rsync - Jan 2002 - Fix for "Warning: unexpected read size of 0 in map_ptr" in rsync 2.5.2

I upgraded to rsync 2.5.2 and started to get the error
    Warning: unexpected read size of 0 in map_ptr
when sending just about any file from a Solaris 2.5.1 system to Linux
systems, and then the file that is sent ends up being the wrong size.

Through gdb and trial & error I discovered that if I back out the security
fix in rsync.h at
   
http://cvs.samba.org/cgi-bin/cvsweb/rsync/rsync.h.diff?r1=1.116&r2=1.117&f=h
and recompile match.c on only the sender side then the problem goes away.
Further trial & error resulted in the patch below to fix the problem.

Martin, I think this is probably serious enough to warrant an immediate new
release.

- Dave Dykstra


--- match.c.O	Tue Jan 29 15:31:37 2002
+++ match.c	Tue Jan 29 15:31:54 2002
@@ -246,7 +246,7 @@
 		   match. The 3 reads are caused by the
 		   running match, the checksum update and the
 		   literal send. */
-		if (offset-last_match >= CHUNK_SIZE+s->n && 
+		if (offset-last_match >= CHUNK_SIZE+(int)s->n && 
 		    (end-offset > CHUNK_SIZE)) {
 			matched(f,s,buf,offset - s->n, -2);
 		}

I happened to notice that Redhat fixed this on Sunday already:
    http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=58878

I wish one of them would have reported it to rsync@samba.org.

- Dave Dykstra

On Tue, Jan 29, 2002 at 03:36:18PM -0600, Dave Dykstra
wrote:
> I upgraded to rsync 2.5.2 and started to get the error
>     Warning: unexpected read size of 0 in map_ptr
> when sending just about any file from a Solaris 2.5.1 system to Linux
> systems, and then the file that is sent ends up being the wrong size.
> 
> Through gdb and trial & error I discovered that if I back out the
security
> fix in rsync.h at
>    
http://cvs.samba.org/cgi-bin/cvsweb/rsync/rsync.h.diff?r1=1.116&r2=1.117&f=h
> and recompile match.c on only the sender side then the problem goes away.
> Further trial & error resulted in the patch below to fix the problem.
> 
> Martin, I think this is probably serious enough to warrant an immediate new
> release.
> 
> - Dave Dykstra
> 
> 
> --- match.c.O	Tue Jan 29 15:31:37 2002
> +++ match.c	Tue Jan 29 15:31:54 2002
> @@ -246,7 +246,7 @@
>  		   match. The 3 reads are caused by the
>  		   running match, the checksum update and the
>  		   literal send. */
> -		if (offset-last_match >= CHUNK_SIZE+s->n && 
> +		if (offset-last_match >= CHUNK_SIZE+(int)s->n && 
>  		    (end-offset > CHUNK_SIZE)) {
>  			matched(f,s,buf,offset - s->n, -2);
>  		}