>
> Aside from the git blame pain, there will also be a period of time
> when cherry-picking patches back to old versions will be tediously
> conflicting, potentially forcing cherr-picking of the style cleanup
> patches too. If the cleanup patches are fine grained it might not
> be too terrible though.
>
> So we have pain with current state and we have pain with use of
> clang-format. The difference is the current pain is long term
> ongoing pain, while the clang-format pain will be an initial
> hit whose impact will slowly fade over time.
>
> Personally I think it would be worth the change in the long
> term. I should really put my money where my mouth is though and
> propose it for libvirt too.
When the repository getting bigger and bigger, a automatically formatter are more
and more needed, 
LLVM/Chrome(Blink)/Rust and may big project are already using formatter,
python also did that, qemu are getting bigger and bigger everyday.
>
> Regards,
> Daniel
> --
> |: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
> |: https://libvirt.org         -o-            https://fstop138.berrange.com :|
> |: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
>


--
         此致

罗勇刚
Yours
    sincerely,
Yonggang Luo

reply via email to

[Prev in Thread] Current Thread [Next in Thread]