# PSFILTER  Change Log
# CHANGES,v 1.14 1997/12/03 08:47:48 papowell Exp

Version 1.2.3
   Added a -Tmaxresponse=nnn option.
   (Patch sent by: Adam Tilghman <agt@ieng9.ucsd.edu>)

Version 1.2.2
   Modified the page counting code as follows:
    1. at end of job,  reads the page count
    2. pauses endpause seconds
    3. rereads the page count, and repeats from 2. until stable.
   Experimentally verified that a endpause value of 5 seconds appears
    to be practical.

Version 1.2.1
   New baseline distribution
   Added the following -T options:
     forcepagecount  - even if -Tstatus=off, forces sending page count
          postscript program.
     forceps - forces sending a small (null) PostScript job to trick
          autosensing printers into thinking that PostScript jobs are
          coming.
     nosync - ^T will not produce information messages
   If you want to get page counts from printers that do not accept the
     Adobe conventions of ^T producing a status message, then:
       ifhp -Tstatus=off -Tforcepagecount -Tforceps -Tnosync
   Apparently the port 9101 (UDP only) AppSocket port is
	 used by Tektronix printers using PhaserShare Ethernet
	 card, such as Phaser340/350/560 to force and respond with a UDP
     status response.  The -Tudp_status_port=printer_ip%port will now
     open a connection to the printer_ip address, on the specified port,
     and send it \n.  It will also only listen for status responses
     on this port as well.  Note that the method used to get page counts
     sends a PostScript file to the printer;  the results are still
     returned on the usual TCP port.  Note that the UDP responses are NOT
     terminated by \n or \r - just packets.  This is truly an ugly printer
     to use for PostScript printing.
     (Reported and suggested by: Bertrand Decouty <Bertrand.Decouty@irisa.fr>)
       (Editorial comment by: Patrick Powell <papowell@astart.com>)

Version 1.1.14
	psfilter's sync_pr() routine was getting confused by printer
	status messages like the following:
	psfilter: %%[ job: x15; status: waiting ]%%
	It would only check the first entry in the status list and so would
	miss the STATUS field completely, even though that's what it was
	waiting for.
	(Reported by: Adam Tilghman <agt@ieng9.ucsd.edu>)
    Support for Adobe Tagged Binary Communications Protcol has been added.
    The -Ttbcp value or -Ztbcp (LPRng -Z option) can be used.
   
Version 1.1.13
    sync.c - read_pr() did not save input read, sync_pr() did not
   have a correct do{}while(!response)
     (Reported by: Wayne Knowles <w.knowles@niwa.cri.nz>)

Version 1.1.12
    Makefile now has a 'exec_prefix' configuration variable, to conform
    to GNU configure conventions.
    psbanner(1) man page now has psbanner as title.
      (Reported by: "Tarjei T. Jensen" <jensentt@uredd.sjo.statkart.no>)

Version 1.1.11
	testif documented as testf - fixed.
	(Reported by: John Perkins <john@cs.wisc.edu>)

Version 1.1.10
	Fixed up format of some status messages to aid in debugging.
	New version of portable.h, acconfig.h, etc.

Version 1.1.9
    autoconf 2.12 configuration file,  handles more systems

Version 1.1.8
	filter() and sync() static variables not initialized on each call.
	  (Reported by: Christian Kuehnke" <Christian.Kuehnke@arbi.Informatik.Uni-Oldenburg.DE)

Version 1.1.7
    Put in BINDEST directory for install
	Added ability to specify origination port to -Tdev=... in hex and
		decimal.

Version 1.1.6
    Makefile.bsd for BSD style makes
	filter.c had bad log() entries
		(Reported by: Zack Weinberg <zack@rabi.phys.columbia.edu>)
Version 1.1.5
    Check of sys_siglist[] definition.
    Updated configure.in, configure, config.h

Version 1.1.4
    filter() - now uses fcntl to check to see if stdout readable.

Version 1.1.3
	Makefile updated to make monitor as well as filters.
	  (Per Foreby <perf@efd.lth.se>)
	filter() - now checks to see if stdout readable, and if not,
	  suppresses getting status information.

Version 1.1.2
	Minor portability changes for various operating systems - syslog.h
	  and others, for example.

Version 1.1.1
    Fixed end of job accounting glitch.

Version 1.0.9
	Pagecount at end of job now pauses slightly before the actual
		pagecount is obtained - this allows the last job to finish.
		Apparently on several HP LaserJet4 and other printers,
		the pagecount is readable before the last page is fed through.
		This means the page count is off by one.

Version 1.0.8
    Fixed problem with trying to read from a file descriptor
		which was not open for reading - terminates with an error

Version 1.0.7
	documentation updates, correction and revision
	Fixed form feed at end of page problem.

Version 1.0.6
	Charles Hedrick <hedrick@klinzhai.rutgers.edu>
		Massive set of fixes for status gathering and printing.
	Patrick Powell <papowell@sdsu.edu>
	Added summary file (one line status file), send summary to UPD port as well.
		monitor and sendtest programs for testing status reporting.
	Note: README printcap information for 'ps=status', (printer status file)
	    was wrong, corrected

Version 1.0.5
	changed longjmp() and setjmp() to siglongjmp() and sigsetjmp()
	where appropriate.

Version 1.0.4 and below
   Basic functionality
