[cups.bugs] [MOD] STR #4150: TBCP quoting broken in combination with PJL

[Stefan Brüns]

DO NOT REPLY TO THIS MESSAGE.  INSTEAD, POST ANY RESPONSES TO THE LINK BELOW.

[STR New]

currently, a Job in Postscript format will be get a PJL preambel and
postambel, given the corresponding options in the PPD are set.

This is done by the pstops filter.

Afterwards, the data stream might be mangled with the TBCP port monitor.

The second step is broken:

As the TBCP filter does not know the end of the PS data, it cannot
recognize the UEL (ESC %-12345X)  sequence in front of the PJL EOJ as
such, but quotes the ESC character of the UEL.

The copying of PJL sequences is broken, as the TBCP start marker (0x01
0x4d) is inserted before the "PJL ENTER LANGUAGE".

If a job containing two files (i.e. lp file1.ps file2.ps) is submitted,
the two jobs are send concatenated to the backend, whereas the broken PJL
EOJ sequence (prefixed with a broken UEL) causes the PS interpreter at
least on Kyoceras to drop the remaining data.

The TBCP quoting should be part of the pstops filter, and be applied to
all PS jobs with PJL header (see Adobe Serial and Parallel Communications
Protocols Specification 5009:
"The tagged binary communicationsprotocol works well, and Adobe recommends
its use, on printers that have a language independent software component
and arbitrates which language will interpret an incoming job, such as
Hewlett Packard’s Printer Job Language (PJL). ")

TBCP should be applied if a PS file does not already contain a TBCP start
marker.

Link: http://www.cups.org/str.php?L4150
Version: 1.6-current