Re: [Outreachy kernel] [PATCH v2] staging: media: atomisp: pci: Change line break to avoid an open parenthesis at the end of the line

[Dan Carpenter <dan.carpenter@oracle.com>]

On Fri, Apr 16, 2021 at 11:37:28AM +0300, Sakari Ailus wrote:
> Hi Dan,
> 
> On Fri, Apr 16, 2021 at 08:49:41AM +0300, Dan Carpenter wrote:
> > On Fri, Apr 16, 2021 at 12:21:58AM +0300, Sakari Ailus wrote:
> > > On Thu, Apr 15, 2021 at 08:59:41PM +0100, Matthew Wilcox wrote:
> > > > On Thu, Apr 15, 2021 at 08:57:04PM +0100, Matthew Wilcox wrote:
> > > > > On Thu, Apr 15, 2021 at 10:49:55PM +0300, Sakari Ailus wrote:
> > > > > > On Thu, Apr 15, 2021 at 06:14:09PM +0100, Matthew Wilcox wrote:
> > > > > > > On Thu, Apr 15, 2021 at 02:08:19PM -0300, Aline Santana Cordeiro wrote:
> > > > > > > > -const struct atomisp_format_bridge *get_atomisp_format_bridge_from_mbus(
> > > > > > > > -    u32 mbus_code);
> > > > > > > > +const struct atomisp_format_bridge*
> > > > > > > > +get_atomisp_format_bridge_from_mbus(u32 mbus_code);
> > > > > > > 
> > > > > > > No, this does not match coding style.  Probably best to break the
> > > > > > > 80-column guideline in this instance.  Best would be to have a function
> > > > > > 
> > > > > > Having the return type on the previous line is perfectly fine. There should
> > > > > > be a space before the asterisk though.
> > > > > 
> > > > > No, it's not.  Linus has ranted about that before.
> > > > 
> > > > Found it.  https://lore.kernel.org/lkml/1054519757.161606@palladium.transmeta.com/
> > > 
> > > Two decades ago, really?
> > > 
> > > This is simply one of the practical means how you split long function
> > > declarations and avoid overly long lines. Not my favourite though, but
> > > still better than those long lines.
> > 
> > I've always thought we allow either style, but it has to be done
> > consistently within the file.  I was pretty sure that was policy but
> > it's another thing that goes back decades so I don't have a reference.
> > It shouldn't be about breaking up long lines.
> > 
> > > 
> > > My personal preference would be to wrap at the opening parenthesis and
> > > indent by just a tab, but I know many people who disagree with that...
> > 
> > If you're running into the 80 character limit, then it's fine to use
> > two tabs.  I think we have been rejecting patches that push align the
> > parameters but push past the 80 character limit.  Using one tab is
> > confusing because it makes the decalarations line up with the code.
> 
> Interesting. Do you have an example of this? I've thought checkpatch.pl
> gave a warning if the line ended with an opening parenthesis no matter
> what.

The prefered style is still aligning with the parentheses but if you
have to choose between a warning about going over the limit or a warning
about aligning then probably it's fine to not align.

I can't find an example, but I'm pretty sure we've been rejecting
patches that align the parameters but now go over the 80/100 char limit.

regards,
dan carpenter