cups-lpd: compatibility with PLP

[H. Blischke]

Patrick Powll's PLP (Public Line Printer) behaves, even if configured to use "Berkeley protocol", slightly different from what RFC 1179, as pointed out in the following syslog snippet:


Mar  6 21:21:29 obelix cups-lpd[20107]: connect from clienthost
Mar  6 21:21:29 obelix cups-lpd[20107]: Connection from clienthost
Mar  6 21:21:29 obelix cups-lpd[20107]: Receive print job for printername
Mar  6 21:21:29 obelix cups-lpd[20107]: Receive print job for printername
Mar  6 21:21:29 obelix cups-lpd[20107]: [03] 254631 dfA589clienthost
Mar  6 21:21:29 obelix cups-lpd[20107]: [02] 137 cfZ589clienthost
Mar  6 21:21:29 obelix cups-lpd[20107]: [00] (* final EOD *)
Mar  6 21:21:29 obelix cups-lpd[20107]: Control file:
Mar  6 21:21:29 obelix cups-lpd[20107]:         D1236370629
Mar  6 21:21:29 obelix cups-lpd[20107]:         Hclienthost
Mar  6 21:21:29 obelix cups-lpd[20107]:         Proot
Mar  6 21:21:29 obelix cups-lpd[20107]:         N/tmp/temp.manu
Mar  6 21:21:29 obelix cups-lpd[20107]:         J/tmp/temp.manu
Mar  6 21:21:29 obelix cups-lpd[20107]:         CZ
Mar  6 21:21:29 obelix cups-lpd[20107]:         fdfA589clienthost
Mar  6 21:21:29 obelix cups-lpd[20107]:         UdfA589clienthost
Mar  6 21:21:30 obelix cups-lpd[20107]: Print file - job ID = 275915
Mar  6 21:21:30 obelix cups-lpd[20107]: Closing connection

The whole bunch of data and control files is followd by a single zero byte as a final EOF indicator.

The attached patch account for this and additionally implents a "-d" command line switch to force diagnostic output to syslog.