Re: [PATCH v2] checkpatch: add double empty line check

[Andy Whitcroft]

On Tue, Nov 20, 2012 at 05:07:07PM +0200, Eilon Greenstein wrote:
> On Tue, 2012-11-20 at 14:43 +0000, Andy Whitcroft wrote:
> 
> > > > Also this fails if the fragment
> > > > is at the top of the hunk emiting a perl warning.
> > > 
> > > I did not see this warning. Can you please share this example? I tried
> > > adding a couple of empty lines at the beginning of a file and it seemed
> > > to work just fine for me (using perl v5.14.2).lines
> > 
> > Ok, this is actually if it is at the bottom, not the top.  So if you
> > have a range of lines newly added to the bottom of the file.  Leading to
> > this warning:
> > 
> > Use of uninitialized value within @rawlines in pattern match (m//) at
> > ../checkpatch/scripts/checkpatch.pl line 3586.
> 
> Oh... of course, I should have seen that. I did not check changes at the
> end of the file.
> 
> What do you say about something like that (using nextline out of
> rawlines only if it defined):
> diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl
> index 21a9f5d..c0c610c 100755
> --- a/scripts/checkpatch.pl
> +++ b/scripts/checkpatch.pl
> @@ -3579,6 +3579,19 @@ sub process {
>                         WARN("EXPORTED_WORLD_WRITABLE",
>                              "Exporting world writable files is usually an error. Consider more restrictive permissions.
>                 }
> +
> +# check for double empty lines
> +               if ($line =~ /^\+\s*$/) {
> +                       my $nextline = "";
> +                       if (defined($rawlines[$linenr])) {
> +                               $nextline = $rawlines[$linenr];
> +                       }
> +                       if ($nextline =~ /^\s*$/ ||
> +                           $prevline =~ /^\+?\s*$/ && $nextline !~ /^\+\s*$/) {
> +                               CHK("DOUBLE_EMPTY_LINE",
> +                                   "One empty line should be sufficient. Consider removing this one.\n" . $herecurr);
> +                       }
> +               }
>         }
>  
> 
> 

You cannot really rely on nextline even if valid as it may not even be
from this hunk.  This is why in my attempt I detected the top of a long
run of blanks and let the hunk start initialisation reset the detector.

-apw