On Mon, Nov 19, 2012 at 10:18:35PM +0100, Matthias Kilian wrote:
> On Mon, Nov 19, 2012 at 11:09:20AM -0700, Antoine Jacoutot wrote:
> > CVSROOT:    /cvs
> > Module name:        ports
> > Changes by: [email protected]       2012/11/19 11:09:20
> > 
> > Modified files:
> >     print/foomatic-filters: Makefile 
> >     print/foomatic-filters/pkg: README 
> > 
> > Log message:
> > Add a printcap(5) example for remote queues... found out and debugged by
> > deraadt@ the hard way.
> 
> I'm really surprised that this works. After having some looks at
> lpd during the past days, I think that an entry with
> 
>       ...:[email protected]:...:if=${PREFIX}/bin/foomatic-rip:...
> 
> would cause the raw output of foomatic-rip to be sent over to

The raw output of foomatic-rip is the actual print-formatted job that the 
remote printer understands.
Otherwise, you would just send raw print to it (without any filtering) and the 
printer would print rubish.

> 1.2.3.4, port 515. And if there's an lpd listening on that receiving
> and (1.2.3.4, port 515), I would expect it to log a "bad request
> (..) from ..." allmost immediately (or after receiving some bytes)
> and closing the connection, causing a SIGPIPE somewhere in foomatic-rip
> (or the programs invoked by it). That's what I've seen during my
> own tests, and after reading that evil lpd code ;-)
> 
> Ciao,
>       Kili
> 

-- 
Antoine

Reply via email to