small bug in INN response of article command.

Russ Allbery rra at stanford.edu
Sat Jan 27 11:16:39 UTC 2001


Cor Bosman <cor at xs4all.net> writes:

> The diablo folks (im running a 3tb diablo server myself :) decided to
> stay with the RFC. Actually, the problem is very hard to reproduce in
> Netscape. It might not be this specific problem, but a combination of
> issues.

> What happened was that some articles would get garbled while reading on
> a diablo server, and were fine on an INN server. A packet dump showed
> that the only difference was the return code of INN. When I wanted to
> reproduce this later, I could not reproduce it always.  But since I got
> hardly any complaints (only 2 really), I stopped looking at it.

> I guess it's upto INN to decide if you want to go the RFC way or leave
> it.

Since the affected clients appear to be fairly limited, I've just checked
this fix into CURRENT:

diff -u -r1.99 -r1.100
--- article.c   2000/10/05 00:53:57     1.99
+++ article.c   2001/01/27 11:15:06     1.100
@@ -1,4 +1,4 @@
-/*  $Id: article.c,v 1.99 2000/10/05 00:53:57 rra Exp $
+/*  $Id: article.c,v 1.100 2001/01/27 11:15:06 rra Exp $
 **
 **  Article-related routines.
 */
@@ -655,7 +655,7 @@
            return;
        }
        tart=art;
-       Reply("%d %ld %s %s\r\n", what->ReplyCode, art, what->Item, av[1]);
+       Reply("%d %ld %s %s\r\n", what->ReplyCode, art, av[1], what->Item);
        if (what->Type != STstat) {
            ARTsendmmap(what->Type);
        }

I'll run the fixed version on my new reading server, and if there aren't
any problems I'll backport this to STABLE.

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>


More information about the inn-workers mailing list