[RFC patch] checkpatch: test identifier lengths

[Andrew Morton]

On Fri, 16 Feb 2018 09:13:27 -0800 Joe Perches <joe at perches.com> wrote:

> On Fri, 2018-02-16 at 15:55 +0300, Dan Carpenter wrote:
> > On Fri, Feb 16, 2018 at 05:06:34PM +0530, Yash Omer wrote:
> > > This patch fix line should not end with open parenthesis found by checkpatch.plscript.
> > > 
> > > Signed-off-by: Yash Omer <yashomer0007 at gmail.com>
> > > ---
> > >  drivers/staging/nvec/nvec.c | 4 ++--
> > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/drivers/staging/nvec/nvec.c b/drivers/staging/nvec/nvec.c
> > > index 52054a528723..39fb737543b5 100644
> > > --- a/drivers/staging/nvec/nvec.c
> > > +++ b/drivers/staging/nvec/nvec.c
> > > @@ -383,8 +383,8 @@ static void nvec_request_master(struct work_struct *work)
> > >  		msg = list_first_entry(&nvec->tx_data, struct nvec_msg, node);
> > >  		spin_unlock_irqrestore(&nvec->tx_lock, flags);
> > >  		nvec_gpio_set_value(nvec, 0);
> > > -		err = wait_for_completion_interruptible_timeout(
> > > -				&nvec->ec_transfer, msecs_to_jiffies(5000));
> > > +		err = wait_for_completion_interruptible_timeout
> > > +			(&nvec->ec_transfer, msecs_to_jiffies(5000));
> > 
> > The original code is basically fine...  It's OK to ignore checkpatch in
> > this situation.
> 
> Right.

Yes, I'd say that checkpatch is simply wrong here.  I'd prefer that a
function call always have the opening paren hard up against the
function name.  Because I often search for "foo(" to find the callsites
of foo() and I expect that some code-parsing tools do the same thing. 
The "(" is the application of an operator to an identifier.

So I'd vote for simply nuking that checkpatch warning altogether. 
Maybe there are other situations in which it is useful, dunno.