Hi Joe,
On Wed, Sep 04, 2019 at 08:37:08AM -0700, Joe Perches wrote:
> On Wed, 2019-09-04 at 18:04 +0300, Sakari Ailus wrote:
> > On Tue, Sep 03, 2019 at 03:06:07PM +0200, Petr Mladek wrote:
> > > On Mon 2019-09-02 11:32:39, Sakari Ailus wrote:
> []
> > > > diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl
> []
> > > > @@ -5995,7 +5995,8 @@ sub process {
> > > > while ($fmt =~ /(\%[\*\d\.]*p(\w))/g) {
> > > > $specifier = $1;
> > > > $extension = $2;
> > > > - if ($extension !~ /[SsBKRraEhMmIiUDdgVCbGNOxt]/) {
> > > > + if ($extension !~ /[SsBKRraEhMmIiUDdgVCbGNOxtf]/ ||
> > > > + $extension =~ /^f[^w]/) {
> > >
> > > This does not work. $extension seems to have only one character.
> >
> > Good catch. \w indeed matches a single letter; I'll change that to \w+ and
> > change the other uses accordingly.
>
> If you want to make changes to checkpatch, please
> send patches to the checkpatch maintainers.
>
> Don't break other parsing of $2/#extension.
>
> If you really need to know whatever follows the specific
> extension letter use another capture group.
>
> while ($fmt =~ /(\%[\*\d\.]*p(\w)(\w*))/g) {
> $specifier = $1;
> $extension = $2;
> $qualifier = $3;
>
> etc...
>
> Then verify $qualifier or $3 is not undef if necessary
There are just a couple of users, but indeed, the extension handlers work
based on a single letter as well so this way it's better aligned with the
kernel implementation. Forks for me.
--
Regards,
Sakari Ailus
sakari.ailus@xxxxxxxxxxxxxxx
