LWSRV - the simple version
--------------------------
Defines are best set by changing m4.setup and reconfiguring your
makefiles. See [simpleflags] and [lwsrvflags] in m4.setup.
The previous version of lwsrv ran in a single fork and only allowed a
single job at a time. The new version of lwsrv runs in multiple forks
allowing multiple jobs to be spooled concurrently. You can get the
old behavior by running lwsrv with the "-S" switch. A problem that
comes up because lwsrv runs muliforking now is that status messages do
not get updated properly and the mac sees "Starting job" until the job
is submitted.
Some spoolers, like transcript will have problems spooling files with
characters with the 8th bit on. You can make lwsrv quote these chars
by using the "-T quote8bit" option. This differs from the previous
behavior. In addition, some Transcript filters like psrev only allow
line feed as the line terminator instead of both line feed and the
standard on the Macintosh: carriage return. You can use the "-T
crtolf" option to force lwsrv to map carriage return to linefeed.
(This could potentially cause problems with binary data, strings, etc).
If all the "client" programs that spool to lwsrv conform to the Adobe
Systems Document Structuring Conventions, Version 2.0, you can define
"ADOBE_DSC2_CONFORMANT" (SIMPLEFLAGS). Actually, all this does right
now is include the procset at a different point in the file. Turning
this on can cause problems if the client doesn't follow the
conventions correctly. Note: you can also use the "-A on|off" flag to
turn this on and off. The define simply sets the default.
The font list files distributed define the font coordination entries
appropriate for a LaserWriter or LaserWriter Plus. Under old
LaserPreps, there was a "lsf" function defined to return these values.
Under the newer LaserPrep, the code is encapsulated in a query that
can be retrieved by running lwsrv with tracing on. The approriate
code is framed with a "%%?BeginFontListQuery", "%%?EndFontListQuery:
(*)" pair and the last item is indicated with a "*".
Unlike lsrv, lwsrv will prepend an AppleDict (from the LaserPrep).
This allows different users with conflicting LaserPreps on their Macs
to print on the same LaserWriter. However, we do not supply these
files because they contain Apple copyrighted material that it would be
inappropriate to distribute without permission from Apple.
However, it isn't hard to pick out the information. First some
terminology. The LaserPrep contains what Apple calls an AppleDict.
The AppleDict is really known as a ProcSet under the Postscript
document struction conventions.
For newer versions of LaserPrep following version 2.0 of the Adobe
PostScript document structuring conventions, unknown ProcSets are
automatically captured (with the ExitServer code stripped).
The Procset is recorded into a file called "FoundProcSet.<time>". The
first line of the file is "%%BeginProcSet: <tag>" where <tag>
describes the particular procset. This line is critical since this is
how lwsrv determines that this is a valid ProcSet.
Unfortunately, the LaserPrep's "FoundProcSet" files are not usable
without some restructuring. The problem lies in the two special
operators that are downloaded as hex - the way it is done, it assumes
that an end of file follows the end of ProcSet. You can also this on
a case by case basis by editing the appropriate prep files as follows.
Remove these bits of code for "stretch" and "smooth". They start with:
currentfile ok userdict/stretch known not and{eexec}{flushfile}ifelse
.. long hex string ..
.. some more hex strings ..
cleartomark
If you want these operators to be installed into your LaserWriter,
then the best thing to do is to put the above bits of code into
another file with the proper "exit to server" code (systemdict begin
<password> exittoserver) and have it print at regular intervals. This
should work since this stretch and smooth code seems to be independent
of the LaserPrep version.
For older LaserPreps. The structuring isn't as nice, so you'll have a
harder time figuring out where the dictionary ends and the document
begins. Also, the exit to server may still be in the FoundProcSet.
(Note: in the computer center here, we have macs on our appletalk
networks with the laserwriters and depend on the macs to download the
"common" version of the laserprep. lwsrv is given a dummy version of
that laserprep which is empty. It is told about other versions, so
those people can print and be happy. In the computer science
department, they have the laserwriters connected via serial lines for
the most part and they use lwsrv to spool from the macs. Since the
appledicts aren't permanently installed, the lw (not pluses) have
enough memory to do tex too!)
An alternative to editing the procset files is to make lwsrv
preprend the following PostScript code that executes the two
"files" (e.g. stream with 2 eofs) as a single job by running lwsrv
with the "-e" option (e for eexec may be inline):
2 {(%stdin) (r) file cvx exec} repeat
in front of the job and puts end of file ("%%EOF") after the prologue
code. This makes two assumptions:
(a) you are using papif without the EOF code turned off
(default is off under the NO_STRUCT conditional)
or you are using another spooler that handles it
(b) You are running without ADOBE_DSC_CONFORMANT turned on.
At the present time we don't recommend doing this.
These are the contents of the former NiCE NeXT User Group NeXTSTEP/OpenStep software archive, currently hosted by Netfuture.ch.