#!/usr/bin/perl
#
#  You may need to change the above path.
#
#-----------------------------------------------------------------------------
#
#  Copyright (C) 1996-1998 James Macnicol
#
#  This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by the
# Free Software Foundation; either version 2, or (at your option) any later
# version.
#
#  This program is distributed in the hope that it will be useful, but
# WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTIBILITY
# or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
# for more details.
#
#-----------------------------------------------------------------------------
#
# type1inst : Generate a "fonts.scale" file for Type 1 fonts in PFB format
# for use with your favourite X server.  Also generate a "Fontmap" for use
# with ghostscript.
#
# cd to the directory you want to install fonts in and invoke this script.
# Options:
#
#          -samples   Create sample PS files for each font
#          -nox       Do not create fonts.scale and fonts.dir for X11
#          -nogs      Do not create Fontmap for GhostScript
#          -quiet     Don't print anything on the stdout, just to the log
#                     (see also next section).
#          -silent    Same as -quiet (for backwards compatiblity)
#          -q         Same as -quiet
#          -nolog     Don't create a log file
#          -version   Print version info and quit
#          -v         Same as -version
#
#
#
# THIS IS BETA SOFTWARE!  PLEASE READ THE "README" FILE!!!
#
# Direct all correspondence regarding this software to 
#
#               J.Macnicol@student.anu.edu.au
#
#
# Good luck!
#
#
# James Macnicol
#
#-----------------------------------------------------------------------------

# Version number and date information

# NOTE THAT MY E-MAIL ADDRESS HAS CHANGED (AS OF VERSION 0.6.1) !!!!

$version = "0.6.1";
$versiondate = "11th February 1998";
$emailaddress = "james.macnicol\@mailexcite.com";
$copyright = "Copyright (C) 1996-1998 James Macnicol ($emailaddress)";

#
# Map identifying strings in /Notice into foundry names.  Separate identifier
# from name with a :.  Someone let me know if this is a problem (i.e. foundry
# has a : in its name which really ought to be there (although I may not
# believe it) ; we'll change it to ! or something.
#
# You probably want to put foundries which license type from others near the
# top of this list (e.g. Adobe).  If the name of the original source of the
# face is listed lower down then it will be used that instead.  It's just that
# Adobe does have its own faces too, but more often than not they are
# licensed.  Doing it this way will make it work out correctly in either case.
#


@foundries = (
	      "Adobe:adobe",
              "Allied Corporation:allied",
	      "Publishers' Paradise:paradise",
	      "PUBLISHERS' PARADISE:paradise",
	      "Bigelow & Holmes:b&h",
	      "Bitstream:bitstream",
	      "Corel Corporation:corel",
	      "International Typeface Corporation:itc",
	      "IBM:ibm",
	      "LETRASET:letraset",
	      "Monotype Corporation:monotype",
	      "SoftMaker:softmaker",
	      "URW:urw",
	      "Jonathan Brecher:brecher",
	      "Brendel Informatik:brendel",
	      "A. Carr:carr",
	      "FontBank:fontbank",
	      "Hershey:hershey",
	      "A.S.Meit:meit",
	      "Andrew s. Meit:meit",
	      "S.G. Moye:moye",
	      "S. G. Moye:moye",
	      "D. Rakowski:rakowski",
	      "David Rakowski:rakowski",
	      "Reasonable Solutions:reasonable",
	      "Southern Software:southern",
	      "Title Wave:titlewave",
	      "ZSoft:zsoft",
	      "Digiteyes Multimedia:digiteyes",
	      "MWSoft:mwsoft",
	      "MacroMind:macromind",
	      "Three Islands Press:3ip",
	      "Hank Gillette:gillette",
	      "Doug Miles:miles",
	      "Richard Mitchell:mitchell");

# Note: Hershey is the public Hershey fonts which come with Ghostscript.
# These cause no end of problems since they look inside like funny PS
# programs rather than standard fonts.  The current version of type1inst will
# refuse to process such fonts.  Older versions (< 0.6) tended to fall over
# when these were present.

# Note 2 : Some of these are obviously names of people only, not companies.
# They are generally public domain fonts.

# Note 3 : Publisher's Paradise did not produce a majority of the fonts that
# contain their name in the /Notice field, rather they distributed them on
# their BBS.  Unfortunately there is no other identifying info in these fonts.

#
# These are font weights.  Some are synonyms, e.g. regular for medium.  It
# has been suggested we map "thin" to "light", however there are some font
# families which have both "thin" and "light" variants.  An example is
# Linotype's Helvetica Neue.  Please let me know if you find a font where
# assuming "semi", and "demi" to be the same fails.
#

@weights = (
	    "book:book",
	    "demibold:demibold",
	    "semibold:demibold",
	    "demi:demibold",
	    "semi:demibold",
	    "extrabold:extrabold",
	    "boldface:bold",
 	    "bold:bold",
	    "heavyface:heavyface",
	    "heavy:heavy",
	    "ultrablack:ultrablack",
	    "extrablack:extrablack",
	    "ultra:ultra", # it's gonna break some widths...
	    "black:black",
	    "extralight:extralight",
 	    "light:light",
	    "thin:thin",
	    "super:super",
	    "thin:thin",
	    "light:light",
	    "semi:demi",
	    "bold:bold",
	    "heavy:heavy",
	    "black:black",
	    "normal:medium",
	    "regular:regular",
	    "roman:regular" # this too might break something...
	    );

#
# Likewise for slants
#

@slants = (
	   "italic:i",
	   "roman:r",
	   "regular:r",
	   # "it:i",
	   "cursive:i",
	   "kursiv:i",
	   "oblique:o",
	   "obl:o",
	   "slanted:o",
	   # Cyrillic fonts
	   "upright:r",
	   "inclined:i");

#
# Style.  Wondering if we should put "serif" in here somehow....?
#
# I haven't put "ultracondensed" here since I think they're two different
# things, i.e. Garamond Ultra Condensed is very bold but condensed.

@styles = (
	   "extracondensed:extracondensed",
	   "condensed:condensed",
	   "cond:condensed",
	   "sans:sans",
	   "wide:wide",
	   "cn:condensed",
	   "narrow:narrow",
	   "extracompressed:extracompressed",
	   "compressed:compressed",
	   "extraextended:extraextended",
	   "extended:extended",
	   "expanded:expanded",
	   "normal:normal");

#
# Additional styles. Refer to the line that puts together $xline.
#

@addstyles = ("alt:alternate",
	      "beginning:beginning",
	      "display:display",
	      "dfr:dfr",
	      "ending:ending",
	      # "exp" and "ep" seems to be sometimes part of a fonts name,
	      # sometimes part of additional classification. I'm crying... :-(
	      "ep:expert",
	      "exp:expert",
	      "ornaments:ornaments",
	      "osf:oldstylefigures",
	      "outline:outline",
	      "sc:smallcaps",
	      "shaded:shaded",
	      "shadowed:shadowed",
	      "stencil:stencil",
	      "swash:swash",
	      "sw:swash",
	      "one:one",
	      "two:two",
	      "three:three",
	      "four:four",
	      # Some fonts use just "a" to mean a font with alternate 
	      # character set. 
	      "a:alternate");

#
# Write a message to the stdout and/or the log file depending on what the
# user chose.
#

sub log_msg {
    ($msg) = @_;

    if (! $silent) {
	print STDOUT "$msg";
    }
    if ($dologfile) {
	print LOG "$msg";
    }
}

sub log_only_msg {
    ($msg) = @_;

    if ($dologfile) {
	print LOG "$msg";
    }
}

#
# Die with a bug message
#

sub die_bug {
    ($msg) = @_;

    die("BUG: $msg\nIf you have not modified the script in a way which might have\ncaused this error you are encouraged to report it as a bug to\n\n$emailaddress\n\n");
}

#
# Print out a string with a given minimum width.  This is used to make the
# Fontmap entries look nice.
#

sub print_min_width {
    ($stream, $minwidth, $string) = @_;
    $_ = $string;
    $strlength = length($string);
    # Print the string
    print $stream $string;
    # Now pad out the rest of the space if the string is short.
    if ($strlength < $minwidth) {
	for ($i = 0; $i < ($minwidth - $strlength); $i = $i + 1) {
	    print $stream " ";
	}
    }
}

#
# Indicate progress through the directory on the command line
#

sub print_progress {
  $totalfonts = $numpffonts + $numgsfonts + $badfonts;
  if (! $silent) {
      if (($totalfonts % 10) == 0) {
	  print "[$totalfonts]\n";
      }
  }
}

#
# Put the processing stuff into a procedure since we want to do the same for
# .pfb, .pfa and .gsf files (once .pfb's are decompressed).
#
# Argument : filename.
# Returns : X font description, name of font for Fontmap
#

sub process_font {
  ($fname) = @_;
  local($xline);

  # Check to see if this is a ghostscript font
  if ($fname =~ /\.gsf\s*$/) {
      $gsfont = 1;
  } else {
      $gsfont = 0;
  }

  # Default is not MultipleMaster
  $mm = 0;

  open(IN, $fname) || die "cannot open $file for reading";
  # An unlikely name to check to see we get a fontname out of the file.
  $fontname = "abcXYZ:!@#";
  $foundry = "unknown";
  $notice = "No notice given.";
  while(<IN>) {
      if (/\/isFixedPitch\s+(.+)\s+def\s*/) {
	  if ($1 =~ /true/) {
	      $fixedpitch = "m";
	  } else {
	      $fixedpitch = "p";
	  }
      }

      # I think that we should accept the manufacturers classification.
      # Try to extract this from FontName only if it's missing.
      # (It shouldn't. There are other reasons why this won't work, though.)
      if (/\/FamilyName\s*\((.+)\)\s+readonly\s+def\s*/) {
	  $familyname = $1;

	  # Convert to lower case (because case is insignificant).
	  # Spaces are acceptable according to XLFD.
	  $familyname =~ tr/A-Z/a-z/;
      }
      # Previous applies to this also... This might make xfontsels list a
      # a little cluttered, though. Perhaps it would be better to map it
      # to standard strings like you do. It's named $weight_add because
      # you already used $weight...
      if (/\/Weight\s*\((.+)\)\s+readonly\s+def\s*/) {
	  $weight_add = $1;

	  # Convert to lower case. Spaces are acceptable according to XLFD?
	  # Remove for consistency (as there would be any left after my
	  # slaughtering).
	  $weight_add =~ tr/A-Z/a-z/;
	  $weight_add =~ s/\s*//g;

	  # Remember if it's a MultipleMaster font
	  $mm = 1 if ($weight_add =~ /^all$/);
	  # Strange. This field seems to contain also width sometimes... remove it.
          $numstyles = @styles;
          for ($x = 0; $x < $numstyles; $x = $x + 1) {
              $ident = $styles[$x];
              @fields = split(/:/, $ident);
              $numfields = @fields;
              if ($numfields != 2) {
                  die_bug("The style identification \"$ident\" is bad\n");
              }
	      # Remove matched word from the font's name
              $weight_add =~ s/$fields[0]//;
          }
      }
      # FullName might contain useful information in determining
      # the properties of a font.
      if (/\/FullName\s*\((.+)\)\s+readonly\s+def\s*/) {
	  $fullname = $1;

	  # Convert to lower case
	  $fullname =~ tr/A-Z/a-z/;

	  # Some names got extra numerical information at the start.
	  $fullname =~ s/^\d*\s*(.+)/$1/;
      }
      # Note : some fonts have a suspect /FontName declaration where there
      # is no space between /FontName and the name of the font itself....
      if (/\/FontName\s*[\/\(]([^\)]+)\)?\s+def\s*/) {
	  $fontname = $1;

	  # Remove any embedded spaces
	  # (Probably unnecessary. If I remember it right, it can't contain any spaces,
	  # because it's a PostScript identifier/keyword or what's the right term...)
	  $fontname =~ s/\s//g;

	  # Save a copy of original full name for later
	  $fontnamecopy = $fontname;

	  # Convert to lower case
	  $fontname =~ tr/A-Z/a-z/;

	  # There are fonts like Mendoza Roman, Baskerville Book etc, where what
	  # looks like weight is part of the font's name, not it's weight.
	  # Split the name into fontname and fontstyle instead and handle them separate.
	  ($fontname, $fontstyle) = split(/-/, $fontname);

	  # Remove -s
	  $fontname =~ s/-//g;
	  $fontstyle =~ s/-//g;


	  # Check for weight modifiers (medium, bold, demi, light etc.)
	  $weight = "medium";
	  $numweights = @weights;
	  for ($x = 0; $x < $numweights; $x = $x + 1) {
	      $ident = $weights[$x];
	      @fields = split(/:/, $ident);
	      $numfields = @fields;
	      if ($numfields != 2) {
		  die_bug("The weight identification \"$ident\" is bad\n");
	      }
	      if ($fontstyle =~ /$fields[0]/) {
		  $weight = $fields[1];
	      } elsif ($weight_add) {
		  # Try any possible way
		  $weight = $weight_add;
	      }
	      # Remove matched word from the font's name
	      $fontstyle =~ s/$fields[0]//;
	  }

	  # Check for slant (italic, roman, oblique)
	  $slant = "r";

	  $numslants = @slants;
	  for ($x = 0; $x < $numslants; $x = $x + 1) {
	      $ident = $slants[$x];
	      @fields = split(/:/, $ident);
	      $numfields = @fields;
	      if ($numfields != 2) {
		  die_bug("The slant identification \"$ident\" is bad\n");
	      }
	      if ($fontstyle =~ /$fields[0]/) {
		  $slant = $fields[1];
	      }
	      # Remove matched word from the font's name
	      $fontstyle =~ s/$fields[0]//;
	  }

	  # Check for style (condensed, normal, sans, or wide)
	  $style = "normal";

	  $numstyles = @styles;
	  for ($x = 0; $x < $numstyles; $x = $x + 1) {
	      $ident = $styles[$x];
	      @fields = split(/:/, $ident);
	      $numfields = @fields;
	      if ($numfields != 2) {
	          die_bug("The style identification \"$ident\" is bad\n");
	      }
              if ($fontstyle =~ /$fields[0]/) {
	          $style = $fields[1];
	      }
	      # Remove matched word from the font's name
              $fontstyle =~ s/$fields[0]//;
	  }

	  # Check for additional styles (alternate, smallcaps, oldstylefigures etc.)
	  $addstyle = "";

          $numaddstyles = @addstyles;
          for ($x = 0; $x < $numaddstyles; $x = $x + 1) {
              $ident = $addstyles[$x];
              @fields = split(/:/, $ident);
              $numfields = @fields;
              if ($numfields != 2) {
                  die_bug("The additional style identification \"$ident\" is bad.\n");
              }
              if ($fontstyle =~ /$fields[0]/) {
                  $addstyle = $fields[1];
              }
              # Remove matched word from the font's name
              $fontstyle =~ s/$fields[0]//;
	  }
      }
      if (/^\/Encoding\s+(\S+)\s*/) {
	  if ($1 =~ /StandardEncoding/) {
	      $encoding = "iso8859-1";
	  } else {
	      # This needs work
	      $encoding = "adobe-fontspecific";
	  }
      }

      if (/^\s*\/Notice\s*(.*)$/) {
	  $notice = $1;

	  $notice =~ s/readonly def//g;

	  $numfoundries = @foundries;
	  for ($x = 0; $x < $numfoundries; $x = $x + 1) {
	      $ident = $foundries[$x];
	      @fields = split(/:/, $ident);
	      $numfields = @fields;
	      if ($numfields != 2) {
		  die_bug("The foundry identification \"$ident\" is bad.\n");
	      }
	      if ($notice =~ /$fields[0]/) {
		  $foundry = $fields[1];
	      }
	  }
      }

      # MultipleMaster fonts have this field.
      if (/\/BlendAxisTypes\s+\[([^\]]+)\]\s*def/) {
	  $axis = $1;
	  # Remove axises we don't need
	  $axis =~ s/\/Weight\s+//;
	  $axis =~ s/\/Width\s+//;
	  # Are there still some axises left?
	  if ($axis =~ /\//) {
	      # Remove trailing spaces
	      $axis =~ s/^(.*?)\s*$/$1/;
	      $axis =~ s/\/\S+/0/g;
	      $axis= "[$axis]";
	  }
      }

      # Break out of loop if we've passed the interesting stuff.
      # And time to try another way to find out the fontname.
      if ((! $gsfont) && (/currentfile\s+eexec/)) {
	  &try_another_way();
	  # This is for .pfa and .pfb fonts
	  last;
      } elsif (($gsfont) && (/currentdict\s+end/)) {
	  &try_another_way();
	  # This is for ghostscript .gsf fonts.  Why don't all these have a
	  # currentfile eexec ?
	  last;
      }
  }
  close(IN);

  # I use quite different mechanism to get fontname etc. However it's done,
  # the results are hard to get right. Should it be a command-line option?
  # Now I try both ways.

  # familyname, use fontname
  $familyname = ($anotherway ? $familyname : $fontname);

  # Oh, we are dealing with a MultipleMaster font...
  if ($mm) {
      $weight = "0";
      $style = "0";
      $addstyle .= $axis;
  }

  if ($familyname =~ /abcXYZ\:\!\@\#/) {
      log_only_msg("\n");
      log_only_msg("$filename : could not determine font name\n");
      log_only_msg("\n");
      $badfonts = $badfonts + 1;
      &print_progress();
      return;
  }

  if (($dox) && (! $gsfont) && ($foundry =~ /unknown/)) {
      $nofoundry = $nofoundry + 1;
      log_only_msg("\n");
      log_only_msg("$filename ($fontnamecopy) : foundry not matched\n");
      log_only_msg("  /Notice said : \"$notice\"\n");
      log_only_msg("\n");
  } elsif ($dox) {
#      log_only_msg("$filename ($fontnamecopy) : okay\n");
  }

  if (($dox) && (! $gsfont)) {
      # Addstyle is any extra information needed to uniquely identify a variation of a font
      # in it's family, like "alternate" (ACaslon-AltRegular) or "one" (EuropeanPi-One).
      # Changed fontname to familyname because it describes that field better, but that's
      # just my opinion...
      $xline = "-$foundry-$familyname-$weight-$slant-$style-$addstyle-0-0-0-0-$fixedpitch-0-$encoding";
  }

  # Update count of each type
  if ($gsfont) {
      $numgsfonts = $numgsfonts + 1;
  } else {
      $numpffonts = $numpffonts + 1;
  }

  &print_progress();

  ($xline, $fontnamecopy);
}


#
# An alternative way to get fontname
#

sub try_another_way {
    # Strip familyname from fullname. This seems to work most of time.
    # Some fontnames have extra numerical information after familyname.
    # Strip it if it's longer than two numbers.
    # Otherwise, it's probably part of additional style classification.
    # In a few cases it IS part of the name, and this algorithm should break.
    # Sometimes there's a strange string of *'s somewhere. Get rid of it.
    $fullname =~ s/\*//g;
    print STDERR "1: ${fullname}:\n" if $debug;
    if ($fullname =~ s/^$familyname\s*(\d\d+)?\s*(.*)/$2/) {
	# Wow. It worked. Let's continue and remove excess whitespace.
	$anotherway = 1;
	$fullname =~ s/\s+//g;
	
	# familyname can now stripped of -s
	$familyname =~ s/-//g; # Or space?
	print STDERR "2: ${fullname}:\n" if $debug;

	# Check for weight modifiers (medium, bold, demi, light etc.)
	$weight = "medium";
	$numweights = @weights;
	for ($x = 0; $x < $numweights; $x = $x + 1) {
	    $ident = $weights[$x];
	    @fields = split(/:/, $ident);
	    $numfields = @fields;
	    if ($numfields != 2) {
		die_bug("The weight identification \"$ident\" is bad\n");
	    }
	    if ($fullname =~ /$fields[0]/) {
		$weight = $fields[1];
		$weight =~ s/-//g;
	    }
	    # Remove matched word from the font's name
	    $fullname =~ s/$fields[0]//;
	}
	
    	print STDERR "3: ${fullname}:\n" if $debug;

	# Check for slant (italic, oblique)
	$slant = "r";

	$numslants = @slants;
	for ($x = 0; $x < $numslants; $x = $x + 1) {
	    $ident = $slants[$x];
	    @fields = split(/:/, $ident);
	    $numfields = @fields;
	    if ($numfields != 2) {
		die_bug("The slant identification \"$ident\" is bad\n");
	    }
	    if ($fullname =~ /$fields[0]/) {
		$slant = $fields[1];
		$slant =~ s/-//g;
	    }
	    # Remove matched word from the font's name
	    $fullname =~ s/$fields[0]//;
	}
	print STDERR "4: ${fullname}:\n" if $debug;
	# Check for style (normal or sans)
	$style = "normal";

	$numstyles = @styles;
	for ($x = 0; $x < $numstyles; $x = $x + 1) {
	    $ident = $styles[$x];
	    @fields = split(/:/, $ident);
	    $numfields = @fields;
	    if ($numfields != 2) {
		die_bug("The style identification \"$ident\" is bad\n");
	    }
	    if ($fullname =~ /$fields[0]/) {
		$style = $fields[1];
		$style =~ s/-//g;
	    }
	    # Remove matched word from the font's name
	    $fullname =~ s/$fields[0]//;
	}

	# What's left of fullname is probably additional style information.
	# Some fontnames have some strange numerical information here too.
	# If it's just one number, it usually refers to some variant of the
	# fontfamily, otherwise, just get rid of it.
	$fullname = "" if ($fullname =~ /^\d\d+$/);
	print STDERR "5: ${fullname}:\n" if $debug;
	$addstyle = $fullname;
	$addstyle =~ s/-//g;
    } else {
	$anotherway = 0;
    }
}

#
# Makes associative array out of current entries in fonts.scale
#

sub read_fonts_scale {
    local($finish, %rv, $line, $filename, $fontname);

    $finish = open(SCALE, "fonts.scale") ? 0 : 1;
    if ($finish == 1) {
	%rv;
    }

    log_only_msg("Reading fonts.scale ....");

    # First line should be an integer saying how many fonts there are.
    # Discard.
    $line = <SCALE>;
    if (! $line =~ /\s*[0-9]+\s*/) {
	log_only_msg("Warning : first line of fonts.scale is bad\n");
    }

    while (<SCALE>) {
	# Very rough pattern
	if (/\s*(\S+)\s+(.+)\s*/) {
	    chop;
	    $filename = $1;
	    $fontname = $2;
	    if (! -e $filename) {
		$numxremoved++;
		log_only_msg("Removed fonts.scale entry \"$_\" since the file did not exist\n");
		next;
	    }
	    if ($rv{$filename}) {
		$numxduplicates++;
		log_only_msg("Warning : fonts.scale already contains a line for file \"$filename\"\n");
		log_only_msg("  the line \"$_\" has been ignored\n");
	    } else {
		$rv{$filename} = $fontname;
	    }
	} else {
	    log_only_msg(" Couldn't understand line : \n");
	    log_only_msg("    \"$_\"\n");
	}
    }
    close(SCALE);

    log_only_msg("Done.\n");

    %rv;
}

#
# Write out an associative array into fonts.scale, making a backup copy
# first.
#

sub write_fonts_scale {
    (%fontdata) = @_;
    local($numentries, $key);

    # First, make backup copy
    if (-e "fonts.scale") {
	system ("cp -f fonts.scale fonts.scale.bak");
    }

    log_only_msg("Writing fonts.scale....\n");

    $numentries = keys(%fontdata);
    open(SCALE, ">fonts.scale") || die("Can't open fonts.scale!\n");
    print SCALE "$numentries\n";
    foreach $key (sort(keys %fontdata)) {
	print_min_width(SCALE, 12, $key);
	print SCALE " ";
	print SCALE "$fontdata{$key}\n";
    }
    close(SCALE);
    system ("chmod 0755 fonts.scale") && log_msg("Coudln't chmod \"fonts.scale\" ... continuing on anyway\n");

    log_only_msg(" Done.\n");
}

#
# Read the current Fontmap and return associative array with data.
#

sub read_fontmap {
    local(%rv, $finish, $fontname, $filename);

    $finish = open(FONTMAP, "Fontmap") ? 0 : 1;
    if ($finish) {
	%rv;
    }

    log_only_msg("Reading Fontmap ....\n");

    while (<FONTMAP>) {
	if (/\/+(\S+)\s+\((.*)\)\s+;\s+/) {
	    chop;
	    $fontname = $1;
	    $filename = $2;
	    if (! -e $filename) {
		$numgsremoved++;
		log_only_msg("Removed Fontmap entry \"$_\" since the file did not exist\n");
		next;
	    }
	    if ($rv{$filename}) {
		# Entry already exists
		$numgsduplicates++;
		log_only_msg("Warning : the Fontmap already contains a line for file \"$filename\"\n");
		log_only_msg("  the line \"$_\" has been ignored\n");
	    } else {
		$rv{$filename} = $fontname;
	    }
	} else {
	    $numgsbarf++;
	    log_only_msg("Couldn't understand line :\n");
	    log_only_msg("  $_\n");
	}
    }

    close(FONTMAP);

    log_only_msg("Done.\n");

    %rv;
}

#
# Write associative array containing font data to Fontmap
#

sub write_fontmap {
    (%fontdata) = @_;
    local($numentries, $key);

    # First, make backup copy
    if (-e "Fontmap") {
	system ("cp -f Fontmap Fontmap.bak");
    }

    log_only_msg("Writing Fontmap....");

    $numentries = keys(%fontdata);
    open(FONTMAP, ">Fontmap") || die("Couldn't open Fontmap!\n");
    foreach $key (sort(keys %fontdata)) {
	print_min_width(FONTMAP, 40, "/$fontdata{$key}");
	print FONTMAP " ";
	print FONTMAP "($key)\t;\n";
    }
    close(FONTMAP);
    system ("chmod 0755 Fontmap") && log_msg("Couldn't chmod \"Fontmap\" ... continuing on anyway\n");

    log_only_msg(" Done.\n");
}

#
# Add a font (either X or gs) to hash table 
#

sub add_font_to_aarray {
    ($fname, $text, %aa) = @_;

    if (($text =~ /^\s*$/) || ($fname =~ /^\s*$/)) {
	# This will occur if the font is a dud (e.g. a Hershey font).  We
	# assume that $badfonts has been incremented and we just return.
	%aa;
    }

    if (! $aa{$fname}) {
	$aa{$fname} = $text;
    }

    %aa;
}

#
# Create sample text using each font
#

sub font_sample {
    ($filename, $fontname, $height) = @_;
    local($text, $alltext, $samplefile);

    if (($filename =~ /^\s*$/) || ($fontname =~ /^s*$/)) {
	print "font_sample: $filename, $fontname\n";
	die_bug("Bad argument(s) to font_sample()!\n");
    }

# Here we create a full page sample for the current font.  It contains
# a large point-size version, a normal sized version, and a small version.

    $text = <<"TEXT";
%!
%%EndComments
/$samplefont findfont 
18 scalefont
setfont
newpath
200 715 moveto
(File : $filename) show
200 695 moveto
(Font Name : $fontname) show
% t1embed : $filename $fontname
closepath

/$fontname findfont 
60 scalefont        
setfont             
newpath        
40 640 moveto
(ABCDE) show
40 575 moveto
(FGHIJK) show
40 510 moveto
(LMNOP) show
40 445 moveto
(QRSTU) show
40 380 moveto
(VWXYZ) show
40 305 moveto
(abcdefghijklm) show
40 240 moveto
(nopqrstuvwxyz) show
40 175 moveto           
(1234567890) show
closepath

/$fontname findfont 
12 scalefont        
setfont             
newpath        
50 148 moveto
(A B C D E F G H I J K L M N O P Q R S T U V W X Y Z) show
50 132 moveto
(a b c d e f g h i j k l m n o p q r s t u v w x y z) show
50 116 moveto           
(1 2 3 4 5 6 7 8 9 0 \! \$ \% \& \\\( \\\) \; \: \< \> ) show
closepath

/$fontname findfont 
4 scalefont        
setfont             
newpath        
50 99 moveto
(A B C D E F G H I J K L M N O P Q R S T U V W X Y Z) show
50 93 moveto
(a b c d e f g h i j k l m n o p q r s t u v w x y z) show
50 87 moveto           
(1 2 3 4 5 6 7 8 9 0 \! \$ \% \& \\\( \\\) \; \: \< \> ) show
closepath
showpage
TEXT

    $samplefile = $fontname . ".ps";
    open(SAMPLE, ">samples/$samplefile") || 
    die("Couldn't open samples/$samplefile\n");
    print SAMPLE "$text\n";
    close(SAMPLE);
    system("chmod 0755 samples/$samplefile") && log_msg("Couldn't chmod individual sample file \"samples/$samplefile\" ... continuing on anyway\n");

# For the "allfont.ps" files we use a standard font for the font name so
# that in the case of non-alpha fonts we still can still read the name of
# the font (eg symbol or dingbats).

    if ($height == 700) {
	$allsample = "samples/allfont-$allcount.ps";
	$allcount = $allcount + 1;

	log_only_msg("Creating new sample file \"$allsample\"....");
	open(ALLSAMPLE, ">$allsample") || 
	die("Couldn't open $allsample\n");
	log_only_msg("done\n");
	print ALLSAMPLE "%!\n";
	print ALLSAMPLE "%%EndComments\n";
    }

    $alltext = <<"ALLTEXT";

% t1embed : $filename $fontname
/$samplefont findfont
12 scalefont
setfont
newpath
30 $height moveto
($fontname : ) show
/$fontname findfont 
20 scalefont        
setfont             
(AbCdEfGhIjKlMnOpQrStUvWxYz 0123456789) show
closepath
ALLTEXT

    print ALLSAMPLE "$alltext\n";
    $height = $height - 32;
    if ($height < 100) {
	print ALLSAMPLE "showpage\n";
	close(ALLSAMPLE);
	system("chmod 0755 $allsample") && log_msg("Couldn't chmod all sample sheet \"$allsample\" ... continuing on anyway\n");
	$height = 700;
    }

    ($height);
}

#
# Some users have had problems with perl's file globbing not working.  This
# gets a shell to do it for us.  It matches all files with the extension
# specified in the parameter, i.e. if pat = "foo" then it matches all of
# *.foo .
#

sub do_glob {
    ($pat) = @_;
    local($raw, @fnames);
    open(SHELL, "echo *.$pat|") || die("Couldn't open shell in do_glob\n");
    $raw = <SHELL>;
    $raw =~ s/\*\.$pat//;
    @fnames = split(/\s/,$raw);
    close(SHELL);
    (@fnames);
}

# ------------------------------------------------------------------------
# Start of program proper
# ------------------------------------------------------------------------

# Process command line arguments
$workdir = 0;
$dox = 1;
$dogs = 1;
$silent = 0;
$samples = 0;
$dologfile = 1;
@argvcopy = (@ARGV);
$numargs = @ARGV;
for ($x = 0; $x < $numargs; $x = $x + 1) {
    $arg = $ARGV[$x];
    if ($arg =~ /-nox/) {
	$dox = 0;
    } elsif ($arg =~ /-nogs/) {
	$dogs = 0;
    } elsif ($arg =~ /-silent/) {
	$silent = 1;
    } elsif ($arg =~ /-quiet/) {
	$silent = 1;
    } elsif ($arg =~ /-q/) {
	$silent = 1;
    } elsif ($arg =~ /-samples/) {
	$samples = 1;
    } elsif ($arg =~ /-nolog/) {
	$dologfile = 0;
    } elsif ($arg =~ /-d/) {
	$x++;
	$workdir = $ARGV[$x];
    } elsif ($arg =~ /-version/) {
	die("type1inst version $version ($versiondate)\n$copyright\n");
    } elsif ($arg =~ /-v/) {
	die("type1inst version $version ($versiondate)\n$copyright\n");
    } else {
	die("Usage: $0 [-silent] [-quiet] [-q] [-nox] [-nogs] [-samples] [-version] [-v]\n");
    }
}
if ((! $dox) && (! $dogs) && (! $samples)) {
    die("$0: Nothing to do!\n");
}

if ($workdir) {
  chdir $workdir || die "Cannot change to \"$workdir\"";
}

# Open logfile
if ($dologfile) {
    open(LOG, ">type1inst.log") || die "Cannot open log file \"type1inst.log\"";
}

log_only_msg("type1inst Version $version ($versiondate)\n");
log_only_msg("$copyright\n\n");
open (DATE, "date|") || die("Couldn't run \"date\"\n");
$currenttime = <DATE>;
log_only_msg("Run started at $currenttime\n");
close(DATE);

# Setup directory for font samples
if ($samples) {
    if (! -e "samples") {
	# Create directory for sample text PS files
	log_only_msg("Creating directory for samples ...\n");
	system("mkdir samples");
	system("chmod 0755 samples") && log_msg("Coudln't chmod \"samples\" directory\n");
	
    } elsif (-f "samples") {
	die("$0: remove file \"samples\" or do not use -samples option\n");
    } else {
	log_msg("Clearing samples directory\n");
	system("rm -f samples/*.ps");
    }
    $height = 700;
    $samplefont = "Helvetica";
    $allcount = 0;
    $allsample = "samples/allfont-$allcount.ps";
    log_only_msg("Creating new sample file \"$allsample\"....");
    open(ALLSAMPLE, ">$allsample") || die("Couldn't open all sample file \"$allsample\"\n");
    log_only_msg("done\n");
    print ALLSAMPLE "%!\n";
    print ALLSAMPLE "%%EndComments\n";
}


# Counts how many fonts we come across
$numpffonts = 0;
$numgsfonts = 0;
$nofoundry = 0;
$badfonts = 0;
$numskipped = 0;
$numxremoved = 0;
$numgsremoved = 0;
$numxduplicates = 0;
$numgsduplicates = 0;
$numxbarf = 0;
$numgsbarf = 0;

if (! $silent) {
    print "type1inst Version $version ($versiondate)\n";
    print "$copyright\n\n";
}

$totalfonts = 0;
foreach $filename (do_glob("pfa")) {
    $totalfonts++;
}
foreach $filename (do_glob("pfb")) {
    $totalfonts++;
}
foreach $filename (do_glob("pfa.gz")) {
    $totalfonts++;
}
foreach $filename (do_glob("pfb.gz")) {
    $totalfonts++;
}
foreach $filename (do_glob("gsf")) {
    $totalfonts++;
}
if (! $silent) {
    if ($totalfonts == 0) {
	die("There are no PostScript fonts in this directory\n");
    } elsif ($totalfonts == 1) {
	print "There is 1 PostScript font in this directory\n";
    } else {
	print "There are a total of $totalfonts PostScript fonts in this directory\n";
    }
}

if ($dox) {
    %fs = &read_fonts_scale();
}
if (($dogs) || ($samples)) {
    %fm = &read_fontmap();
}

# Process ASCII PS fonts
foreach $filename (do_glob("pfa")) {
    if (($dox && (! $fs{$filename})) ||
	(($dogs || $samples) && (! $fm{$filename}))) {
	($x, $gs) = &process_font($filename);
	if ($dox) {
	    %fs = &add_font_to_aarray($filename, $x, %fs);
	}
	if (($dogs) || ($samples)) {
            %fm = &add_font_to_aarray($filename, $gs, %fm);
	}
    } else {
	$numpffonts = $numpffonts + 1;
	$numskipped = $numskipped + 1;
	&print_progress();
    }
    if ($samples) {
	($height) = &font_sample($filename, $fm{$filename}, $height);
    }
}

# Process binary PS fonts
foreach $filename (do_glob("pfb")) {
    if (($dox && (! $fs{$filename})) ||
	(($dogs || $samples) && (! $fm{$filename}))) {
	system("pfbtops $filename > foo");
	($x, $gs) = &process_font("foo");
	system("rm foo");
	if ($dox) {
	    %fs = &add_font_to_aarray($filename, $x, %fs);
	}
	if ($dogs || $samples) {
	    %fm = &add_font_to_aarray($filename, $gs, %fm);
	}
    } else {
	$numpffonts = $numpffonts + 1;
	$numskipped = $numskipped + 1;
	&print_progress();
    }
    if ($samples) {
	($height) = &font_sample($filename, $fm{$filename}, $height);
    }
}

# Process binary PS fonts
foreach $filename (do_glob("pfa.gz")) {
    if (($dox && (! $fs{$filename})) ||
	(($dogs || $samples) && (! $fm{$filename}))) {
	system("gunzip -c $filename > foo");
	($x, $gs) = &process_font("foo");
	system("rm foo");
	if ($dox) {
	    %fs = &add_font_to_aarray($filename, $x, %fs);
	}
	if ($dogs || $samples) {
	    %fm = &add_font_to_aarray($filename, $gs, %fm);
	}
    } else {
	$numpffonts = $numpffonts + 1;
	$numskipped = $numskipped + 1;
	&print_progress();
    }
    if ($samples) {
	($height) = &font_sample($filename, $fm{$filename}, $height);
    }
}

# Process binary PS fonts
foreach $filename (do_glob("pfb.gz")) {
    if (($dox && (! $fs{$filename})) ||
	(($dogs || $samples) && (! $fm{$filename}))) {
	system("gunzip -c $filename | pfbtops > foo");
	($x, $gs) = &process_font("foo");
	system("rm foo");
	if ($dox) {
	    %fs = &add_font_to_aarray($filename, $x, %fs);
	}
	if ($dogs || $samples) {
	    %fm = &add_font_to_aarray($filename, $gs, %fm);
	}
    } else {
	$numpffonts = $numpffonts + 1;
	$numskipped = $numskipped + 1;
	&print_progress();
    }
    if ($samples) {
	($height) = &font_sample($filename, $fm{$filename}, $height);
    }
}

# Process Ghostscript fonts
if ($dogs || $samples) {
    foreach $filename (do_glob("gsf")) {
	if (! $fm{$filename}) {
	    ($x, $gs) = &process_font($filename);
	    %fm = &add_font_to_aarray($filename, $gs, %fm);
	} else {
	    $numgsfonts = $numgsfonts + 1;
	    $numskipped = $numskipped + 1;
	    &print_progress();
	}
	if ($samples) {
	    ($height) = &font_sample($filename, $fm{$filename}, $height);
	}
    }
}

if ($dox) {
    &write_fonts_scale(%fs);
    system("mkfontdir");      # Generate fonts.dir
    system("chmod 0755 fonts.dir") && log_msg("Couldn't chmod \"fonts.dir\" ... continuing on anyway\n");
}
if ($dogs) {
    &write_fontmap(%fm);
}

# Finish up the all font sample file
if ($samples) {
    log_only_msg("Finished font sample files\n");
    if ($height < 700) {
	print ALLSAMPLE "showpage\n";
	close(ALLSAMPLE);
	system("chmod 0755 $allsample") && log_msg("Couldn't chmod \"$allsample\" ... continuing on anyway\n");
    }
}

# Report
if (! $silent) {
    $totalfonts = $numpffonts + $numgsfonts + $badfonts;

    # List statistics
    print "-------------------------------------------------------\n";
    if ($totalfonts == 0) {
	print "No fonts were found in this directory\n";
    } elsif ($totalfonts == 1) {
	print "1 font was found in this directory\n";
    } else {
	print "$totalfonts fonts found\n";
    }
    if ($numpffonts == 1) {
	print "1 was a PostScript font\n";
    } elsif ($numpffonts > 1) {
	print "$numpffonts were standard PostScript fonts\n";
    }
    if ($numgsfonts == 1) {
	print "1 was a Ghostscript font\n";
    } elsif ($numgsfonts > 1) {
	print "$numgsfonts were Ghostscript fonts\n";
    }
    if ($numskipped == 1) {
	print "\n";
	print "I skipped one of these fonts because it already had\n";
	print "an overriding entry in both fonts.scale and/or Fontmap\n";
	print "(X Windows font or Ghostscript font respectively).\n";
    } elsif ($numskipped > 1) {
	print "\n";
	print "I skipped $numskipped of these fonts because they already\n";
	print "had overriding entries in both fonts.scale and/or Fontmap\n";
	print "(X Windows fonts or Ghostscript fonts respectively).\n";
    }

    # Print error messages
    $wereerrors = 0;
    if ($badfonts > 0) {
	$wereerrors = 1;
	print "-------------------------------------------------------\n";
	if ($badfonts == 1) {
	    print "I couldn't extract a font name for 1 font in\n";
	} else {
	    print "I couldn't extract font names for $ badfonts fonts in\n";
	}
	print "this directory.  This means the font file had a non-standard\n";
	print "format which this program doesn't know about or cannot do\n";
	print "anything with.  Check the README file to find out more.\n";
    }
    if ($dox) {
	if ($nofoundry > 0) {
	    $wereerrors = 1;
	    print "-------------------------------------------------------\n";
	    print "For $nofoundry of these I couldn't figure out which foundry\n";
	    print "the font is from.  Thus, these fonts will appear under the\n";
	    print "foundry unknown, i.e. X font name -unknown-*.\n";
	    print "Please consult the README file to see what this means.\n";
	}

	if ($numxremoved > 0) {
	    $wereerrors = 1;
	    print "-------------------------------------------------------\n";
	    if ($numxremoved == 1) {
		print "While reading the existing fonts.scale file I saw 1 entry\n";
	    } else {
		print "While reading the existing fonts.scale file I saw $numxremoved entries\n";
	    }
	    print "which mentioned a filename which now does not exist.  Most likely\n";
	    print "you removed or renamed the file.  I ignored these entries.\n";
	}
	if ($numxbarf > 0) {
	    $wereerrors = 1;
	    print "-------------------------------------------------------\n";
	    if ($numxbarf == 1) {
		print "There was a line in fonts.scale I couldn't understand.\n";
	    } else {
		print "There were $numxbarf lines in fonts.scale which I couldn't understand\n";
	    }
	    print "These were ignored.\n";
	}
    }
    if ($dogs) {
	if ($numgsremoved > 0) {
	    $wereerrors = 1;
	    print "-------------------------------------------------------\n";
	    if ($numgsremoved == 1) {
		print "While reading the existing Fontmap file I saw 1 entry\n";
	    } else {
		print "While reading the existing Fontmap file I saw $numgsremoved entries\n";
	    }
	    print "which mentioned a filename which now does not exist.  Most likely\n";
	    print "you removed or renamed the file.  I ignored these entries.\n";
	}
	if ($numgsbarf > 0) {
	    $wereerrors = 1;
	    print "-------------------------------------------------------\n";
	    if ($numgsbarf == 1) {
		print "There was a line in Fontmap I couldn't understand.\n";
	    } else {
		print "There were $numgsbarf lines in Fontmap which I couldn't understand\n";
	    }
	    print "These were ignored.\n";
	}
    }

    if ($wereerrors) {
	print "-------------------------------------------------------\n";
	print "\n";
	print "A log of errors is located in the file \"type1inst.log\"\n";
	print "\n";
    }
}