doc/rdsrc.pl

   1 #!/usr/bin/perl
   2
   3 # Read the source-form of the NASM manual and generate the various
   4 # output forms.
   5
   6 # TODO:
   7 #
   8 # Ellipsis support would be nice.
   9
  10 # Source-form features:
  11 # ---------------------
  12 #
  13 # Bullet \b
  14 #   Bullets the paragraph. Rest of paragraph is indented to cope. In
  15 #   HTML, consecutive groups of bulleted paragraphs become unordered
  16 #   lists.
  17 #
  18 # Emphasis \e{foobar}
  19 #   produces `_foobar_' in text and italics in HTML, PS, RTF
  20 #
  21 # Inline code \c{foobar}
  22 #   produces ``foobar'' in text, and fixed-pitch font in HTML, PS, RTF
  23 #
  24 # Display code
  25 # \c  line one
  26 # \c   line two
  27 #   produces fixed-pitch font where appropriate, and doesn't break
  28 #   pages except sufficiently far into the middle of a display.
  29 #
  30 # Chapter, header and subheader
  31 # \C{intro} Introduction
  32 # \H{whatsnasm} What is NASM?
  33 # \S{free} NASM Is Free
  34 #   dealt with as appropriate. Chapters begin on new sides, possibly
  35 #   even new _pages_. (Sub)?headers are good places to begin new
  36 #   pages. Just _after_ a (sub)?header isn't.
  37 #   The keywords can be substituted with \K and \k.
  38 #
  39 # Keyword \K{cintro} \k{cintro}
  40 #   Expands to `Chapter 1', `Section 1.1', `Section 1.1.1'. \K has an
  41 #   initial capital whereas \k doesn't. In HTML, will produce
  42 #   hyperlinks.
  43 #
  44 # Web link \W{http://foobar/}{text} or \W{mailto:me@here}\c{me@here}
  45 #   the \W prefix is ignored except in HTML; in HTML the last part
  46 #   becomes a hyperlink to the first part.
  47 #
  48 # Literals \{ \} \\
  49 #   In case it's necessary, they expand to the real versions.
  50 #
  51 # Nonbreaking hyphen \-
  52 #   Need more be said?
  53 #
  54 # Source comment \#
  55 #   Causes everything after it on the line to be ignored by the
  56 #   source-form processor.
  57 #
  58 # Indexable word \i{foobar} (or \i\e{foobar} or \i\c{foobar}, equally)
  59 #   makes word appear in index, referenced to that point
  60 #   \i\c comes up in code style even in the index; \i\e doesn't come
  61 #   up in emphasised style.
  62 #
  63 # Indexable non-displayed word \I{foobar} or \I\c{foobar}
  64 #   just as \i{foobar} except that nothing is displayed for it
  65 #
  66 # Index rewrite
  67 # \IR{foobar} \c{foobar} operator, uses of
  68 #   tidies up the appearance in the index of something the \i or \I
  69 #   operator was applied to
  70 #
  71 # Index alias
  72 # \IA{foobar}{bazquux}
  73 #   aliases one index tag (as might be supplied to \i or \I) to
  74 #   another, so that \I{foobar} has the effect of \I{bazquux}, and
  75 #   \i{foobar} has the effect of \I{bazquux}foobar
  76 #
  77 # Metadata
  78 # \M{key}{something}
  79 #   defines document metadata, such as authorship, title and copyright;
  80 #   different output formats use this differently.
  81 #
  82 # Include subfile
  83 # \&{filename}
  84 #  Includes filename. Recursion is allowed.
  85 #
  86
  87 use IO::File;
  88
  89 $diag = 1, shift @ARGV if $ARGV[0] eq "-d";
  90
  91 ($out_format) = @ARGV;
  92
  93 $| = 1;
  94
  95 $tstruct_previtem = $node = "Top";
  96 $nodes = ($node);
  97 $tstruct_level{$tstruct_previtem} = 0;
  98 $tstruct_last[$tstruct_level{$tstruct_previtem}] = $tstruct_previtem;
  99 $MAXLEVEL = 10;  # really 3, but play safe ;-)
 100
 101 # Read the file; pass a paragraph at a time to the paragraph processor.
 102 print "Reading input...";
 103 $pname = "para000000";
 104 @pnames = @pflags = ();
 105 $para = undef;
 106 while (defined($_ = <STDIN>)) {
 107   $_ = &untabify($_);
 108   &check_include($_);
 109 }
 110 &got_para($para);
 111 print "done.\n";
 112
 113 # Now we've read in the entire document and we know what all the
 114 # heading keywords refer to. Go through and fix up the \k references.
 115 print "Fixing up cross-references...";
 116 &fixup_xrefs;
 117 print "done.\n";
 118
 119 # Sort the index tags, according to the slightly odd order I've decided on.
 120 print "Sorting index tags...";
 121 &indexsort;
 122 print "done.\n";
 123
 124 if ($diag) {
 125   print "Writing index-diagnostic file...";
 126   &indexdiag;
 127   print "done.\n";
 128 }
 129
 130 # OK. Write out the various output files.
 131 if ($out_format eq 'txt') {
 132     print "Producing text output: ";
 133     &write_txt;
 134     print "done.\n";
 135 } elsif ($out_format eq 'html') {
 136     print "Producing HTML output: ";
 137     &write_html;
 138     print "done.\n";
 139 } elsif ($out_format eq 'texi') {
 140     print "Producing Texinfo output: ";
 141     &write_texi;
 142     print "done.\n";
 143 } elsif ($out_format eq 'hlp') {
 144     print "Producing WinHelp output: ";
 145     &write_hlp;
 146     print "done.\n";
 147 } elsif ($out_format eq 'dip') {
 148     print "Producing Documentation Intermediate Paragraphs: ";
 149     &write_dip;
 150     print "done.\n";
 151 } else {
 152     die "$0: unknown output format: $out_format\n";
 153 }
 154
 155 sub untabify($) {
 156   my($s) = @_;
 157   my $o = '';
 158   my($c, $i, $p);
 159
 160   $p = 0;
 161   for ($i = 0; $i < length($s); $i++) {
 162     $c = substr($s, $i, 1);
 163     if ($c eq "\t") {
 164       do {
 165         $o .= ' ';
 166         $p++;
 167       } while ($p & 7);
 168     } else {
 169       $o .= $c;
 170       $p++;
 171     }
 172   }
 173   return $o;
 174 }
 175 sub check_include {
 176   local $_ = shift;
 177   if (/\\& (\S+)/) {
 178      &include($1);
 179   } else {
 180      &get_para($_);
 181   }
 182 }
 183 sub get_para($_) {
 184   chomp;
 185   if (!/\S/ || /^\\(IA|IR|M)/) { # special case: \IA \IR \M imply new-paragraph
 186     &got_para($para);
 187     $para = undef;
 188   }
 189   if (/\S/) {
 190     s/\\#.*$//; # strip comments
 191     $para .= " " . $_;
 192   }
 193 }
 194 sub include {
 195   my $name = shift;
 196   my $F = IO::File->new($name)
 197      or die "Cannot open $name: $!";
 198   while (<$F>) {
 199      &check_include($_);
 200   }
 201 }
 202 sub got_para {
 203   local ($_) = @_;
 204   my $pflags = "", $i, $w, $l, $t;
 205   return if !/\S/;
 206
 207   @$pname = ();
 208
 209   # Strip off _leading_ spaces, then determine type of paragraph.
 210   s/^\s*//;
 211   $irewrite = undef;
 212   if (/^\\c[^{]/) {
 213     # A code paragraph. The paragraph-array will contain the simple
 214     # strings which form each line of the paragraph.
 215     $pflags = "code";
 216     while (/^\\c (([^\\]|\\[^c])*)(.*)$/) {
 217       $l = $1;
 218       $_ = $3;
 219       $l =~ s/\\{/{/g;
 220       $l =~ s/\\}/}/g;
 221       $l =~ s/\\\\/\\/g;
 222       push @$pname, $l;
 223     }
 224     $_ = ''; # suppress word-by-word code
 225   } elsif (/^\\C/) {
 226     # A chapter heading. Define the keyword and allocate a chapter
 227     # number.
 228     $cnum++;
 229     $hnum = 0;
 230     $snum = 0;
 231     $xref = "chapter-$cnum";
 232     $pflags = "chap $cnum :$xref";
 233     die "badly formatted chapter heading: $_\n" if !/^\\C{([^}]*)}\s*(.*)$/;
 234     $refs{$1} = "chapter $cnum";
 235     $node = "Chapter $cnum";
 236     &add_item($node, 1);
 237     $xrefnodes{$node} = $xref; $nodexrefs{$xref} = $node;
 238     $xrefs{$1} = $xref;
 239     $_ = $2;
 240     # the standard word-by-word code will happen next
 241   } elsif (/^\\A/) {
 242     # An appendix heading. Define the keyword and allocate an appendix
 243     # letter.
 244     $cnum++;
 245     $cnum = 'A' if $cnum =~ /[0-9]+/;
 246     $hnum = 0;
 247     $snum = 0;
 248     $xref = "appendix-$cnum";
 249     $pflags = "appn $cnum :$xref";
 250     die "badly formatted appendix heading: $_\n" if !/^\\A{([^}]*)}\s*(.*)$/;
 251     $refs{$1} = "appendix $cnum";
 252     $node = "Appendix $cnum";
 253     &add_item($node, 1);
 254     $xrefnodes{$node} = $xref; $nodexrefs{$xref} = $node;
 255     $xrefs{$1} = $xref;
 256     $_ = $2;
 257     # the standard word-by-word code will happen next
 258   } elsif (/^\\H/) {
 259     # A major heading. Define the keyword and allocate a section number.
 260     $hnum++;
 261     $snum = 0;
 262     $xref = "section-$cnum.$hnum";
 263     $pflags = "head $cnum.$hnum :$xref";
 264     die "badly formatted heading: $_\n" if !/^\\[HP]{([^}]*)}\s*(.*)$/;
 265     $refs{$1} = "section $cnum.$hnum";
 266     $node = "Section $cnum.$hnum";
 267     &add_item($node, 2);
 268     $xrefnodes{$node} = $xref; $nodexrefs{$xref} = $node;
 269     $xrefs{$1} = $xref;
 270     $_ = $2;
 271     # the standard word-by-word code will happen next
 272   } elsif (/^\\S/) {
 273     # A sub-heading. Define the keyword and allocate a section number.
 274     $snum++;
 275     $xref = "section-$cnum.$hnum.$snum";
 276     $pflags = "subh $cnum.$hnum.$snum :$xref";
 277     die "badly formatted subheading: $_\n" if !/^\\S{([^}]*)}\s*(.*)$/;
 278     $refs{$1} = "section $cnum.$hnum.$snum";
 279     $node = "Section $cnum.$hnum.$snum";
 280     &add_item($node, 3);
 281     $xrefnodes{$node} = $xref; $nodexrefs{$xref} = $node;
 282     $xrefs{$1} = $xref;
 283     $_ = $2;
 284     # the standard word-by-word code will happen next
 285   } elsif (/^\\IR/) {
 286     # An index-rewrite.
 287     die "badly formatted index rewrite: $_\n" if !/^\\IR{([^}]*)}\s*(.*)$/;
 288     $irewrite = $1;
 289     $_ = $2;
 290     # the standard word-by-word code will happen next
 291   } elsif (/^\\IA/) {
 292     # An index-alias.
 293     die "badly formatted index alias: $_\n" if !/^\\IA{([^}]*)}{([^}]*)}\s*$/;
 294     $idxalias{$1} = $2;
 295     return; # avoid word-by-word code
 296   } elsif (/^\\M/) {
 297     # Metadata
 298     die "badly formed metadata: $_\n" if !/^\\M{([^}]*)}{([^}]*)}\s*$/;
 299     $metadata{$1} = $2;
 300     return; # avoid word-by-word code
 301   } elsif (/^\\b/) {
 302     # A bulleted paragraph. Strip off the initial \b and let the
 303     # word-by-word code take care of the rest.
 304     $pflags = "bull";
 305     s/^\\b\s*//;
 306   } else {
 307     # A normal paragraph. Just set $pflags: the word-by-word code does
 308     # the rest.
 309     $pflags = "norm";
 310   }
 311
 312   # The word-by-word code: unless @$pname is already defined (which it
 313   # will be in the case of a code paragraph), split the paragraph up
 314   # into words and push each on @$pname.
 315   #
 316   # Each thing pushed on @$pname should have a two-character type
 317   # code followed by the text.
 318   #
 319   # Type codes are:
 320   # "n " for normal
 321   # "da" for a dash
 322   # "es" for first emphasised word in emphasised bit
 323   # "e " for emphasised in mid-emphasised-bit
 324   # "ee" for last emphasised word in emphasised bit
 325   # "eo" for single (only) emphasised word
 326   # "c " for code
 327   # "k " for cross-ref
 328   # "kK" for capitalised cross-ref
 329   # "w " for Web link
 330   # "wc" for code-type Web link
 331   # "x " for beginning of resolved cross-ref; generates no visible output,
 332   #      and the text is the cross-reference code
 333   # "xe" for end of resolved cross-ref; text is same as for "x ".
 334   # "i " for point to be indexed: the text is the internal index into the
 335   #      index-items arrays
 336   # "sp" for space
 337   while (/\S/) {
 338     s/^\s*//, push @$pname, "sp" if /^\s/;
 339     $indexing = $qindex = 0;
 340     if (/^(\\[iI])?\\c/) {
 341       $qindex = 1 if $1 eq "\\I";
 342       $indexing = 1, s/^\\[iI]// if $1;
 343       s/^\\c//;
 344       die "badly formatted \\c: \\c$_\n" if !/{(([^\\}]|\\.)*)}(.*)$/;
 345       $w = $1;
 346       $_ = $3;
 347       $w =~ s/\\{/{/g;
 348       $w =~ s/\\}/}/g;
 349       $w =~ s/\\-/-/g;
 350       $w =~ s/\\\\/\\/g;
 351       (push @$pname,"i"),$lastp = $#$pname if $indexing;
 352       push @$pname,"c $w" if !$qindex;
 353       $$pname[$lastp] = &addidx($node, $w, "c $w") if $indexing;
 354     } elsif (/^\\[iIe]/) {
 355       /^(\\[iI])?(\\e)?/;
 356       $emph = 0;
 357       $qindex = 1 if $1 eq "\\I";
 358       $indexing = 1, $type = "\\i" if $1;
 359       $emph = 1, $type = "\\e" if $2;
 360       s/^(\\[iI])?(\\e?)//;
 361       die "badly formatted $type: $type$_\n" if !/{(([^\\}]|\\.)*)}(.*)$/;
 362       $w = $1;
 363       $_ = $3;
 364       $w =~ s/\\{/{/g;
 365       $w =~ s/\\}/}/g;
 366       $w =~ s/\\-/-/g;
 367       $w =~ s/\\\\/\\/g;
 368       $t = $emph ? "es" : "n ";
 369       @ientry = ();
 370       (push @$pname,"i"),$lastp = $#$pname if $indexing;
 371       foreach $i (split /\s+/,$w) {  # \e and \i can be multiple words
 372         push @$pname,"$t$i","sp" if !$qindex;
 373         ($ii=$i) =~ tr/A-Z/a-z/, push @ientry,"n $ii","sp" if $indexing;
 374         $t = $emph ? "e " : "n ";
 375       }
 376       $w =~ tr/A-Z/a-z/, pop @ientry if $indexing;
 377       $$pname[$lastp] = &addidx($node, $w, @ientry) if $indexing;
 378       pop @$pname if !$qindex; # remove final space
 379       if (substr($$pname[$#$pname],0,2) eq "es" && !$qindex) {
 380         substr($$pname[$#$pname],0,2) = "eo";
 381       } elsif ($emph && !$qindex) {
 382         substr($$pname[$#$pname],0,2) = "ee";
 383       }
 384     } elsif (/^\\[kK]/) {
 385       $t = "k ";
 386       $t = "kK" if /^\\K/;
 387       s/^\\[kK]//;
 388       die "badly formatted \\k: \\c$_\n" if !/{([^}]*)}(.*)$/;
 389       $_ = $2;
 390       push @$pname,"$t$1";
 391     } elsif (/^\\W/) {
 392       s/^\\W//;
 393       die "badly formatted \\W: \\W$_\n"
 394           if !/{([^}]*)}(\\i)?(\\c)?{(([^\\}]|\\.)*)}(.*)$/;
 395       $l = $1;
 396       $w = $4;
 397       $_ = $6;
 398       $t = "w ";
 399       $t = "wc" if $3 eq "\\c";
 400       $indexing = 1 if $2;
 401       $w =~ s/\\{/{/g;
 402       $w =~ s/\\}/}/g;
 403       $w =~ s/\\-/-/g;
 404       $w =~ s/\\\\/\\/g;
 405       (push @$pname,"i"),$lastp = $#$pname if $indexing;
 406       push @$pname,"$t<$l>$w";
 407       $$pname[$lastp] = &addidx($node, $w, "c $w") if $indexing;
 408     } else {
 409       die "what the hell? $_\n" if !/^(([^\s\\\-]|\\[\\{}\-])*-?)(.*)$/;
 410       die "painful death! $_\n" if !length $1;
 411       $w = $1;
 412       $_ = $3;
 413       $w =~ s/\\{/{/g;
 414       $w =~ s/\\}/}/g;
 415       $w =~ s/\\-/-/g;
 416       $w =~ s/\\\\/\\/g;
 417       if ($w eq "-") {
 418         push @$pname,"da";
 419       } else {
 420         push @$pname,"n $w";
 421       }
 422     }
 423   }
 424   if ($irewrite ne undef) {
 425     &addidx(undef, $irewrite, @$pname);
 426     @$pname = ();
 427   } else {
 428     push @pnames, $pname;
 429     push @pflags, $pflags;
 430     $pname++;
 431   }
 432 }
 433
 434 sub addidx {
 435   my ($node, $text, @ientry) = @_;
 436   $text = $idxalias{$text} || $text;
 437   if ($node eq undef || !$idxmap{$text}) {
 438     @$ientry = @ientry;
 439     $idxmap{$text} = $ientry;
 440     $ientry++;
 441   }
 442   if ($node) {
 443     $idxnodes{$node,$text} = 1;
 444     return "i $text";
 445   }
 446 }
 447
 448 sub indexsort {
 449   my $iitem, $ientry, $i, $piitem, $pcval, $cval, $clrcval;
 450
 451   @itags = map { # get back the original data as the 1st elt of each list
 452              $_->[0]
 453            } sort { # compare auxiliary (non-first) elements of lists
 454              $a->[1] cmp $b->[1] ||
 455              $a->[2] cmp $b->[2] ||
 456              $a->[0] cmp $b->[0]
 457            } map { # transform array into list of 3-element lists
 458              my $ientry = $idxmap{$_};
 459              my $a = substr($$ientry[0],2);
 460              $a =~ tr/A-Za-z0-9//cd;
 461              [$_, uc($a), substr($$ientry[0],0,2)]
 462            } keys %idxmap;
 463
 464   # Having done that, check for comma-hood.
 465   $cval = 0;
 466   foreach $iitem (@itags) {
 467     $ientry = $idxmap{$iitem};
 468     $clrcval = 1;
 469     $pcval = $cval;
 470     FL:for ($i=0; $i <= $#$ientry; $i++) {
 471       if ($$ientry[$i] =~ /^(n .*,)(.*)/) {
 472         $$ientry[$i] = $1;
 473         splice @$ientry,$i+1,0,"n $2" if length $2;
 474         $commapos{$iitem} = $i+1;
 475         $cval = join("\002", @$ientry[0..$i]);
 476         $clrcval = 0;
 477         last FL;
 478       }
 479     }
 480     $cval = undef if $clrcval;
 481     $commanext{$iitem} = $commaafter{$piitem} = 1
 482       if $cval and ($cval eq $pcval);
 483     $piitem = $iitem;
 484   }
 485 }
 486
 487 sub indexdiag {
 488   my $iitem,$ientry,$w,$ww,$foo,$node;
 489   open INDEXDIAG,">index.diag";
 490   foreach $iitem (@itags) {
 491     $ientry = $idxmap{$iitem};
 492     print INDEXDIAG "<$iitem> ";
 493     foreach $w (@$ientry) {
 494       $ww = &word_txt($w);
 495       print INDEXDIAG $ww unless $ww eq "\001";
 496     }
 497     print INDEXDIAG ":";
 498     $foo = " ";
 499     foreach $node (@nodes) {
 500       (print INDEXDIAG $foo,$node), $foo = ", " if $idxnodes{$node,$iitem};
 501     }
 502     print INDEXDIAG "\n";
 503   }
 504   close INDEXDIAG;
 505 }
 506
 507 sub fixup_xrefs {
 508   my $pname, $p, $i, $j, $k, $caps, @repl;
 509
 510   for ($p=0; $p<=$#pnames; $p++) {
 511     next if $pflags[$p] eq "code";
 512     $pname = $pnames[$p];
 513     for ($i=$#$pname; $i >= 0; $i--) {
 514       if ($$pname[$i] =~ /^k/) {
 515         $k = $$pname[$i];
 516         $caps = ($k =~ /^kK/);
 517         $k = substr($k,2);
 518         $repl = $refs{$k};
 519         die "undefined keyword `$k'\n" unless $repl;
 520         substr($repl,0,1) =~ tr/a-z/A-Z/ if $caps;
 521         @repl = ();
 522         push @repl,"x $xrefs{$k}";
 523         foreach $j (split /\s+/,$repl) {
 524           push @repl,"n $j";
 525           push @repl,"sp";
 526         }
 527         pop @repl; # remove final space
 528         push @repl,"xe$xrefs{$k}";
 529         splice @$pname,$i,1,@repl;
 530       }
 531     }
 532   }
 533 }
 534
 535 sub write_txt {
 536   # This is called from the top level, so I won't bother using
 537   # my or local.
 538
 539   # Open file.
 540   print "writing file...";
 541   open TEXT,">nasmdoc.txt";
 542   select TEXT;
 543
 544   # Preamble.
 545   $title = "The Netwide Assembler: NASM";
 546   $spaces = ' ' x ((75-(length $title))/2);
 547   ($underscore = $title) =~ s/./=/g;
 548   print "$spaces$title\n$spaces$underscore\n";
 549
 550   for ($para = 0; $para <= $#pnames; $para++) {
 551     $pname = $pnames[$para];
 552     $pflags = $pflags[$para];
 553     $ptype = substr($pflags,0,4);
 554
 555     print "\n"; # always one of these before a new paragraph
 556
 557     if ($ptype eq "chap") {
 558       # Chapter heading. "Chapter N: Title" followed by a line of
 559       # minus signs.
 560       $pflags =~ /chap (.*) :(.*)/;
 561       $title = "Chapter $1: ";
 562       foreach $i (@$pname) {
 563         $ww = &word_txt($i);
 564         $title .= $ww unless $ww eq "\001";
 565       }
 566       print "$title\n";
 567       $title =~ s/./-/g;
 568       print "$title\n";
 569     } elsif ($ptype eq "appn") {
 570       # Appendix heading. "Appendix N: Title" followed by a line of
 571       # minus signs.
 572       $pflags =~ /appn (.*) :(.*)/;
 573       $title = "Appendix $1: ";
 574       foreach $i (@$pname) {
 575         $ww = &word_txt($i);
 576         $title .= $ww unless $ww eq "\001";
 577       }
 578       print "$title\n";
 579       $title =~ s/./-/g;
 580       print "$title\n";
 581     } elsif ($ptype eq "head" || $ptype eq "subh") {
 582       # Heading or subheading. Just a number and some text.
 583       $pflags =~ /.... (.*) :(.*)/;
 584       $title = sprintf "%6s ", $1;
 585       foreach $i (@$pname) {
 586         $ww = &word_txt($i);
 587         $title .= $ww unless $ww eq "\001";
 588       }
 589       print "$title\n";
 590     } elsif ($ptype eq "code") {
 591       # Code paragraph. Emit each line with a seven character indent.
 592       foreach $i (@$pname) {
 593         warn "code line longer than 68 chars: $i\n" if length $i > 68;
 594         print ' 'x7, $i, "\n";
 595       }
 596     } elsif ($ptype eq "bull" || $ptype eq "norm") {
 597       # Ordinary paragraph, optionally bulleted. We wrap, with ragged
 598       # 75-char right margin and either 7 or 11 char left margin
 599       # depending on bullets.
 600       if ($ptype eq "bull") {
 601         $line = ' 'x7 . '(*) ';
 602         $next = ' 'x11;
 603       } else {
 604         $line = $next = ' 'x7;
 605       }
 606       @a = @$pname;
 607       $wd = $wprev = '';
 608       do {
 609         do { $w = &word_txt(shift @a) } while $w eq "\001"; # nasty hack
 610         $wd .= $wprev;
 611         if ($wprev =~ /-$/ || $w eq ' ' || $w eq '' || $w eq undef) {
 612           if (length ($line . $wd) > 75) {
 613             $line =~ s/\s*$//; # trim trailing spaces
 614             print "$line\n";
 615             $line = $next;
 616             $wd =~ s/^\s*//; # trim leading spaces
 617           }
 618           $line .= $wd;
 619           $wd = '';
 620         }
 621         $wprev = $w;
 622       } while ($w ne '' && $w ne undef);
 623       if ($line =~ /\S/) {
 624         $line =~ s/\s*$//; # trim trailing spaces
 625         print "$line\n";
 626       }
 627     }
 628   }
 629
 630   # Close file.
 631   select STDOUT;
 632   close TEXT;
 633 }
 634
 635 sub word_txt {
 636   my ($w) = @_;
 637   my $wtype, $wmajt;
 638
 639   return undef if $w eq '' || $w eq undef;
 640   $wtype = substr($w,0,2);
 641   $wmajt = substr($wtype,0,1);
 642   $w = substr($w,2);
 643   $w =~ s/<.*>// if $wmajt eq "w"; # remove web links
 644   if ($wmajt eq "n" || $wtype eq "e " || $wtype eq "w ") {
 645     return $w;
 646   } elsif ($wtype eq "sp") {
 647     return ' ';
 648   } elsif ($wtype eq "da") {
 649     return '-';
 650   } elsif ($wmajt eq "c" || $wtype eq "wc") {
 651     return "`${w}'";
 652   } elsif ($wtype eq "es") {
 653     return "_${w}";
 654   } elsif ($wtype eq "ee") {
 655     return "${w}_";
 656   } elsif ($wtype eq "eo") {
 657     return "_${w}_";
 658   } elsif ($wmajt eq "x" || $wmajt eq "i") {
 659     return "\001";
 660   } else {
 661     die "panic in word_txt: $wtype$w\n";
 662   }
 663 }
 664
 665 sub write_html {
 666   # This is called from the top level, so I won't bother using
 667   # my or local.
 668
 669   # Write contents file. Just the preamble, then a menu of links to the
 670   # separate chapter files and the nodes therein.
 671   print "writing contents file...";
 672   open TEXT,">nasmdoc0.html";
 673   select TEXT;
 674   &html_preamble(0);
 675   print "<p>This manual documents NASM, the Netwide Assembler: an assembler\n";
 676   print "targetting the Intel x86 series of processors, with portable source.\n";
 677   print "<p>";
 678   for ($node = $tstruct_next{'Top'}; $node; $node = $tstruct_next{$node}) {
 679     if ($tstruct_level{$node} == 1) {
 680       # Invent a file name.
 681       ($number = lc($xrefnodes{$node})) =~ s/.*-//;
 682       $fname="nasmdocx.html";
 683       substr($fname,8 - length $number, length $number) = $number;
 684       $html_fnames{$node} = $fname;
 685       $link = $fname;
 686       print "<p>";
 687     } else {
 688       # Use the preceding filename plus a marker point.
 689       $link = $fname . "#$xrefnodes{$node}";
 690     }
 691     $title = "$node: ";
 692     $pname = $tstruct_pname{$node};
 693     foreach $i (@$pname) {
 694       $ww = &word_html($i);
 695       $title .= $ww unless $ww eq "\001";
 696     }
 697     print "<a href=\"$link\">$title</a><br>\n";
 698   }
 699   print "<p><a href=\"nasmdoci.html\">Index</a>\n";
 700   print "</body></html>\n";
 701   select STDOUT;
 702   close TEXT;
 703
 704   # Open a null file, to ensure output (eg random &html_jumppoints calls)
 705   # goes _somewhere_.
 706   print "writing chapter files...";
 707   open TEXT,">/dev/null";
 708   select TEXT;
 709   $html_lastf = '';
 710
 711   $in_list = 0;
 712
 713   for ($para = 0; $para <= $#pnames; $para++) {
 714     $pname = $pnames[$para];
 715     $pflags = $pflags[$para];
 716     $ptype = substr($pflags,0,4);
 717
 718     $in_list = 0, print "</ul>\n" if $in_list && $ptype ne "bull";
 719     if ($ptype eq "chap") {
 720       # Chapter heading. Begin a new file.
 721       $pflags =~ /chap (.*) :(.*)/;
 722       $title = "Chapter $1: ";
 723       $xref = $2;
 724       &html_jumppoints; print "</body></html>\n"; select STDOUT; close TEXT;
 725       $html_lastf = $html_fnames{$chapternode};
 726       $chapternode = $nodexrefs{$xref};
 727       $html_nextf = $html_fnames{$tstruct_mnext{$chapternode}};
 728       open TEXT,">$html_fnames{$chapternode}"; select TEXT; &html_preamble(1);
 729       foreach $i (@$pname) {
 730         $ww = &word_html($i);
 731         $title .= $ww unless $ww eq "\001";
 732       }
 733       $h = "<h2><a name=\"$xref\">$title</a></h2>\n";
 734       print $h; print FULL $h;
 735     } elsif ($ptype eq "appn") {
 736       # Appendix heading. Begin a new file.
 737       $pflags =~ /appn (.*) :(.*)/;
 738       $title = "Appendix $1: ";
 739       $xref = $2;
 740       &html_jumppoints; print "</body></html>\n"; select STDOUT; close TEXT;
 741       $html_lastf = $html_fnames{$chapternode};
 742       $chapternode = $nodexrefs{$xref};
 743       $html_nextf = $html_fnames{$tstruct_mnext{$chapternode}};
 744       open TEXT,">$html_fnames{$chapternode}"; select TEXT; &html_preamble(1);
 745       foreach $i (@$pname) {
 746         $ww = &word_html($i);
 747         $title .= $ww unless $ww eq "\001";
 748       }
 749       print "<h2><a name=\"$xref\">$title</a></h2>\n";
 750     } elsif ($ptype eq "head" || $ptype eq "subh") {
 751       # Heading or subheading.
 752       $pflags =~ /.... (.*) :(.*)/;
 753       $hdr = ($ptype eq "subh" ? "h4" : "h3");
 754       $title = $1 . " ";
 755       $xref = $2;
 756       foreach $i (@$pname) {
 757         $ww = &word_html($i);
 758         $title .= $ww unless $ww eq "\001";
 759       }
 760       print "<$hdr><a name=\"$xref\">$title</a></$hdr>\n";
 761     } elsif ($ptype eq "code") {
 762       # Code paragraph.
 763       print "<p><pre>\n";
 764       foreach $i (@$pname) {
 765         $w = $i;
 766         $w =~ s/&/&amp;/g;
 767         $w =~ s/</&lt;/g;
 768         $w =~ s/>/&gt;/g;
 769         print $w, "\n";
 770       }
 771       print "</pre>\n";
 772     } elsif ($ptype eq "bull" || $ptype eq "norm") {
 773       # Ordinary paragraph, optionally bulleted. We wrap, with ragged
 774       # 75-char right margin and either 7 or 11 char left margin
 775       # depending on bullets.
 776       if ($ptype eq "bull") {
 777         $in_list = 1, print "<ul>\n" unless $in_list;
 778         $line = '<li>';
 779       } else {
 780         $line = '<p>';
 781       }
 782       @a = @$pname;
 783       $wd = $wprev = '';
 784       do {
 785         do { $w = &word_html(shift @a) } while $w eq "\001"; # nasty hack
 786         $wd .= $wprev;
 787         if ($w eq ' ' || $w eq '' || $w eq undef) {
 788           if (length ($line . $wd) > 75) {
 789             $line =~ s/\s*$//; # trim trailing spaces
 790             print "$line\n";
 791             $line = '';
 792             $wd =~ s/^\s*//; # trim leading spaces
 793           }
 794           $line .= $wd;
 795           $wd = '';
 796         }
 797         $wprev = $w;
 798       } while ($w ne '' && $w ne undef);
 799       if ($line =~ /\S/) {
 800         $line =~ s/\s*$//; # trim trailing spaces
 801         print "$line\n";
 802       }
 803     }
 804   }
 805
 806   # Close whichever file was open.
 807   &html_jumppoints;
 808   print "</body></html>\n";
 809   select STDOUT;
 810   close TEXT;
 811
 812   print "\n   writing index file...";
 813   open TEXT,">nasmdoci.html";
 814   select TEXT;
 815   &html_preamble(0);
 816   print "<p align=center><a href=\"nasmdoc0.html\">Contents</a>\n";
 817   print "<p>";
 818   &html_index;
 819   print "<p align=center><a href=\"nasmdoc0.html\">Contents</a>\n";
 820   print "</body></html>\n";
 821   select STDOUT;
 822   close TEXT;
 823 }
 824
 825 sub html_preamble {
 826   print "<html><head><title>NASM Manual</title></head>\n";
 827   print "<body><h1 align=center>The Netwide Assembler: NASM</h1>\n\n";
 828   &html_jumppoints if $_[0];
 829 }
 830
 831 sub html_jumppoints {
 832   print "<p align=center>";
 833   print "<a href=\"$html_nextf\">Next Chapter</a> |\n" if $html_nextf;
 834   print "<a href=\"$html_lastf\">Previous Chapter</a> |\n" if $html_lastf;
 835   print "<a href=\"nasmdoc0.html\">Contents</a> |\n";
 836   print "<a href=\"nasmdoci.html\">Index</a>\n";
 837 }
 838
 839 sub html_index {
 840   my $itag, $a, @ientry, $sep, $w, $wd, $wprev, $line;
 841
 842   $chapternode = '';
 843   foreach $itag (@itags) {
 844     $ientry = $idxmap{$itag};
 845     @a = @$ientry;
 846     push @a, "n :";
 847     $sep = 0;
 848     foreach $node (@nodes) {
 849       next if !$idxnodes{$node,$itag};
 850       push @a, "n ," if $sep;
 851       push @a, "sp", "x $xrefnodes{$node}", "n $node", "xe$xrefnodes{$node}";
 852       $sep = 1;
 853     }
 854     $line = '';
 855     do {
 856       do { $w = &word_html(shift @a) } while $w eq "\001"; # nasty hack
 857       $wd .= $wprev;
 858       if ($w eq ' ' || $w eq '' || $w eq undef) {
 859         if (length ($line . $wd) > 75) {
 860           $line =~ s/\s*$//; # trim trailing spaces
 861           print "$line\n";
 862           $line = '';
 863           $wd =~ s/^\s*//; # trim leading spaces
 864         }
 865         $line .= $wd;
 866         $wd = '';
 867       }
 868       $wprev = $w;
 869     } while ($w ne '' && $w ne undef);
 870     if ($line =~ /\S/) {
 871       $line =~ s/\s*$//; # trim trailing spaces
 872       print "$line\n";
 873     }
 874     print "<br>\n";
 875   }
 876 }
 877
 878 sub word_html {
 879   my ($w) = @_;
 880   my $wtype, $wmajt, $pfx, $sfx;
 881
 882   return undef if $w eq '' || $w eq undef;
 883
 884   $wtype = substr($w,0,2);
 885   $wmajt = substr($wtype,0,1);
 886   $w = substr($w,2);
 887   $pfx = $sfx = '';
 888   $pfx = "<a href=\"$1\">", $sfx = "</a>", $w = $2
 889     if $wmajt eq "w" && $w =~ /^<(.*)>(.*)$/;
 890   $w =~ s/&/&amp;/g;
 891   $w =~ s/</&lt;/g;
 892   $w =~ s/>/&gt;/g;
 893   if ($wmajt eq "n" || $wtype eq "e " || $wtype eq "w ") {
 894     return $pfx . $w . $sfx;
 895   } elsif ($wtype eq "sp") {
 896     return ' ';
 897   } elsif ($wtype eq "da") {
 898     return '-'; # sadly, en-dashes are non-standard in HTML
 899   } elsif ($wmajt eq "c" || $wtype eq "wc") {
 900     return $pfx . "<code><nobr>${w}</nobr></code>" . $sfx;
 901   } elsif ($wtype eq "es") {
 902     return "<em>${w}";
 903   } elsif ($wtype eq "ee") {
 904     return "${w}</em>";
 905   } elsif ($wtype eq "eo") {
 906     return "<em>${w}</em>";
 907   } elsif ($wtype eq "x ") {
 908     # Magic: we must resolve the cross reference into file and marker
 909     # parts, then dispose of the file part if it's us, and dispose of
 910     # the marker part if the cross reference describes the top node of
 911     # another file.
 912     my $node = $nodexrefs{$w}; # find the node we're aiming at
 913     my $level = $tstruct_level{$node}; # and its level
 914     my $up = $node, $uplev = $level-1;
 915     $up = $tstruct_up{$up} while $uplev--; # get top node of containing file
 916     my $file = ($up ne $chapternode) ? $html_fnames{$up} : "";
 917     my $marker = ($level == 1 and $file) ? "" : "#$w";
 918     return "<a href=\"$file$marker\">";
 919   } elsif ($wtype eq "xe") {
 920     return "</a>";
 921   } elsif ($wmajt eq "i") {
 922     return "\001";
 923   } else {
 924     die "panic in word_html: $wtype$w\n";
 925   }
 926 }
 927
 928 sub write_texi {
 929   # This is called from the top level, so I won't bother using
 930   # my or local.
 931
 932   # Open file.
 933   print "writing file...";
 934   open TEXT,">nasmdoc.texi";
 935   select TEXT;
 936
 937   # Preamble.
 938   print "\\input texinfo   \@c -*-texinfo-*-\n";
 939   print "\@c \%**start of header\n";
 940   print "\@setfilename ",$metadata{'infofile'},".info\n";
 941   print "\@dircategory ",$metadata{'category'},"\n";
 942   print "\@direntry\n";
 943   printf "* %-28s %s.\n",
 944   sprintf('%s: (%s).', $metadata{'infoname'}, $metadata{'infofile'}),
 945   $metadata{'infotitle'};
 946   print "\@end direntry\n";
 947   print "\@settitle ", $metadata{'title'},"\n";
 948   print "\@setchapternewpage odd\n";
 949   print "\@c \%**end of header\n";
 950   print "\n";
 951   print "\@ifinfo\n";
 952   print $metadata{'summary'}, "\n";
 953   print "\n";
 954   print "Copyright ",$metadata{'year'}," ",$metadata{'author'},"\n";
 955   print "\n";
 956   print $metadata{'license'}, "\n";
 957   print "\@end ifinfo\n";
 958   print "\n";
 959   print "\@titlepage\n";
 960   $title = $metadata{'title'};
 961   $title =~ s/ - / --- /g;
 962   print "\@title ${title}\n";
 963   print "\@author ",$metadata{'author'},"\n";
 964   print "\n";
 965   print "\@page\n";
 966   print "\@vskip 0pt plus 1filll\n";
 967   print "Copyright \@copyright{} ",$metadata{'year'},' ',$metadata{'author'},"\n";
 968   print "\n";
 969   print $metadata{'license'}, "\n";
 970   print "\@end titlepage\n";
 971   print "\n";
 972   print "\@node Top, $tstruct_next{'Top'}, (dir), (dir)\n";
 973   print "\@top ",$metadata{'infotitle'},"\n";
 974   print "\n";
 975   print "\@ifinfo\n";
 976   print $metadata{'summary'}, "\n";
 977   print "\@end ifinfo\n";
 978
 979   $node = "Top";
 980
 981   $bulleting = 0;
 982   for ($para = 0; $para <= $#pnames; $para++) {
 983     $pname = $pnames[$para];
 984     $pflags = $pflags[$para];
 985     $ptype = substr($pflags,0,4);
 986
 987     $bulleting = 0, print "\@end itemize\n" if $bulleting && $ptype ne "bull";
 988     print "\n"; # always one of these before a new paragraph
 989
 990     if ($ptype eq "chap") {
 991       # Chapter heading. Begin a new node.
 992       &texi_menu($node)
 993         if $tstruct_level{$tstruct_next{$node}} > $tstruct_level{$node};
 994       $pflags =~ /chap (.*) :(.*)/;
 995       $node = "Chapter $1";
 996       $title = "Chapter $1: ";
 997       foreach $i (@$pname) {
 998         $ww = &word_texi($i);
 999         $title .= $ww unless $ww eq "\001";
1000       }
1001       print "\@node $node, $tstruct_next{$node}, $tstruct_prev{$node},";
1002       print " $tstruct_up{$node}\n\@unnumbered $title\n";
1003     } elsif ($ptype eq "appn") {
1004       # Appendix heading. Begin a new node.
1005       &texi_menu($node)
1006         if $tstruct_level{$tstruct_next{$node}} > $tstruct_level{$node};
1007       $pflags =~ /appn (.*) :(.*)/;
1008       $node = "Appendix $1";
1009       $title = "Appendix $1: ";
1010       foreach $i (@$pname) {
1011         $ww = &word_texi($i);
1012         $title .= $ww unless $ww eq "\001";
1013       }
1014       print "\@node $node, $tstruct_next{$node}, $tstruct_prev{$node},";
1015       print " $tstruct_up{$node}\n\@unnumbered $title\n";
1016     } elsif ($ptype eq "head" || $ptype eq "subh") {
1017       # Heading or subheading. Begin a new node.
1018       &texi_menu($node)
1019         if $tstruct_level{$tstruct_next{$node}} > $tstruct_level{$node};
1020       $pflags =~ /.... (.*) :(.*)/;
1021       $node = "Section $1";
1022       $title = "$1. ";
1023       foreach $i (@$pname) {
1024         $ww = &word_texi($i);
1025         $title .= $ww unless $ww eq "\001";
1026       }
1027       print "\@node $node, $tstruct_next{$node}, $tstruct_prev{$node},";
1028       print " $tstruct_up{$node}\n";
1029       $hdr = ($ptype eq "subh" ? "\@unnumberedsubsec" : "\@unnumberedsec");
1030       print "$hdr $title\n";
1031     } elsif ($ptype eq "code") {
1032       # Code paragraph. Surround with @example / @end example.
1033       print "\@example\n";
1034       foreach $i (@$pname) {
1035         warn "code line longer than 68 chars: $i\n" if length $i > 68;
1036         $i =~ s/\@/\@\@/g;
1037         $i =~ s/\{/\@\{/g;
1038         $i =~ s/\}/\@\}/g;
1039         print "$i\n";
1040       }
1041       print "\@end example\n";
1042     } elsif ($ptype eq "bull" || $ptype eq "norm") {
1043       # Ordinary paragraph, optionally bulleted. We wrap, FWIW.
1044       if ($ptype eq "bull") {
1045         $bulleting = 1, print "\@itemize \@bullet\n" if !$bulleting;
1046         print "\@item\n";
1047       }
1048       $line = '';
1049       @a = @$pname;
1050       $wd = $wprev = '';
1051       do {
1052         do { $w = &word_texi(shift @a); } while $w eq "\001"; # hack
1053         $wd .= $wprev;
1054         if ($wprev =~ /-$/ || $w eq ' ' || $w eq '' || $w eq undef) {
1055           if (length ($line . $wd) > 75) {
1056             $line =~ s/\s*$//; # trim trailing spaces
1057             print "$line\n";
1058             $line = '';
1059             $wd =~ s/^\s*//; # trim leading spaces
1060           }
1061           $line .= $wd;
1062           $wd = '';
1063         }
1064         $wprev = $w;
1065       } while ($w ne '' && $w ne undef);
1066       if ($line =~ /\S/) {
1067         $line =~ s/\s*$//; # trim trailing spaces
1068         print "$line\n";
1069       }
1070     }
1071   }
1072
1073   # Write index.
1074   &texi_index;
1075
1076   # Close file.
1077   print "\n\@contents\n\@bye\n";
1078   select STDOUT;
1079   close TEXT;
1080 }
1081
1082 # Side effect of this procedure: update global `texiwdlen' to be the length
1083 # in chars of the formatted version of the word.
1084 sub word_texi {
1085   my ($w) = @_;
1086   my $wtype, $wmajt;
1087
1088   return undef if $w eq '' || $w eq undef;
1089   $wtype = substr($w,0,2);
1090   $wmajt = substr($wtype,0,1);
1091   $w = substr($w,2);
1092   $wlen = length $w;
1093   $w =~ s/\@/\@\@/g;
1094   $w =~ s/\{/\@\{/g;
1095   $w =~ s/\}/\@\}/g;
1096   $w =~ s/<.*>// if $wmajt eq "w"; # remove web links
1097   substr($w,0,1) =~ tr/a-z/A-Z/, $capital = 0 if $capital;
1098   if ($wmajt eq "n" || $wtype eq "e " || $wtype eq "w ") {
1099     $texiwdlen = $wlen;
1100     return $w;
1101   } elsif ($wtype eq "sp") {
1102     $texiwdlen = 1;
1103     return ' ';
1104   } elsif ($wtype eq "da") {
1105     $texiwdlen = 2;
1106     return '--';
1107   } elsif ($wmajt eq "c" || $wtype eq "wc") {
1108     $texiwdlen = 2 + $wlen;
1109     return "\@code\{$w\}";
1110   } elsif ($wtype eq "es") {
1111     $texiwdlen = 1 + $wlen;
1112     return "\@emph\{${w}";
1113   } elsif ($wtype eq "ee") {
1114     $texiwdlen = 1 + $wlen;
1115     return "${w}\}";
1116   } elsif ($wtype eq "eo") {
1117     $texiwdlen = 2 + $wlen;
1118     return "\@emph\{${w}\}";
1119   } elsif ($wtype eq "x ") {
1120     $texiwdlen = 0; # we don't need it in this case
1121     $capital = 1; # hack
1122     return "\@ref\{";
1123   } elsif ($wtype eq "xe") {
1124     $texiwdlen = 0; # we don't need it in this case
1125     return "\}";
1126   } elsif ($wmajt eq "i") {
1127     $texiwdlen = 0; # we don't need it in this case
1128     return "\001";
1129   } else {
1130     die "panic in word_texi: $wtype$w\n";
1131   }
1132 }
1133
1134 sub texi_menu {
1135   my ($topitem) = @_;
1136   my $item, $i, $mpname, $title, $wd;
1137
1138   $item = $tstruct_next{$topitem};
1139   print "\@menu\n";
1140   while ($item) {
1141     $title = "";
1142     $mpname = $tstruct_pname{$item};
1143     foreach $i (@$mpname) {
1144       $wd = &word_texi($i);
1145       $title .= $wd unless $wd eq "\001";
1146     }
1147     print "* ${item}:: $title\n";
1148     $item = $tstruct_mnext{$item};
1149   }
1150   print "* Index::\n" if $topitem eq "Top";
1151   print "\@end menu\n";
1152 }
1153
1154 sub texi_index {
1155   my $itag, $ientry, @a, $wd, $item, $len;
1156   my $subnums = "123456789ABCDEFGHIJKLMNOPQRSTU" .
1157                 "VWXYZabcdefghijklmnopqrstuvwxyz";
1158
1159   print "\@ifinfo\n\@node Index, , $FIXMElastnode, Top\n";
1160   print "\@unnumbered Index\n\n\@menu\n";
1161
1162   foreach $itag (@itags) {
1163     $ientry = $idxmap{$itag};
1164     @a = @$ientry;
1165     $item = '';
1166     $len = 0;
1167     foreach $i (@a) {
1168       $wd = &word_texi($i);
1169       $item .= $wd, $len += $texiwdlen unless $wd eq "\001";
1170     }
1171     $i = 0;
1172     foreach $node (@nodes) {
1173       next if !$idxnodes{$node,$itag};
1174       printf "* %s%s (%s): %s.\n",
1175           $item, " " x (40-$len), substr($subnums,$i++,1), $node;
1176     }
1177   }
1178   print "\@end menu\n\@end ifinfo\n";
1179 }
1180
1181 sub write_hlp {
1182   # This is called from the top level, so I won't bother using
1183   # my or local.
1184
1185   # Build the index-tag text forms.
1186   print "building index entries...";
1187   @hlp_index = map {
1188                  my $i,$ww;
1189                  my $ientry = $idxmap{$_};
1190                  my $title = "";
1191                  foreach $i (@$ientry) {
1192                    $ww = &word_hlp($i,0);
1193                    $title .= $ww unless $ww eq "\001";
1194                  }
1195                  $title;
1196                } @itags;
1197
1198   # Write the HPJ project-description file.
1199   print "writing .hpj file...";
1200   open HPJ,">nasmdoc.hpj";
1201   print HPJ "[OPTIONS]\ncompress=true\n";
1202   print HPJ "title=NASM: The Netwide Assembler\noldkeyphrase=no\n\n";
1203   print HPJ "[FILES]\nnasmdoc.rtf\n\n";
1204   print HPJ "[CONFIG]\n";
1205   print HPJ 'CreateButton("btn_up", "&Up",'.
1206             ' "JumpContents(`nasmdoc.hlp'."'".')")';
1207   print HPJ "\nBrowseButtons()\n";
1208   close HPJ;
1209
1210   # Open file.
1211   print "\n   writing .rtf file...";
1212   open TEXT,">nasmdoc.rtf";
1213   select TEXT;
1214
1215   # Preamble.
1216   print "{\\rtf1\\ansi{\\fonttbl\n";
1217   print "\\f0\\froman Times New Roman;\\f1\\fmodern Courier New;\n";
1218   print "\\f2\\fswiss Arial;\\f3\\ftech Wingdings}\\deff0\n";
1219   print "#{\\footnote Top}\n";
1220   print "\${\\footnote Contents}\n";
1221   print "+{\\footnote browse:00000}\n";
1222   print "!{\\footnote DisableButton(\"btn_up\")}\n";
1223   print "\\keepn\\f2\\b\\fs30\\sb0\n";
1224   print "NASM: The Netwide Assembler\n";
1225   print "\\par\\pard\\plain\\sb120\n";
1226   print "This file documents NASM, the Netwide Assembler: an assembler \n";
1227   print "targetting the Intel x86 series of processors, with portable source.\n";
1228
1229   $node = "Top";
1230   $browse = 0;
1231
1232   $newpar = "\\par\\sb120\n";
1233   for ($para = 0; $para <= $#pnames; $para++) {
1234     $pname = $pnames[$para];
1235     $pflags = $pflags[$para];
1236     $ptype = substr($pflags,0,4);
1237
1238     print $newpar;
1239     $newpar = "\\par\\sb120\n";
1240
1241     if ($ptype eq "chap") {
1242       # Chapter heading. Begin a new node.
1243       &hlp_menu($node)
1244         if $tstruct_level{$tstruct_next{$node}} > $tstruct_level{$node};
1245       $pflags =~ /chap (.*) :(.*)/;
1246       $node = "Chapter $1";
1247       $title = $footnotetitle = "Chapter $1: ";
1248       foreach $i (@$pname) {
1249         $ww = &word_hlp($i,1);
1250         $title .= $ww, $footnotetitle .= &word_hlp($i,0) unless $ww eq "\001";
1251       }
1252       print "\\page\n";
1253       printf "#{\\footnote %s}\n", &hlp_sectkw($node);
1254       print "\${\\footnote $footnotetitle}\n";
1255       printf "+{\\footnote browse:%05d}\n", ++$browse;
1256       printf "!{\\footnote ChangeButtonBinding(\"btn_up\"," .
1257              "\"JumpId(\`nasmdoc.hlp',\`%s')\");\n",
1258              &hlp_sectkw($tstruct_up{$node});
1259       print "EnableButton(\"btn_up\")}\n";
1260       &hlp_keywords($node);
1261       print "\\keepn\\f2\\b\\fs30\\sb60\\sa60\n";
1262       print "$title\n";
1263       $newpar = "\\par\\pard\\plain\\sb120\n";
1264     } elsif ($ptype eq "appn") {
1265       # Appendix heading. Begin a new node.
1266       &hlp_menu($node)
1267         if $tstruct_level{$tstruct_next{$node}} > $tstruct_level{$node};
1268       $pflags =~ /appn (.*) :(.*)/;
1269       $node = "Appendix $1";
1270       $title = $footnotetitle = "Appendix $1: ";
1271       foreach $i (@$pname) {
1272         $ww = &word_hlp($i,1);
1273         $title .= $ww, $footnotetitle .= &word_hlp($i,0) unless $ww eq "\001";
1274       }
1275       print "\\page\n";
1276       printf "#{\\footnote %s}\n", &hlp_sectkw($node);
1277       print "\${\\footnote $footnotetitle}\n";
1278       printf "+{\\footnote browse:%05d}\n", ++$browse;
1279       printf "!{\\footnote ChangeButtonBinding(\"btn_up\"," .
1280              "\"JumpId(\`nasmdoc.hlp',\`%s')\");\n",
1281              &hlp_sectkw($tstruct_up{$node});
1282       print "EnableButton(\"btn_up\")}\n";
1283       &hlp_keywords($node);
1284       print "\\keepn\\f2\\b\\fs30\\sb60\\sa60\n";
1285       print "$title\n";
1286       $newpar = "\\par\\pard\\plain\\sb120\n";
1287     } elsif ($ptype eq "head" || $ptype eq "subh") {
1288       # Heading or subheading. Begin a new node.
1289       &hlp_menu($node)
1290         if $tstruct_level{$tstruct_next{$node}} > $tstruct_level{$node};
1291       $pflags =~ /.... (.*) :(.*)/;
1292       $node = "Section $1";
1293       $title = $footnotetitle = "$1. ";
1294       foreach $i (@$pname) {
1295         $ww = &word_hlp($i,1);
1296         $title .= $ww, $footnotetitle .= &word_hlp($i,0) unless $ww eq "\001";
1297       }
1298       print "\\page\n";
1299       printf "#{\\footnote %s}\n", &hlp_sectkw($node);
1300       print "\${\\footnote $footnotetitle}\n";
1301       printf "+{\\footnote browse:%05d}\n", ++$browse;
1302       printf "!{\\footnote ChangeButtonBinding(\"btn_up\"," .
1303              "\"JumpId(\`nasmdoc.hlp',\`%s')\");\n",
1304              &hlp_sectkw($tstruct_up{$node});
1305       print "EnableButton(\"btn_up\")}\n";
1306       &hlp_keywords($node);
1307       print "\\keepn\\f2\\b\\fs30\\sb60\\sa60\n";
1308       print "$title\n";
1309       $newpar = "\\par\\pard\\plain\\sb120\n";
1310     } elsif ($ptype eq "code") {
1311       # Code paragraph.
1312       print "\\keep\\f1\\sb120\n";
1313       foreach $i (@$pname) {
1314         my $x = $i;
1315         warn "code line longer than 68 chars: $i\n" if length $i > 68;
1316         $x =~ s/\\/\\\\/g;
1317         $x =~ s/\{/\\\{/g;
1318         $x =~ s/\}/\\\}/g;
1319         print "$x\\par\\sb0\n";
1320       }
1321       $newpar = "\\pard\\f0\\sb120\n";
1322     } elsif ($ptype eq "bull" || $ptype eq "norm") {
1323       # Ordinary paragraph, optionally bulleted. We wrap, FWIW.
1324       if ($ptype eq "bull") {
1325         print "\\tx360\\li360\\fi-360{\\f3\\'9F}\\tab\n";
1326         $newpar = "\\par\\pard\\sb120\n";
1327       } else {
1328         $newpar = "\\par\\sb120\n";
1329       }
1330       $line = '';
1331       @a = @$pname;
1332       $wd = $wprev = '';
1333       do {
1334         do { $w = &word_hlp((shift @a),1); } while $w eq "\001"; # hack
1335         $wd .= $wprev;
1336         if ($w eq ' ' || $w eq '' || $w eq undef) {
1337           if (length ($line . $wd) > 75) {
1338             $line =~ s/\s*$//; # trim trailing spaces
1339             print "$line \n"; # and put one back
1340             $line = '';
1341             $wd =~ s/^\s*//; # trim leading spaces
1342           }
1343           $line .= $wd;
1344           $wd = '';
1345         }
1346         $wprev = $w;
1347       } while ($w ne '' && $w ne undef);
1348       if ($line =~ /\S/) {
1349         $line =~ s/\s*$//; # trim trailing spaces
1350         print "$line\n";
1351       }
1352     }
1353   }
1354
1355   # Close file.
1356   print "\\page}\n";
1357   select STDOUT;
1358   close TEXT;
1359 }
1360
1361 sub word_hlp {
1362   my ($w, $docode) = @_;
1363   my $wtype, $wmajt;
1364
1365   return undef if $w eq '' || $w eq undef;
1366   $wtype = substr($w,0,2);
1367   $wmajt = substr($wtype,0,1);
1368   $w = substr($w,2);
1369   $w =~ s/\\/\\\\/g;
1370   $w =~ s/\{/\\\{/g;
1371   $w =~ s/\}/\\\}/g;
1372   $w =~ s/<.*>// if $wmajt eq "w"; # remove web links
1373   substr($w,0,length($w)-1) =~ s/-/\\\'AD/g if $wmajt ne "x"; #nonbreakhyphens
1374   if ($wmajt eq "n" || $wtype eq "e " || $wtype eq "w ") {
1375     return $w;
1376   } elsif ($wtype eq "sp") {
1377     return ' ';
1378   } elsif ($wtype eq "da") {
1379     return "\\'96";
1380   } elsif ($wmajt eq "c" || $wtype eq "wc") {
1381     $w =~ s/ /\\\'A0/g; # make spaces non-breaking
1382     return $docode ? "{\\f1 ${w}}" : $w;
1383   } elsif ($wtype eq "es") {
1384     return "{\\i ${w}";
1385   } elsif ($wtype eq "ee") {
1386     return "${w}}";
1387   } elsif ($wtype eq "eo") {
1388     return "{\\i ${w}}";
1389   } elsif ($wtype eq "x ") {
1390     return "{\\uldb ";
1391   } elsif ($wtype eq "xe") {
1392     $w = &hlp_sectkw($w);
1393     return "}{\\v ${w}}";
1394   } elsif ($wmajt eq "i") {
1395     return "\001";
1396   } else {
1397     die "panic in word_hlp: $wtype$w\n";
1398   }
1399 }
1400
1401 sub hlp_menu {
1402   my ($topitem) = @_;
1403   my $item, $kword, $i, $mpname, $title;
1404
1405   $item = $tstruct_next{$topitem};
1406   print "\\li360\\fi-360\n";
1407   while ($item) {
1408     $title = "";
1409     $mpname = $tstruct_pname{$item};
1410     foreach $i (@$mpname) {
1411       $ww = &word_hlp($i, 0);
1412       $title .= $ww unless $ww eq "\001";
1413     }
1414     $kword = &hlp_sectkw($item);
1415     print "{\\uldb ${item}: $title}{\\v $kword}\\par\\sb0\n";
1416     $item = $tstruct_mnext{$item};
1417   }
1418   print "\\pard\\sb120\n";
1419 }
1420
1421 sub hlp_sectkw {
1422   my ($node) = @_;
1423   $node =~ tr/A-Z/a-z/;
1424   $node =~ tr/- ./___/;
1425   $node;
1426 }
1427
1428 sub hlp_keywords {
1429   my ($node) = @_;
1430   my $pfx = "K{\\footnote ";
1431   my $done = 0;
1432   foreach $i (0..$#itags) {
1433     (print $pfx,$hlp_index[$i]), $pfx = ";\n", $done++
1434         if $idxnodes{$node,$itags[$i]};
1435   }
1436   print "}\n" if $done;
1437 }
1438
1439 # Make tree structures. $tstruct_* is top-level and global.
1440 sub add_item {
1441   my ($item, $level) = @_;
1442   my $i;
1443
1444   $tstruct_pname{$item} = $pname;
1445   $tstruct_next{$tstruct_previtem} = $item;
1446   $tstruct_prev{$item} = $tstruct_previtem;
1447   $tstruct_level{$item} = $level;
1448   $tstruct_up{$item} = $tstruct_last[$level-1];
1449   $tstruct_mnext{$tstruct_last[$level]} = $item;
1450   $tstruct_last[$level] = $item;
1451   for ($i=$level+1; $i<$MAXLEVEL; $i++) { $tstruct_last[$i] = undef; }
1452   $tstruct_previtem = $item;
1453   push @nodes, $item;
1454 }
1455
1456 #
1457 # This produces documentation intermediate paragraph format; this is
1458 # basically the digested output of the front end.  Intended for use
1459 # by future backends, instead of putting it all in the same script.
1460 #
1461 sub write_dip {
1462   open(PARAS, "> nasmdoc.dip");
1463   foreach $k (keys(%metadata)) {
1464       print PARAS 'meta :', $k, "\n";
1465       print PARAS $metadata{$k},"\n";
1466   }
1467   for ($para = 0; $para <= $#pnames; $para++) {
1468       print PARAS $pflags[$para], "\n";
1469       print PARAS join("\037", @{$pnames[$para]}, "\n");
1470   }
1471   foreach $k (@itags) {
1472       print PARAS 'indx :', $k, "\n";
1473       print PARAS join("\037", @{$idxmap{$k}}), "\n";
1474   }
1475   close(PARAS);
1476 }