Bug 56503

Created attachment 37988 [details]
gnumeric print output

Created attachment 37989 [details]
psnup -2 -pletter -d foo.ps output

Created attachment 37990 [details]
mpage -2 foo.ps output

I confirm that foo2.ps is broken. I can look into this problem
closer if Red Hat agrees to support the project.

Reconfirmed that this is still broken on FC3 with psutils-1.17-19.

[rangek@asimov psnup]$ psnup -2 -pletter -d foo.ps foo_new.ps
[1] Wrote 1 pages, 55400 bytes
[rangek@asimov psnup]$ diff -u foo2bad.ps foo_new.ps

That should be psutils-1.17-23 in FC3.  (I had a window on a FC9 box up.)

I have a fix for this problem. It is submitted to both psnup and cups
maintainers, however, it is not clear to me if the patch is accepted. The
problem is in cups, not in psnup, because the postscript files print when
submitted directly to the printer. However, the cups maintainer is saying that
psnup is not creating complient files. The fix is to install an additional
filter in cups (two files, see below) and change the filter in the
/etc/cups/mime.conv file:
#application/postscript  application/vnd.cups-postscript 66      pstops
application/postscript  application/vnd.cups-postscript 66      badpstops

This will make postscript files filter through the attached scripts, which
just move the psutils stuff down in the %%BeginSetup section. I do not
know if that makes the file perfect postscript, but at least cusp can print
it.

Since everyone thinks redhat should support people submitting patches, I
like to comment that it took me 2 days to fix this problem. 
Send the money to New Orleans.


File "/usr/lib/cups/filter/badpostscript":
#!/usr/bin/perl -w
# simple perl script to fix buggy postscript produced by psutils
# (c) GNU GPL, Peter Englmaier, 2005.

$mode=0;
$found=0;
my @saved;
while (<STDIN>) {
   if ($mode == 0) {
     if ($_ =~ "%%BeginProcSet: PStoPS.*") {
        $mode=1;
        $found=1;
        @saved = (@saved, $_);
     } else {
        print "$_";
     }
   } elsif ($mode == 1) {
     @saved = (@saved, $_);
     if ($_ =~ "%%EndProcSet.*") {
        $mode=2;
     }
   } else {
     print "$_";
     if ($found && $_ =~ "%%BeginSetup.*") {
       for (@saved) {
          print "$_";
      }
       $found=0; # do not repeat
     }
   }
}
exit 0

Wrapper script "/usr/lib/cups/badpstops"
#!/bin/sh
FILTERDIR=/usr/lib/cups/filter

# See if we have a filename on the command-line...
if test -z "$6"; then
   ifile="-"
else
        ifile="$6"
fi

cat $ifile | $FILTERDIR/badpostscript |
$FILTERDIR/pstops "$1" "$2" "$3" "$4" "$5"

Fedora Core 3 is now maintained by the Fedora Legacy project for security
updates only. If this problem is a security issue, please reopen and
reassign to the Fedora Legacy product. If it is not a security issue and
hasn't been resolved in the current FC5 updates or in the FC6 test
release, reopen and change the version to match.

Please try to reproduce this bug with latest version of psutils in FC5.
Otherwise I'm going to close this bug

FWW, here is what cupstestdsc (the CUPS utility for checking PostScript DSC
validity) says about the original gnumeric output:

foo.ps: FAIL
    Bad %%Page: on line 23!
        REF: Page 53, %%Page:
    Bad %%Page: on line 2289!
        REF: Page 53, %%Page:
    Missing or bad %%BoundingBox: comment!
        REF: Page 39, %%BoundingBox:
    Missing or bad %%Pages: comment!
        REF: Page 43, %%Pages:
    Missing or bad %%Page: comments!
        REF: Page 53, %%Page:
    Warning: obsolete DSC version 2.0 in file!
    Warning: no %%EndComments comment in file!

Closing this as NOTABUG seems just plain wrong.  What is the motivation behind
this change?