This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: Code formatting [Re: [patch] s390*: watchpoints regression [repost]]
- From: Mark Kettenis <mark dot kettenis at xs4all dot nl>
- To: jan dot kratochvil at redhat dot com
- Cc: brobecker at adacore dot com, eliz at gnu dot org, gdb-patches at sourceware dot org, pedro at codesourcery dot com
- Date: Sun, 18 Dec 2011 14:52:09 +0100 (CET)
- Subject: Re: Code formatting [Re: [patch] s390*: watchpoints regression [repost]]
- References: <20111218115931.GA22952@host2.jankratochvil.net>
> Date: Sun, 18 Dec 2011 12:59:31 +0100
> From: Jan Kratochvil <jan.kratochvil@redhat.com>
>
> On Sun, 18 Dec 2011 11:46:34 +0100, Joel Brobecker wrote:
> > Regarding the extra curly braces, I think it's OK to leave them out,
> > like so:
> >
>
> On Sun, 18 Dec 2011 12:37:59 +0100, Eli Zaretskii wrote:
> > > It then requires new { brackets }
> >
> > I don't think you need braces. The compiler certainly doesn't.
>
> While this is bikeshedding in its purest form I can jump in.
>
>
> > if ([...])
> > /* This is a comment that ... */
> > return;
>
> This is a bug from the first sight as there are two C statements attached to
> an `if' conditional. Two statements always need a block. This is a bug.
>
> I really do not have time to interrupt myself each time, several times
> a minute, looking at the code starting examining what those two statements
> semantically are, and therefore if they really require a block or not.
I agree with Jan here.