From owner-freebsd-doc  Fri Dec 29 15:41:01 1995
Return-Path: owner-doc
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id PAA23275
          for doc-outgoing; Fri, 29 Dec 1995 15:41:01 -0800 (PST)
Received: from parody.tecc.co.uk (parody.tecc.co.uk [193.128.6.83])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id PAA23262
          for <doc@freebsd.org>; Fri, 29 Dec 1995 15:40:56 -0800 (PST)
Received: (from james@localhost) by parody.tecc.co.uk (8.6.12/8.6.12) id OAA00366; Fri, 29 Dec 1995 14:37:15 GMT
Date: Fri, 29 Dec 1995 14:37:13 +0000 ()
From: James Raynard <james@parody.tecc.co.uk>
To: doc@freebsd.org
Subject: Handbook - printer setup
Message-ID: <Pine.BSF.3.91.951229135243.322B-100000@parody.tecc.co.uk>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-doc@freebsd.org
Precedence: bulk

I spent most of last night trying to set up a printer for the first time
ever under Unix following the Handbook step-by-step. On the whole, it was 
quite a good set of instructions, but there were one or two points that 
may be worth picking up.

Due to the staircase effect, the "Printer on a Parallel Port" test 
appeared to only produce one line of output on my DeskJet 520, which was
a bit alarming until I realised what was happening. It does say "You should 
see something print. Don't worry if the text doesn't look right; we'll 
fix such things later"; perhaps this could be changed to "You should see 
something print. Don't worry if the text doesn't look right, or if it
disappears at the end of the first line; we'll fix such things later."?

The bit that really had me tearing my hair out was when I tried to use
the suggested text filter - I kept getting errors 'execv of 
/usr/local/libexec/hpif failed'. After checking the paths and were correct, 
checking that the filter was world-readable and executable and successfully 
piping into and out of the filter by hand I was at a complete loss.

I then noticed that the supplied /usr/libexec/lpf didn't have this problem.
It did cross my mind that it could be a problem with something being setuid.
However, the Handbook says that the text filter is run by lpd, which is not
setuid or setgid on my system, so I assumed it was something else.

As I had by now completely run out of ideas, I compiled a C program which 
did the same as hpif. This worked! So the only explanation I can think of
is that it is a setuid problem and it's actually one of the other lp?
programs which execv's the filter (sorry, I know nothing about how print
spoolers work).

Anyway, something needs changing in this part of the documentation. If it's
any use, here's the C program I used:-

#include <stdio.h>

#define LINELEN 255
#define INITSTR "\033&k2G"

main() {
    char buf[LINELEN];

    if (printf(INITSTR) == 0)
        exit(2);

    while (fgets(buf, LINELEN, stdin))
        printf(buf);

    printf("\f");
    return 0;
}

(Yes, this is the first time I've ever checked the return value from 
printf 8-)

And here's the one I used for Ghostscript (I used different printer
names for text and gs, so I didn't bother with the magic number test):-

#include <stdio.h>

#define LINELEN 255

main() {
    char buf[LINELEN];
    FILE *fp;

    if ((fp = popen("/usr/local/bin/gs -dSAFER -dNOPAUSE -q \
-sDEVICE=deskjet -sOutputFile=- -", "w")) == NULL)
        exit(2);
    while (fread(buf, 1, LINELEN, stdin) == LINELEN)
        if(fwrite(buf, 1, LINELEN, fp) != LINELEN)
            exit(2);

    pclose(fp);
    printf("\f");
    return 0;
}

This had me tearing my hair out as well, due to a Ghostscript bug which
I've reported to them (it was using an undocumented feature of the 
DeskJet500 that doesn't work on my DJ520), but that's another story!

Cheers
James

Segmentation fault (core dumped): cannot find file '.signature'