Folke Ashberg writes:
>> If there's a problem, it must lie elsewhere. It'snot here.
>
> not my opinion. Using sendmail there are no problems. And i think this
You don't get it. The problem was not in the rewrite function. The was
nothing wrong with it. The problem was elsewhere - the parser did not
correctly set the ending position of the MIME section in the original
message.
The following patch fixes the real problem. Now, I just have to make sure
that nothing else depended on the original broken logic.
--
Sam
Index: rfc2045/rfc2045.c
===================================================================
RCS file: /cvsroot/courier/libs/rfc2045/rfc2045.c,v
retrieving revision 1.28
diff -U3 -r1.28 rfc2045.c
--- rfc2045/rfc2045.c 2001/11/26 04:03:51 1.28
+++ rfc2045/rfc2045.c 2002/01/24 18:06:48
@@ -441,7 +441,7 @@
if (p->lastpart->workclosed)
{
- update_counts(p, p->endpos+cnt, p->endpos+cnt, 1);
+ update_counts(p, p->endpos+cnt, p->endpos+n, 1);
return;
}
/* Leftover trash -- workclosed is set when the final
