[Linux-kernel-mentees] [PATCH v2] checkpatch: add new exception to repeated word check

[Joe Perches]

On Sat, 2020-10-17 at 10:02 +0530, Dwaipayan Ray wrote:
> On Sat, Oct 17, 2020 at 8:26 AM Joe Perches <joe at perches.com> wrote:
> > On Wed, 2020-10-14 at 11:35 -0700, Joe Perches wrote:
> > > On Wed, 2020-10-14 at 23:42 +0530, Dwaipayan Ray wrote:
> > > > On Wed, Oct 14, 2020 at 11:33 PM Joe Perches <joe at perches.com> wrote:
> > > > > On Wed, 2020-10-14 at 22:07 +0530, Dwaipayan Ray wrote:
> > > > > > Recently, commit 4f6ad8aa1eac ("checkpatch: move repeated word test")
> > > > > > moved the repeated word test to check for more file types. But after
> > > > > > this, if checkpatch.pl is run on MAINTAINERS, it generates several
> > > > > > new warnings of the type:
> > > > > 
> > > > > Perhaps instead of adding more content checks so that
> > > > > word boundaries are not something like \S but also
> > > > > not punctuation so that content like
> > > > > 
> > > > >         git git://
> > > > >         @size size
> > > > > 
> > > > > does not match?
> > > > > 
> > > > > 
> > > > Hi,
> > > > So currently the words are trimmed of non alphabets before the check:
> > > > 
> > > > while ($rawline =~ /\b($word_pattern) (?=($word_pattern))/g) {
> > > > my $first = $1;
> > > > my $second = $2;
> > > > 
> > > > where, the word_pattern is:
> > > > my $word_pattern = '\b[A-Z]?[a-z]{2,}\b';
> > > 
> > > I'm familiar.
> > > 
> > > > So do you perhaps recommend modifying this word pattern to
> > > > include the punctuation as well rather than trimming them off?
> > > 
> > > Not really, perhaps use the capture group position
> > > markers @- @+ or $-[1] $+[1] and $-[2] $+[2] with the
> > > substr could be used to see what characters are
> > > before and after the word matches.
> > 
> > Perhaps something like:
> > ---
> >  scripts/checkpatch.pl | 12 +++++++++++-
> >  1 file changed, 11 insertions(+), 1 deletion(-)
> > 
> > diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl
> > index fab38b493cef..a65eb40a5539 100755
> > --- a/scripts/checkpatch.pl
> > +++ b/scripts/checkpatch.pl
> > @@ -3054,15 +3054,25 @@ sub process {
> > 
> >                                 my $first = $1;
> >                                 my $second = $2;
> > +                               my $start_pos = $-[1];
> > +                               my $end_pos = $+[2];
> > 
> >                                 if ($first =~ /(?:struct|union|enum)/) {
> >                                         pos($rawline) += length($first) + length($second) + 1;
> >                                         next;
> >                                 }
> > 
> > -                               next if ($first ne $second);
> > +                               next if (lc($first) ne lc($second));
> >                                 next if ($first eq 'long');
> > 
> > +                               my $start_char = "";
> > +                               my $end_char = "";
> > +                               $start_char = substr($rawline, $start_pos - 1, 1) if ($start_pos > 0);
> > +                               $end_char = substr($rawline, $end_pos, 1) if (length($rawline) > $end_pos);
> > +
> > +                               next if ($start_char =~ /^\S$/);
> > +                               next if ($end_char !~ /^[\.\,\s]?$/);
> > +
> >                                 if (WARN("REPEATED_WORD",
> >                                          "Possible repeated word: '$first'\n" . $herecurr) &&
> >                                     $fix) {
> > 
> > 
> 
> Hi Joe,
> Thank you for the insight. I was also doing something similar:
> 
> ---
>  scripts/checkpatch.pl | 16 ++++++++++++++++
>  1 file changed, 16 insertions(+)
> 
> diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl
> index f1a4e61917eb..82497a71ac96 100755
> --- a/scripts/checkpatch.pl
> +++ b/scripts/checkpatch.pl
> @@ -595,6 +595,7 @@ our @mode_permission_funcs = (
>  );
> 
>  my $word_pattern = '\b[A-Z]?[a-z]{2,}\b';
> +my $punctuation_chars = '[,:;@\.\-]';
> 
>  #Create a search pattern for all these functions to speed up a loop below
>  our $mode_perms_search = "";
> @@ -3065,6 +3066,21 @@ sub process {
>   next if ($first ne $second);
>   next if ($first eq 'long');
> 
> + # check for character before and after the word matches
> + my $ca_first = substr($rawline, $-[1]-1, 1);
> + my $cb_first = substr($rawline, $+[1], 1);
> + my $ca_second = substr($rawline, $-[2]-1, 1);
> + my $cb_second = substr($rawline, $+[2], 1);
> +
> + if ($ca_first ne $ca_second || $cb_first ne $cb_second) {
> + if ($ca_first =~ /$punctuation_chars/ ||
> +     $ca_second =~ /$punctuation_chars/ ||
> +     $cb_first =~ /$punctuation_chars/ ||
> +     $cb_second =~ /$punctuation_chars/) {
> + next;
> + }
> + }
> +
>   if (WARN("REPEATED_WORD",
>   "Possible repeated word: '$first'\n" . $herecurr) &&
>       $fix) {
> 
> Does it look okay to you??

Not really, as ca_second and cb_first are both known
to be the same position and known to be a single space.