Updated the documentation building tools to work with a recent linuxdoc-tools package.

Note that this is (and always was) a terrible hack and we really should
replace it with something reasonable which wouldn't need changing every
time linuxdoc-tools evolve.

I also needed to include a patched version of LinuxDocTools.pm, because the
original one explicitly refused to work with a non-linuxdoc DTD. The authors
of linuxdoc recommend to use sgmltools-lite in such cases, but it would mean
rewritting our formatting rules to the DSSSL language which I don't dare to
speak about here :)

3 files changed, 541 insertions, 63 deletions

diff --git a/doc/sbase/dist/fmt_html.pl b/doc/sbase/dist/fmt_html.pl
index f42e631..b372a12 100644
--- a/doc/sbase/dist/fmt_html.pl
+++ b/doc/sbase/dist/fmt_html.pl
@@ -7,17 +7,17 @@
 #
 #  © Copyright 1996, Cees de Groot
 #
-package SGMLTools::fmt_html;
+package LinuxDocTools::fmt_html;
 use strict;
 
-use SGMLTools::CharEnts;
-use SGMLTools::Vars;
+use LinuxDocTools::CharEnts;
+use LinuxDocTools::Vars;
 
-use SGMLTools::FixRef;
-my $fixref = $SGMLTools::FixRef::fixref;
+use LinuxDocTools::FixRef;
+my $fixref = $LinuxDocTools::FixRef::fixref;
 
-use SGMLTools::Html2Html;
-my $html2html = $SGMLTools::Html2Html::html2html;
+use LinuxDocTools::Html2Html;
+my $html2html = $LinuxDocTools::Html2Html::html2html;
 
 my $html = {};
 $html->{NAME} = "html";
@@ -25,14 +25,22 @@ $html->{HELP} = "";
 $html->{OPTIONS} = [
    { option => "split", type => "l", 
      'values' => [ "0", "1", "2" ], short => "s" },
+   { option => "toc", type => "l", 
+     'values' => [ "0", "1", "2" ], short => "T" },
    { option => "dosnames", type => "f", short => "h" },
-   { option => "imagebuttons", type => "f", short => "I"}
+   { option => "imagebuttons", type => "f", short => "I"},
+   { option => "header", type => "s", short => "H"},
+   { option => "footer", type => "s", short => "F"}
 ];
 $html->{'split'}  = 1;
+$html->{'toc'}  = -1;
 $html->{dosnames}  = 0;
 $html->{imagebuttons}  = 0;
+$html->{header}  = "";
+$html->{footer}  = "";
 $html->{preNSGMLS} = sub {
   $global->{NsgmlsOpts} .= " -ifmthtml ";
+  $global->{NsgmlsPrePipe} = "cat $global->{file}";
 };
 
 $Formats{$html->{NAME}} = $html;
@@ -132,11 +140,15 @@ $html->{postASP} = sub
   #  Run through html2html, preserving stdout
   #  Also, handle prehtml.sed's tasks
   #
+  my $filter = "";
+#  $filter = "|$main::progs->{NKF} -e" if ($global->{language} eq "ja");
   open SAVEOUT, ">&STDOUT";
-  open STDOUT, ">$filename.$ext" or die qq(Cannot open "$filename.$ext");
+  open STDOUT, "$filter>$filename.$ext" or die qq(Cannot open "$filename.$ext");
 
   &{$html2html->{init}}($html->{'split'}, $ext, $img, $filename,
-                        $fixref->{filenum}, $fixref->{lrec});
+                        $fixref->{filenum}, $fixref->{lrec},
+			$html->{'header'}, $html->{'footer'}, $html->{'toc'},
+                        $global->{tmpbase}, $global->{debug});
   LINE: foreach (@file) {
       s,<P></P>,,g; 			# remove empty <P></P> containers
       foreach my $pat (keys %{$html2html->{rules}}) {
diff --git a/doc/sbase/dist/fmt_latex2e.pl b/doc/sbase/dist/fmt_latex2e.pl
index 03b07f5..1f12174 100644
--- a/doc/sbase/dist/fmt_latex2e.pl
+++ b/doc/sbase/dist/fmt_latex2e.pl
@@ -7,12 +7,15 @@
 #
 #  © Copyright 1996, Cees de Groot
 #
-package SGMLTools::fmt_latex2e;
+#  Support for PDF files: added by Juan Jose Amor, January 2000
+#  © Copyright 2000, Juan Jose Amor
+#
+package LinuxDocTools::fmt_latex2e;
 use strict;
 
-use SGMLTools::CharEnts;
-use SGMLTools::Vars;
-use SGMLTools::Lang;
+use LinuxDocTools::CharEnts;
+use LinuxDocTools::Vars;
+use LinuxDocTools::Lang;
 
 use File::Copy;
 
@@ -24,23 +27,48 @@ $latex2e->{HELP} = <<EOF;
 EOF
 $latex2e->{OPTIONS} = [
    { option => "output", type => "l", 
-     'values' => [ "dvi", "tex", "ps" ], short => "o" },
+     'values' => [ "dvi", "tex", "ps", "pdf" ], short => "o" },
    { option => "bibtex", type => "f",  short => "b" },
    { option => "makeindex", type => "f",  short => "m" },
    { option => "pagenumber", type => "i", short => "n" },
-   { option => "quick",  type => "f",  short => "q" }
+   { option => "quick",  type => "f",  short => "q" },
+   { option => "dvips",  type => "l",
+     'values' => [ "dvips", "dvi2ps", "jdvi2kps" ], short => "s" },
+   { option => "latex",  type => "l",  
+     'values' => [ "latex", "hlatexp", "platex", "jlatex" ], short => "x" }
 ];
-$latex2e->{output} = "dvi";
+$latex2e->{output} = "tex";
 $latex2e->{pagenumber} = 1;
 $latex2e->{quick}  = 0;
 $latex2e->{bibtex}  = 0;
 $latex2e->{makeindex} = 0;
+$latex2e->{latex} = "unknown";
+$latex2e->{dvips} = "unknown";
+$Formats{$latex2e->{NAME}} = $latex2e;
+
 $latex2e->{preNSGMLS} = sub {
   $global->{NsgmlsOpts} .= " -ifmttex ";
-};
 
-$Formats{$latex2e->{NAME}} = $latex2e;
+  #  for Japanese jlatex users
+  if ($global->{language} eq "ja" && $latex2e->{latex} eq "unknown") {
+  	$latex2e->{latex} = "jlatex";
+	$latex2e->{dvips} = "dvi2ps";
+  # for Japanese platex users
+  #	$latex2e->{latex} = "platex";
+  #	$latex2e->{dvips} = "dvips";
+  }
+
+  # for Korean users
+  if ($global->{language} eq "ko" && $latex2e->{latex} eq "unknown") {
+  	$latex2e->{latex} = "hlatexp";
+  }
 
+  # default process command
+  $latex2e->{latex} = "latex" if ($latex2e->{latex} eq "unknown");
+  $latex2e->{dvips} = "dvips" if ($latex2e->{dvips} eq "unknown");
+
+  $global->{NsgmlsPrePipe} = "cat $global->{file} ";
+};
 
 # extra `\\' here for standard `nsgmls' output
 my %latex2e_escapes;
@@ -59,6 +87,7 @@ $latex2e_escapes{'<'} = '{$<$}';	# wouldn't happen, but that's what'd be
 $latex2e_escapes{'|'} = '{$|$}';
 
 my $in_verb;
+my $remove_comment; # added 2000 Jan 25 by t.sano
 
 # passed to `parse_data' below in latex2e_preASP
 my $latex2e_escape = sub {
@@ -95,13 +124,22 @@ $latex2e->{preASP} = sub
   # a VERB or CODE environment or not
   $in_verb = 0;
 
+  # switch to remove empty line from TeX source or not, depending
+  # on whether we're in a HEADING or ABSTRACT environment or not
+  $remove_comment = 0;
+
   while (<$infile>)
     {
       if (/^-/)
         {
 	    my ($str) = $';
 	    chop ($str);
-	    print $outfile "-" . parse_data ($str, $char_maps, $latex2e_escape) . "\n";
+	    $_ = parse_data ($str, $char_maps, $latex2e_escape);
+	    if ($remove_comment)
+	      {
+		s/(\s+\\n)+//;
+	      }
+	    print $outfile "-" . $_ . "\n";
         }
       elsif (/^A/)
         {
@@ -111,9 +149,12 @@ $latex2e->{preASP} = sub
 	  if ($type eq "CDATA")
 	    {
 	      # CDATA attributes get translated also
-	      if ($name eq "URL" or $name eq "ID")
+	      if ($name eq "URL" or $name eq "ID" or $name eq "CA")
 	        {
 		  # URL for url.sty is a kind of verbatim...
+		  # CA is used in "tabular" element.
+		  # Thanks to Evgeny Stambulchik, he posted this fix
+		  # on sgml-tools list. 2000 May 17, t.sano
 		  my $old_verb = $in_verb;
 		  $in_verb = 1;
 		  $value = parse_data ($value, $ascii_char_maps, 
@@ -141,6 +182,19 @@ $latex2e->{preASP} = sub
 	  $in_verb = 0;
 	  $char_maps = $tex_char_maps;
 	}
+      elsif (/^\((HEADING|ABSTRACT)/)
+        {
+	  print $outfile $_;
+	  # empty lines (comment in sgml source) do harm 
+          # in HEADING or ABSTRACT
+	  $remove_comment = 1;
+	}
+      elsif (/^\)(HEADING|ABSTRACT)/)
+        {
+	  print $outfile $_;
+	  # leaving HEADING or ABSTRACT section
+	  $remove_comment = 0;
+	}
       else
         {
 	  print $outfile $_;
@@ -148,6 +202,24 @@ $latex2e->{preASP} = sub
     }
 };
 
+# return the string of the name of the macro for urldef
+sub latex2e_defnam($)
+{
+    my ($num) = @_;
+
+    if ($num > 26*26*26) {
+       die "Too many URLs!\n";
+    }
+
+    my $anum = ord("a");
+
+    my $defnam = chr ($anum + ($num / 26 / 26)) .
+                 chr ($anum + ($num / 26 % 26)) .
+                 chr ($anum + ($num % 26));
+
+    return ($defnam);
+};
+
 #
 #  Take the sgmlsasp output, and make something
 #  useful from it.
@@ -156,29 +228,117 @@ $latex2e->{postASP} = sub
 {
   my $infile = shift;
   my $filename = $global->{filename};
-  $ENV{TEXINPUTS} .= ":$main::LibDir";
+  my $tmplatexdir = $global->{tmpbase} . "-latex-" . $$ . ".dir";
+  my $tmplatexnam = $tmplatexdir . "/" . $filename;
+  my @epsfiles = ();
+  my @texlines = ();
+  my @urldefines = ();
+  my @urlnames = ();
+  my $urlnum = 0;
+  my $tmpepsf;
+  my $saved_umask = umask;
+  $ENV{TEXINPUTS} .= ":$main::DataDir";
+
+  umask 0077;
+  mkdir ($tmplatexdir, 0700) || return -1;
 
   #
-  #  Set the correct \documentclass options. The if statement is just
-  #  a small optimization.
+  # check epsfile is specified in source file
+  # check nameurl specified in source file
+  #
+  {
+      my $epsf;
+      open SGMLFILE, "<$filename.sgml";
+      while (<SGMLFILE>)
+        {
+	# for epsfile
+           if ( s/^\s*<eps\s+file=(.*)>/$1/ )
+             {
+               s/\s+angle=.*//;
+               s/\s+height=.*//;
+               s/\"//g;
+               $epsf = $_;
+               chop ( $epsf );
+               push @epsfiles, $epsf;
+             }
+           if ($latex2e->{output} eq "pdf")
+             {
+	       if ( s/^\s*<img\s+src=(.*)>/$1/ )
+                 {
+                   s/\"//g;
+                   $epsf = $_;
+                   chop ( $epsf );
+                   push @epsfiles, $epsf;
+                 }
+             }
+        }
+      close SGMLFILE;
+  }
+  {
+      my $urlid;
+      my $urlnam;
+      my $urldef;
+      while (<$infile>)
+	{
+	   push @texlines, $_;
+	# for nameurl
+	   if ( /\\nameurl/ )
+	     {
+		($urlid, $urlnam) = ($_ =~ /\\nameurl{(.*)}{(.*)}/);
+		print $urlnum . ": " . $urlid . "\n" if ( $global->{debug} );
+
+		$urldef = latex2e_defnam($urlnum) . "url";
+		s/\\nameurl{.*}{.*}/{\\em $urlnam} {\\tt \\$urldef}/;
+		push @urlnames, $_;
+		push @urldefines, "\\urldef{\\$urldef} \\url{$urlid}\n";
+		$urlnum++;
+	     }
+        }
+      close $infile;
+  }
+
+  open OUTFILE, ">$tmplatexnam.tex";
+  #
+  #  Set the correct \documentclass options.
   #
-  if ($global->{language} ne "en" ||
-      $global->{papersize} ne "a4" ||
-      $latex2e->{pagenumber} != 1 ||
-      $global->{pass} ne "" ||
-      $latex2e->{makeindex})
     {
       my $langlit = ISO2English ($global->{language});
       $langlit = ($langlit eq 'english') ? "" : ",$langlit"; 
       my $replace = $global->{papersize} . 'paper' .   $langlit;
-      open OUTFILE, ">$filename.tex";
-      while (<$infile>)
+      my $hlatexopt = "";
+      $global->{charset} = "nippon" if ($global->{language} eq "ja");
+      $global->{charset} = "euc-kr" if ($global->{language} eq "ko");
+      $replace = $global->{papersize} . 'paper' if ($global->{charset} eq "nippon") || ($global->{charset} eq "euc-kr");
+      while (defined($texlines[0]))
         {
+	  $_ = shift @texlines;
           if (/^\\documentclass/) 
 	    {
-	      s/\\documentclass\[.*\]/\\documentclass\[$replace\]/;
+	      if ($global->{language} ne "en" ||
+		  $global->{papersize} ne "a4")
+		{
+		   s/\\documentclass\[.*\]/\\documentclass\[$replace\]/;
+		}
+	      if ($global->{charset} eq "nippon") {
+		 if ($latex2e->{latex} eq "platex") {
+		   s/{article}/{jarticle}/;
+		 } elsif ($latex2e->{latex} eq "jlatex") {
+		   s/{article}/{j-article}/;
+		 }
+	      }
               $_ = $_ . "\\makeindex\n" if ($latex2e->{makeindex});
-            }  
+            }
+          if (/^\\usepackage.epsfig/ && ($global->{charset} eq "euc-kr"))
+            {
+              $hlatexopt = "[noautojosa]" if ($latex2e->{latex} eq "hlatexp");
+              $_ = $_ . "\\usepackage" . "$hlatexopt" . "{hangul}\n"
+            }
+          if ((/\\usepackage.t1enc/) &&
+		 (($global->{charset} eq "nippon") || 
+		  ($global->{charset} eq "euc-kr")))
+            {
+		s/^/%%/;
+            }
           if (/%end-preamble/)
 	    {
 	      if ($latex2e->{pagenumber}) 
@@ -192,20 +352,45 @@ $latex2e->{postASP} = sub
 	        }
 	      $_ = $_ . $global->{pass} . "\n" if ($global->{pass});
 	    }
+       	  if (/\\nameurl/ && $latex2e->{output} ne "pdf")
+            {
+		$_ = shift @urlnames;
+	    }
 	  print OUTFILE;
+          if (/%end-preamble/)
+	    {
+		if ($urlnum && $latex2e->{output} ne "pdf")
+		  {
+			while (defined($urldefines[0]))
+			  {
+				$_ = shift @urldefines;
+				print OUTFILE;
+			  }
+		  }
+	    }
 	}
-      close OUTFILE;
-    }
-  else
-    {
-      copy ($infile, "$filename.tex");
     }
+  close OUTFILE;
 
   #
   #  LaTeX, dvips, and assorted cleanups.
   #
   if ($latex2e->{output} eq "tex")
     { 
+# comment out, because this backup action is not documented yet.
+#
+#      if ( -e "$filename.tex" ) {
+#         rename ("$filename.tex", "$filename.tex.back");
+#      }
+
+      umask $saved_umask;
+      copy ("$tmplatexnam.tex", "$filename.tex");
+      if ( ! $global->{debug} )
+        {
+          unlink ("$tmplatexnam.tex");
+          rmdir ($tmplatexdir) || return -1;
+        }
+
       return 0; 
     }
 
@@ -215,30 +400,169 @@ $latex2e->{postASP} = sub
   # references have been resolved.  This avoids large numbers of
   # spurious warnings.
   #
-  my ($latexcommand) = "latex '\\nonstopmode\\input{$filename.tex}'";
+  my $current_dir;
+  chop ($current_dir = `pwd`);
+  print $current_dir . "\n" if ( $global->{debug} );
+
+  #
+  # copy epsfiles specified in tex file
+  #
+  for my $epsf ( @epsfiles )
+   {
+      $tmpepsf = $tmplatexdir . "/" . $epsf; 
+      print $epsf . " " . $tmpepsf . "\n" if ( $global->{debug} );
+      copy ("$epsf", "$tmpepsf") or die "can not copy graphics\n";
+   }
+
+  #
+  # go to the temporary directory
+  chdir ($tmplatexdir);
+
+  my ($latexcommand) = "$latex2e->{latex} '\\nonstopmode\\input{$filename.tex}'";
+
+  #
+  # We run pdflatex instead of latex if user selected pdf output
+  #
+  if ($latex2e->{output} eq "pdf")
+    {
+      $latexcommand = "pdflatex '\\nonstopmode\\input{$filename.tex}'";
+    }
+
+  # 
+  # run hlatex if hlatexp is used
+  # for pdf: how about status?(for hlatex and hlatexp)
+  #
+  if ($latex2e->{latex} eq "hlatexp")
+    {
+      #$latex2e->{output} = "ps" if ($latex2e->{output} eq "pdf");
+      $latexcommand = "hlatex '\\nonstopmode\\input{$filename.tex}'";
+    }
+
+  #
+  # We use jlatex for Japanese encoded (euc-jp) characters.
+  # pdf support for Japanese are not yet. use ps for the time being.
+  #
+  if ($global->{charset} eq "nippon")
+    {
+      $latex2e->{output} = "ps" if ($latex2e->{output} eq "pdf");
+      $latexcommand = "$latex2e->{latex} '\\nonstopmode\\input{$filename.tex}'"
+    }
   my ($suppress) = $latex2e->{quick} ? "" : ' >/dev/null';
 
   system $latexcommand . $suppress  || die "LaTeX problem\n";
   $latex2e->{bibtex} && system "bibtex $filename.tex";
   $latex2e->{quick} || system $latexcommand . ' >/dev/null';
   $latex2e->{quick} || system $latexcommand;
-  if ($global->{debug} == 0)
+  if ( ! $global->{debug} )
     {
-      my @suffixes = qw(log blg aux toc lof lot dlog bbl);
+      my @suffixes = qw(log blg aux toc lof lot dlog bbl out);
       for my $suf (@suffixes)
         {
-          unlink "$filename.$suf";
+          unlink "$tmplatexnam.$suf";
         }
     }
+  #
+  # go back to the working directory
+  chdir ($current_dir);
+  #
+  # output dvi file
   if ($latex2e->{output} eq "dvi")
     {
-      $global->{debug} || unlink "$filename.tex";
+# comment out, because this backup action is not documented yet.
+#
+#      if ( -e "$filename.dvi" )
+#        {
+#          rename ("$filename.dvi", "$filename.dvi.back");
+#        }
+      umask $saved_umask;
+      copy ("$tmplatexnam.dvi", "$filename.dvi");
+      if ( $global->{debug} )
+        {
+          print "Temporary files are in $tmplatexdir\n";
+          print "Please check there and remove them manually.\n";
+        } else {
+          unlink ("$tmplatexnam.tex", "$tmplatexnam.dvi");
+          for my $epsf ( @epsfiles )
+            {
+               $tmpepsf = $tmplatexdir . "/" . $epsf; 
+               print $tmpepsf . "\n" if ( $global->{debug} );
+               unlink ("$tmpepsf");
+            }
+          rmdir ($tmplatexdir) || return -1;
+        }
+      return 0;
+    }
+  #
+  # output pdf file
+  if ($latex2e->{output} eq "pdf")
+    {
+# comment out, because this backup action is not documented yet.
+#
+#      if ( -e "$filename.pdf" )
+#        {
+#          rename ("$filename.pdf", "$filename.pdf.back");
+#        }
+      umask $saved_umask;
+      copy ("$tmplatexnam.pdf", "$filename.pdf");
+      if ( $global->{debug} )
+        {
+          print "Temporary files are in $tmplatexdir\n";
+          print "Please check there and remove them manually.\n";
+        } else {
+          unlink ("$tmplatexnam.tex", "$tmplatexnam.pdf");
+          for my $epsf ( @epsfiles )
+            {
+               $tmpepsf = $tmplatexdir . "/" . $epsf; 
+               print $tmpepsf . "\n" if ( $global->{debug} );
+               unlink ("$tmpepsf");
+            }
+          rmdir ($tmplatexdir) || return -1;
+        }
       return 0;
     }
-  `dvips -q -t $global->{papersize} -o $filename.ps $filename.dvi`;
-  $global->{debug} || unlink ("$filename.dvi", "$filename.tex");
+  #
+  # convert dvi into ps using dvips command
+  chdir ($tmplatexdir);
+  if ($latex2e->{dvips} eq "dvi2ps")
+    {
+      `dvi2ps -q -o $global->{papersize} -c $tmplatexnam.ps $filename.dvi`;
+    }
+  elsif ($latex2e->{dvips} eq "jdvi2kps")
+    {
+      `jdvi2kps -q -pa $global->{papersize} -o $tmplatexnam.ps $filename.dvi`;
+    }
+  else
+    {
+      `dvips -R -q -t $global->{papersize} -o $tmplatexnam.ps $filename.dvi`;
+    }
+
+  chdir ($current_dir);
 
+# comment out, because this backup action is not documented yet.
+#
+#  if ( -e "$filename.ps" )
+#    {
+#      rename ("$filename.ps", "$filename.ps.back");
+#    }
+  umask $saved_umask;
+  copy ("$tmplatexnam.ps", "$filename.ps");
+  unlink ("$tmplatexnam.ps");
+  if ( $global->{debug} )
+    {
+      print "Temporary files are in $tmplatexdir\n";
+      print "Please check there and remove them manually.\n";
+    } else {
+      unlink ("$tmplatexnam.tex", "$tmplatexnam.dvi", "$tmplatexnam.ps");
+      for my $epsf ( @epsfiles )
+        {
+           $tmpepsf = $tmplatexdir . "/" . $epsf; 
+           print $tmpepsf . "\n" if ( $global->{debug} );
+           unlink ("$tmpepsf");
+        }
+      rmdir ($tmplatexdir) || return -1;
+    }
   return 0;
+
 };
 
 1;
diff --git a/doc/sbase/dist/fmt_txt.pl b/doc/sbase/dist/fmt_txt.pl
index 81c2b15..87d70d3 100644
--- a/doc/sbase/dist/fmt_txt.pl
+++ b/doc/sbase/dist/fmt_txt.pl
@@ -7,24 +7,27 @@
 #
 #  © Copyright 1996, Cees de Groot
 #
-package SGMLTools::fmt_txt;
+package LinuxDocTools::fmt_txt;
 use strict;
 
 use File::Copy;
 use Text::EntityMap;
-use SGMLTools::CharEnts;
-use SGMLTools::Lang;
-use SGMLTools::Vars;
+use LinuxDocTools::CharEnts;
+use LinuxDocTools::Lang;
+use LinuxDocTools::Vars;
+use LinuxDocTools::Utils qw(create_temp);
 
 my $txt = {};
 $txt->{NAME} = "txt";
 $txt->{HELP} = "";
 $txt->{OPTIONS} = [
    { option => "manpage", type => "f", short => "m" },
-   { option => "filter",  type => "f", short => "f" }
+   { option => "filter",  type => "f", short => "f" },
+   { option => "blanks",  type => "i", short => "b" }
 ];
 $txt->{manpage} = 0;
 $txt->{filter}  = 0;
+$txt->{blanks}  = 3;
 
 $Formats{$txt->{NAME}} = $txt;
 
@@ -44,19 +47,26 @@ $txt->{preNSGMLS} = sub
       $global->{charset} = "latin1" if $global->{charset} eq "latin";
     }
 
+
   #
   #  Is there a cleaner solution than this? Can't do it earlier,
   #  would show up in the help messages...
   #
+  #  the language support ja.
+  #  the charset  support nippon.
+  #
   $global->{format} = $global->{charset};
+  $global->{charset} = "nippon" if $global->{language} eq "ja";
   $global->{format} = "groff" if $global->{format} eq "ascii";
+  $global->{format} = "groff" if $global->{format} eq "nippon";
+  $global->{format} = "groff" if $global->{format} eq "euc-kr";
   $ENV{SGML_SEARCH_PATH} =~ s/txt/$global->{format}/;
 
   $Formats{"groff"} = $txt;
   $Formats{"latin1"} = $txt;
   $Formats{"man"} = $txt;
 
-  return 0;
+  $global->{NsgmlsPrePipe} = "cat $global->{file} " ;
 };
 
 
@@ -81,16 +91,48 @@ $txt->{preASP} = sub
 {
   my ($infile, $outfile) = @_;
   my (@toc, @lines);
-  if ($txt->{manpage})
+  my $char_maps = load_char_maps ('.2tr', [ Text::EntityMap::sdata_dirs() ]);
+  if ( $global->{charset} eq "latin1" )
+   {
+    $char_maps = load_char_maps ('.2l1tr', [ Text::EntityMap::sdata_dirs() ]);
+   }
+
+   if ($txt->{manpage})
     {
-      copy ($infile, $outfile);
+     while (<$infile>)
+      {
+       if (/^-/)
+        {
+         my ($str) = $';
+         chop ($str);
+         print $outfile "-" . 
+                 parse_data ($str, $char_maps, $txt_escape) . "\n";
+         next;
+        }
+       elsif (/^A/)
+        {
+         /^A(\S+) (IMPLIED|CDATA|NOTATION|ENTITY|TOKEN)( (.*))?$/
+             || die "bad attribute data: $_\n";
+         my ($name,$type,$value) = ($1,$2,$4);
+         if ($type eq "CDATA")
+          {
+            # CDATA attributes get translated also
+            $value = parse_data ($value, $char_maps, $txt_escape);
+          }
+         print $outfile "A$name $type $value\n";
+         next;
+        }
+        #
+        #  Default action if not skipped over with next: copy in to out.
+        #
+        print $outfile $_;
+      }
+
       return;
     }
 
   # note the conversion of `sdata_dirs' list to an anonymous array to
   # make a single argument
-  my $char_maps = load_char_maps ('.2tr', [ Text::EntityMap::sdata_dirs() ]);
-  $char_maps = load_char_maps ('.2l1tr', [ Text::EntityMap::sdata_dirs() ]) if $global->{charset} eq "latin1";
 
   #
   #  Build TOC. The file is read into @lines in the meantime, we need to
@@ -103,6 +145,8 @@ $txt->{preASP} = sub
   push (@toc, ")P\n");
   push (@toc, "(VERB\n");
   my (@prevheader, @header);
+  my $appendix = 0;
+  my $nonprint = 0;
   while (<$infile>)
     {
       push (@lines, $_);
@@ -111,16 +155,27 @@ $txt->{preASP} = sub
         {
 	  @prevheader = @header;
 	  @header = @header[0..$1];
-	  $header[$1]++;
+	  if ($appendix == 1) 
+            {
+	      $header[$1] = "A";
+	      $appendix = 0;
+            } else 
+            {
+	      $header[$1]++;
+	    }
+        }
+      if (/^\(APPEND(.*)/) 
+        {
+	  $appendix = 1;
         }
       if (/^\(HEADING/) 
         {
 	  $_ = <$infile>;
+	  s/\\n/ /g;
 	  push (@lines, $_);
 	  chop;
 	  s/^-//;
 	  $_ = join(".",@header) . " " . $_;
-	  s/\\n/ /g;
 	  s/\(\\[0-9][0-9][0-9]\)/\\\1/g;
 
 	  if (!$#header) 
@@ -130,15 +185,63 @@ $txt->{preASP} = sub
 	      $_ = "\\n" . $_ unless (!$#prevheader);
 	      # put a . and a space after top level sections
 	      s/ /. /;
-	      $_ = "-" . $_ . "\\n";
+#####	      $_ = "-" . $_ . "\\n";
+	      $_ = "-" . $_;
 	    } 
 	  else 
 	    {
 	      # subsections get indentation matching hierarchy
 	      $_ = "-" . "   " x $#header . $_;
 	    }
-	  push(@toc, parse_data ($_, $char_maps, $txt_escape), "\\n\n");
-      }   
+
+#	remove tags from a toc
+	  s/\)TT//g;
+	  s/\(TT//g;
+	  s/\)IT//g;
+	  s/\(IT//g;
+	  s/\)EM//g;
+	  s/\(EM//g;
+	  s/\)BF//g;
+	  s/\(BF//g;
+	  s/AID * CDATA.*$//g;
+	  s/\)LABEL//g;
+	  s/\(LABEL//g;
+
+	  push(@toc, parse_data ($_, $char_maps, $txt_escape));
+
+	  $_ = <$infile>;
+	  while (!/^\)HEADING/) {
+	    s/\\n/ /g; ####
+	    push(@lines, $_);
+	    chop;
+	    s/^-//;
+
+#	remove tags from a toc
+	    s/\)TT//g;
+	    s/\(TT//g;
+	    s/\)IT//g;
+	    s/\(IT//g;
+	    s/\)EM//g;
+	    s/\(EM//g;
+	    s/\)BF//g;
+	    s/\(BF//g;
+	    s/AID * CDATA.*$//g;
+	    s/\)LABEL//g;
+	    s/\(LABEL//g;
+
+#	remove NIDX, NCDX from a toc entry
+	    if (/^\(NIDX$/ || /^\(NCDX$/) { $nonprint = 1; }
+	    if (/^\)NIDX$/ || /^\)NCDX$/) { $nonprint = 1; }
+
+#	  $_ = "-" . $_ . "\\n";
+	    push(@toc, parse_data ($_, $char_maps, $txt_escape))
+	      if (! $nonprint);
+	    $_ = <$infile>;
+	  }
+	  s/\\n/ /g; ###
+	  push(@lines, $_);
+	  push(@toc, "\\n\n");
+      }
     }
   push (@toc, ")VERB\n");
   push (@toc, "(HLINE\n");
@@ -233,8 +336,9 @@ $txt->{postASP} = sub
     }
   else
     {
+      create_temp("$global->{tmpbase}.txt.1");
       $outfile = new FileHandle 
-	  "|$main::progs->{GROFF} -T $global->{pass} $global->{charset} -t $main::progs->{GROFFMACRO} >$global->{tmpbase}.txt.1";
+	  "|$main::progs->{GROFF} $global->{pass} -T $global->{charset} -t $main::progs->{GROFFMACRO} >\"$global->{tmpbase}.txt.1\"";
     }
 
   #
@@ -265,17 +369,36 @@ $txt->{postASP} = sub
     {
       $outfile->open (">$global->{filename}.txt");
       $groffout = new FileHandle "<$global->{tmpbase}.txt.1";
+      my $count = 0;
       if ($txt->{filter})
         {
 	  while (<$groffout>)
 	    {
+	      s/[^\cH][^\cH]\cH\cH//g;
 	      s/.//g;
-	      print $outfile $_;
+              if ($txt->{blanks})
+                {
+                  $count = &{$txt->{cutblank}}($count, $outfile, $_);
+                }
+              else
+                {
+                  print $outfile $_;
+                }
 	    }
 	}
       else
         {
-	  copy ($groffout, $outfile);
+          if ($txt->{blanks})
+            {
+              while (<$groffout>)
+                {
+                  $count = &{$txt->{cutblank}}($count, $outfile, $_);
+                }
+            }
+          else
+            {
+	      copy ($groffout, $outfile);
+            }
 	}
     }
   $groffout->close;
@@ -284,4 +407,23 @@ $txt->{postASP} = sub
   return 0;
 };
 
+$txt->{cutblank} = sub
+{
+  my ($num, $out, $in) = @_;
+  if ( $in =~ /^$/ )
+    {
+      $num++;
+    }
+  else
+    {
+      $num = 0;
+    }
+  if ( $num <= $txt->{blanks} )
+    {
+      print $out $in;
+    }
+
+  return ($num);
+};
+
 1;