Re: [PATCH RFC] checkpatch: add new warnings to author signoff checks.

[Lukas Bulwahn]

On Mon, 5 Oct 2020, Dwaipayan Ray wrote:

> The author signed-off-by checks are currently very vague.
> Cases like same name or same address are not handled separately.
> 
> For example, running checkpatch on commit be6577af0cef
> ("parisc: Add atomic64_set_release() define to avoid CPU soft lockups"),
> gives:
> 
> WARNING: Missing Signed-off-by: line by nominal patch author
> 'John David Anglin <dave.anglin@xxxxxxxx>'
> 
> The signoff line was:
> "Signed-off-by: Dave Anglin <dave.anglin@xxxxxxxx>"
> 
> Clearly the author has signed off but with a slightly different version
> of his name. A more appropriate warning would have been to point out
> at the name mismatch instead.
> 
> Introduced three new types of warnings:
> 
> 1) Address matches, but names are different.
>    "James Watson <james@xxxxxxxxx>", "James <james@xxxxxxxxx>"
> 
> 2) Name matches, but addresses are different.
>    "James Watson <james@xxxxxxxxxx>", "James Watson <james@xxxxxxxxx>"
> 
> 3) Name matches, but addresses without mail extensions are same.
>    "James Watson <james@xxxxxxxxx>", "James Watson <james+a@xxxxxxxxx>"
> 
> For the 3rd class, a --strict check message is generated, and for the
> other two, warnings are generated.
>

I will start running an evaluation on checkpatch.pl before and after this 
patch to get some insight on this change.
 
> Signed-off-by: Dwaipayan Ray <dwaipayanray1@xxxxxxxxx>
> ---
>  scripts/checkpatch.pl | 57 ++++++++++++++++++++++++++++++++++++++++---
>  1 file changed, 53 insertions(+), 4 deletions(-)
> 
> diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl
> index 31624bbb342e..80feb15f93cb 100755
> --- a/scripts/checkpatch.pl
> +++ b/scripts/checkpatch.pl
> @@ -2347,6 +2347,7 @@ sub process {
>  	my $signoff = 0;
>  	my $author = '';
>  	my $authorsignoff = 0;
> +	my $authorsignoff_ctx = '';
>  	my $is_patch = 0;
>  	my $is_binding_patch = -1;
>  	my $in_header_lines = $file ? 0 : 1;
> @@ -2674,9 +2675,34 @@ sub process {
>  		if ($line =~ /^\s*signed-off-by:\s*(.*)/i) {
>  			$signoff++;
>  			$in_commit_log = 0;
> -			if ($author ne '') {
> +			if ($author ne ''  && $authorsignoff != 1) {
>  				if (same_email_addresses($1, $author)) {
>  					$authorsignoff = 1;
> +				} else {
> +					my $ctx = $1;
> +					my ($email_name, $email_comment, $email_address, $comment1) = parse_email($ctx);
> +					my ($author_name, $author_comment, $author_address, $comment2) = parse_email($author);
> +
> +					if($email_address eq $author_address) {
> +						$authorsignoff_ctx = $ctx;
> +						$authorsignoff = 2;
> +					} elsif ($email_name eq $author_name) {
> +						$authorsignoff_ctx = $ctx;
> +						$authorsignoff = 3;
> +
> +						my $address1 = $email_address;
> +						my $address2 = $author_address;
> +
> +						if ($address1 =~ /(\S+)\+\S+(\@.*)/) {
> +							$address1 = $1.$2;
> +						}
> +						if ($address2 =~ /(\S+)\+\S+(\@.*)/) {
> +							$address2 = $1.$2;
> +						}
> +						if($address1 eq $address2) {
> +							$authorsignoff = 4;
> +						}
> +					}
>  				}
>  			}
>  		}
> @@ -6891,9 +6917,32 @@ sub process {
>  		if ($signoff == 0) {
>  			ERROR("MISSING_SIGN_OFF",
>  			      "Missing Signed-off-by: line(s)\n");
> -		} elsif (!$authorsignoff) {
> -			WARN("NO_AUTHOR_SIGN_OFF",
> -			     "Missing Signed-off-by: line by nominal patch author '$author'\n");
> +		} elsif ($authorsignoff != 1) {
> +			# authorsignoff values:
> +			# 0 -> missing sign off
> +			# 1 -> sign off present
> +			# 2 -> address matches, name different
> +			# 3 -> name matches, address different
> +			# 4 -> name matches, address matches without extension
> +
> +			my $ctx_msg = "'Signed-off-by: $authorsignoff_ctx' should be:\n'Signed-off-by: $author'";
> +
> +			if($authorsignoff == 0) {
> +				WARN("NO_AUTHOR_SIGN_OFF",
> +					"Missing Signed-off-by: line by nominal patch author '$author'\n");
> +			}

I think in this case, we could actually turn this into an ERROR; now that 
we have the refined cases (2,3,4) on which we would just warn or 'note' 
with --strict checks.

Lukas

> +			elsif($authorsignoff == 2) {
> +				WARN("NO_AUTHOR_SIGN_OFF",
> +					"Author name mismatch:\n$ctx_msg\n");
> +			}
> +			elsif($authorsignoff == 3) {
> +				WARN("NO_AUTHOR_SIGN_OFF",
> +					"Author address mismatch:\n$ctx_msg\n");
> +			}
> +			elsif($authorsignoff == 4) {
> +				CHK("NO_AUTHOR_SIGN_OFF",
> +					"Author mail extension mismatch:\n$ctx_msg\n");
> +			}
>  		}
>  	}
>  
> -- 
> 2.27.0
> 
>