Hi Ken & Paul,
Kenneth R Westerback wrote on Tue, Nov 15, 2011 at 07:09:03PM -0500:
> On Wed, Nov 16, 2011 at 01:54:24AM +0200, Paul Irofti wrote:
>> I see this when packaging, maybe you can talk to ingo and fix it?
>> This did not happen with the former package (from looking at the
>> archives).
>>
>> an1/sbcl.1:44: a newline character is not allowed in an escape name
As that port wants to USE_GROFF, voila a non-fatal error message
from groff. And that's probably already an exaggeration on the
part of groff; a mere warning ought to be sufficient.
When running groff during pkg_create(1), we already inhibit *some*
warnings (-Wall). Maybe we should also inhibit all the remaining
warnings (-Ww) and also inhibit non-fatal errors (-E). *If* you
want to audit manuals for quality to help upstream with patches,
mandoc -Tlint is probably the better choice, compared to
groff -wall, and i assume that at most some porters want to spend
their time on that.
See below for a patch. Testing it on sbcl contributed considerably
to the heating of the hackroom; but yes, it does rid us of the error
message.
OK?
Ingo
> I saw that message but somehow missed that it was man page related.
> I will indeed ping Ingo.
Hm, it looks like Ingo woke up without being ping'ed.
Index: OpenBSD/PackingElement.pm
===================================================================
RCS file: /cvs/src/usr.sbin/pkg_add/OpenBSD/PackingElement.pm,v
retrieving revision 1.200
diff -u -p -r1.200 PackingElement.pm
--- OpenBSD/PackingElement.pm 23 Aug 2011 10:32:27 -0000 1.200
+++ OpenBSD/PackingElement.pm 16 Nov 2011 09:47:25 -0000
@@ -593,8 +593,8 @@ sub format
chdir($dir) or die "Can't chdir to $dir";
},
OpenBSD::Paths->groff,
- qw(-Tascii -mandoc -Wall -mtty-char -P -c), @extra, '--',
- $file);
+ qw(-mandoc -mtty-char -E -Ww -Tascii -P -c),
+ @extra, '--', $file);
} else {
die "Can't parse source name $fname";
}
