Subject: Re: [netatalk-admins] Postscript and Netatalk problem
From: Andras Kadinger (bandit@freeside.elte.hu)
Date: Fri Nov 06 1998 - 11:24:24 EST
PayPC System Mail Subscriber wrote:
>
> Lee Blevins said in [netatalk-admins] Postscript and Netatalk problem at
> 06/Nov/1998 (Fri) 22:14:24.
> >
> > My instinct tells me this is cr/lf related since the ps file is a text
> > file.
>
> I strongly suspect your instincts are correct. Try configuring your netatalk
> to NOT convert cr/lfs on any TEXT files.... If your scitex is expecting (at
> least) CRs as line terminators, it could be very cranky.
Also don't forget please, a PostScript stream is not necessarily a text
file, but _can_ contain binary parts (images, sometimes font
definitions, etc.). However, line-endings themselves are not a concern,
since the PostScript language can freely accept and mix all three
popular line ending conventions: LF (unix), CR (Mac), CR/LF (DOS); even
LF/CR should be supported now. I don't know what 'cranky' means, but I
would argue that an implementation which doesn't support all variations
is broken. CR/LF conversion usually causes problems only, when it
happens in binary data; ASCII streams should survive.
PostScript interpreters usually emit error messages similar to this:
error: <error type> offending command: <command or text causing the
error>
example:
error: undefined offending command: this_is_undefined
Here, the error is 'undefined' (undefined command, that is a command not
known to the interpreter), and the offending command (the command/text
causing the error) is 'this_is_undefined'.
Sincerely,
Andras Kadinger
bandit@freeside.elte.hu
This archive was generated by hypermail 2b28 : Sat Dec 18 1999 - 16:33:40 EST