truss bug + PATCH, pls review

[Sam Lawrance]

On Thu, 2005-03-03 at 07:32 -0700, Scott Long wrote:
> Alfred Perlstein wrote:
> > Can someone review this?  I think 'u' is incorrectly
> > added to instead of assigned to.  This causes the initial
> > calculation to be garage based and screws up displaying
> > poll information.
> > 
> > I'd like this to be MFC'd before 5.4 if possible.
> > 
> > Index: syscalls.c
> > ===================================================================
> > RCS file: /home/ncvs/src/usr.bin/truss/syscalls.c,v
> > retrieving revision 1.45
> > diff -u -r1.45 syscalls.c
> > --- syscalls.c	5 Sep 2004 05:27:30 -0000	1.45
> > +++ syscalls.c	2 Mar 2005 21:19:13 -0000
> > @@ -414,7 +414,7 @@
> >  	(POLLIN | POLLPRI | POLLOUT | POLLERR | POLLHUP | POLLNVAL | \
> >  	 POLLRDNORM |POLLRDBAND | POLLWRBAND | POLLINIGNEOF) 
> >  
> > -	  u += snprintf(tmp + used, per_fd,
> > +	  u = snprintf(tmp + used, per_fd,
> >  	    "%s%d 0x%hx%s%s%s%s%s%s%s%s%s ",
> >  	    i > 0 ? " " : "",
> >  	    pfd[i].fd,
> > 

While you're on the subject of Trussburger, perhaps someone can comment
on this Problem Report Meal. A McPatch is included.

http://www.freebsd.org/cgi/query-pr.cgi?pr=71749

PS. It really is a McPatch.