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 # PS output:
   9 # - show page numbers in printed output
  10 # - think about double-sided support (start all chapters on RHS,
  11 #   ie odd-numbered, pages).
  12 #
  13 # Ellipsis support would be nice.
  14
  15 # Source-form features:
  16 # ---------------------
  17 #
  18 # Bullet \b
  19 #   Bullets the paragraph. Rest of paragraph is indented to cope. In
  20 #   HTML, consecutive groups of bulleted paragraphs become unordered
  21 #   lists.
  22 #
  23 # Emphasis \e{foobar}
  24 #   produces `_foobar_' in text and italics in HTML, PS, RTF
  25 #
  26 # Inline code \c{foobar}
  27 #   produces ``foobar'' in text, and fixed-pitch font in HTML, PS, RTF
  28 #
  29 # Display code
  30 # \c  line one
  31 # \c   line two
  32 #   produces fixed-pitch font where appropriate, and doesn't break
  33 #   pages except sufficiently far into the middle of a display.
  34 #
  35 # Chapter, header and subheader
  36 # \C{intro} Introduction
  37 # \H{whatsnasm} What is NASM?
  38 # \S{free} NASM Is Free
  39 #   dealt with as appropriate. Chapters begin on new sides, possibly
  40 #   even new _pages_. (Sub)?headers are good places to begin new
  41 #   pages. Just _after_ a (sub)?header isn't.
  42 #   The keywords can be substituted with \K and \k.
  43 #
  44 # Keyword \K{cintro} \k{cintro}
  45 #   Expands to `Chapter 1', `Section 1.1', `Section 1.1.1'. \K has an
  46 #   initial capital whereas \k doesn't. In HTML, will produce
  47 #   hyperlinks.
  48 #
  49 # Web link \W{http://foobar/}{text} or \W{mailto:me@here}\c{me@here}
  50 #   the \W prefix is ignored except in HTML; in HTML the last part
  51 #   becomes a hyperlink to the first part.
  52 #
  53 # Literals \{ \} \\
  54 #   In case it's necessary, they expand to the real versions.
  55 #
  56 # Nonbreaking hyphen \-
  57 #   Need more be said?
  58 #
  59 # Source comment \#
  60 #   Causes everything after it on the line to be ignored by the
  61 #   source-form processor.
  62 #
  63 # Indexable word \i{foobar} (or \i\e{foobar} or \i\c{foobar}, equally)
  64 #   makes word appear in index, referenced to that point
  65 #   \i\c comes up in code style even in the index; \i\e doesn't come
  66 #   up in emphasised style.
  67 #
  68 # Indexable non-displayed word \I{foobar} or \I\c{foobar}
  69 #   just as \i{foobar} except that nothing is displayed for it
  70 #
  71 # Index rewrite
  72 # \IR{foobar} \c{foobar} operator, uses of
  73 #   tidies up the appearance in the index of something the \i or \I
  74 #   operator was applied to
  75 #
  76 # Index alias
  77 # \IA{foobar}{bazquux}
  78 #   aliases one index tag (as might be supplied to \i or \I) to
  79 #   another, so that \I{foobar} has the effect of \I{bazquux}, and
  80 #   \i{foobar} has the effect of \I{bazquux}foobar
  81
  82 $diag = 1, shift @ARGV if $ARGV[0] eq "-d";
  83
  84 $| = 1;
  85
  86 $tstruct_previtem = $node = "Top";
  87 $nodes = ($node);
  88 $tstruct_level{$tstruct_previtem} = 0;
  89 $tstruct_last[$tstruct_level{$tstruct_previtem}] = $tstruct_previtem;
  90 $MAXLEVEL = 10;  # really 3, but play safe ;-)
  91
  92 # Read the file; pass a paragraph at a time to the paragraph processor.
  93 print "Reading input...";
  94 $pname = "para000000";
  95 @pnames = @pflags = ();
  96 $para = undef;
  97 while (<>) {
  98   chomp;
  99   if (!/\S/ || /^\\I[AR]/) { # special case: \I[AR] implies new-paragraph
 100     &got_para($para);
 101     $para = undef;
 102   }
 103   if (/\S/) {
 104     s/\\#.*$//; # strip comments
 105     $para .= " " . $_;
 106   }
 107 }
 108 &got_para($para);
 109 print "done.\n";
 110
 111 # Now we've read in the entire document and we know what all the
 112 # heading keywords refer to. Go through and fix up the \k references.
 113 print "Fixing up cross-references...";
 114 &fixup_xrefs;
 115 print "done.\n";
 116
 117 # Sort the index tags, according to the slightly odd order I've decided on.
 118 print "Sorting index tags...";
 119 &indexsort;
 120 print "done.\n";
 121
 122 if ($diag) {
 123   print "Writing index-diagnostic file...";
 124   &indexdiag;
 125   print "done.\n";
 126 }
 127
 128 # OK. Write out the various output files.
 129 print "Producing text output: ";
 130 &write_txt;
 131 print "done.\n";
 132 print "Producing HTML output: ";
 133 &write_html;
 134 print "done.\n";
 135 print "Producing PostScript output: ";
 136 &write_ps;
 137 print "done.\n";
 138 print "Producing Texinfo output: ";
 139 &write_texi;
 140 print "done.\n";
 141 print "Producing WinHelp output: ";
 142 &write_hlp;
 143 print "done.\n";
 144
 145 sub got_para {
 146   local ($_) = @_;
 147   my $pflags = "", $i, $w, $l, $t;
 148   return if !/\S/;
 149
 150   @$pname = ();
 151
 152   # Strip off _leading_ spaces, then determine type of paragraph.
 153   s/^\s*//;
 154   $irewrite = undef;
 155   if (/^\\c[^{]/) {
 156     # A code paragraph. The paragraph-array will contain the simple
 157     # strings which form each line of the paragraph.
 158     $pflags = "code";
 159     while (/^\\c (([^\\]|\\[^c])*)(.*)$/) {
 160       $l = $1;
 161       $_ = $3;
 162       $l =~ s/\\{/{/g;
 163       $l =~ s/\\}/}/g;
 164       $l =~ s/\\\\/\\/g;
 165       push @$pname, $l;
 166     }
 167     $_ = ''; # suppress word-by-word code
 168   } elsif (/^\\C/) {
 169     # A chapter heading. Define the keyword and allocate a chapter
 170     # number.
 171     $cnum++;
 172     $hnum = 0;
 173     $snum = 0;
 174     $xref = "chapter-$cnum";
 175     $pflags = "chap $cnum :$xref";
 176     die "badly formatted chapter heading: $_\n" if !/^\\C{([^}]*)}\s*(.*)$/;
 177     $refs{$1} = "chapter $cnum";
 178     $node = "Chapter $cnum";
 179     &add_item($node, 1);
 180     $xrefnodes{$node} = $xref; $nodexrefs{$xref} = $node;
 181     $xrefs{$1} = $xref;
 182     $_ = $2;
 183     # the standard word-by-word code will happen next
 184   } elsif (/^\\A/) {
 185     # An appendix heading. Define the keyword and allocate an appendix
 186     # letter.
 187     $cnum++;
 188     $cnum = 'A' if $cnum =~ /[0-9]+/;
 189     $hnum = 0;
 190     $snum = 0;
 191     $xref = "appendix-$cnum";
 192     $pflags = "appn $cnum :$xref";
 193     die "badly formatted appendix heading: $_\n" if !/^\\A{([^}]*)}\s*(.*)$/;
 194     $refs{$1} = "appendix $cnum";
 195     $node = "Appendix $cnum";
 196     &add_item($node, 1);
 197     $xrefnodes{$node} = $xref; $nodexrefs{$xref} = $node;
 198     $xrefs{$1} = $xref;
 199     $_ = $2;
 200     # the standard word-by-word code will happen next
 201   } elsif (/^\\H/) {
 202     # A major heading. Define the keyword and allocate a section number.
 203     $hnum++;
 204     $snum = 0;
 205     $xref = "section-$cnum.$hnum";
 206     $pflags = "head $cnum.$hnum :$xref";
 207     die "badly formatted heading: $_\n" if !/^\\[HP]{([^}]*)}\s*(.*)$/;
 208     $refs{$1} = "section $cnum.$hnum";
 209     $node = "Section $cnum.$hnum";
 210     &add_item($node, 2);
 211     $xrefnodes{$node} = $xref; $nodexrefs{$xref} = $node;
 212     $xrefs{$1} = $xref;
 213     $_ = $2;
 214     # the standard word-by-word code will happen next
 215   } elsif (/^\\S/) {
 216     # A sub-heading. Define the keyword and allocate a section number.
 217     $snum++;
 218     $xref = "section-$cnum.$hnum.$snum";
 219     $pflags = "subh $cnum.$hnum.$snum :$xref";
 220     die "badly formatted subheading: $_\n" if !/^\\S{([^}]*)}\s*(.*)$/;
 221     $refs{$1} = "section $cnum.$hnum.$snum";
 222     $node = "Section $cnum.$hnum.$snum";
 223     &add_item($node, 3);
 224     $xrefnodes{$node} = $xref; $nodexrefs{$xref} = $node;
 225     $xrefs{$1} = $xref;
 226     $_ = $2;
 227     # the standard word-by-word code will happen next
 228   } elsif (/^\\IR/) {
 229     # An index-rewrite.
 230     die "badly formatted index rewrite: $_\n" if !/^\\IR{([^}]*)}\s*(.*)$/;
 231     $irewrite = $1;
 232     $_ = $2;
 233     # the standard word-by-word code will happen next
 234   } elsif (/^\\IA/) {
 235     # An index-alias.
 236     die "badly formatted index alias: $_\n" if !/^\\IA{([^}]*)}{([^}]*)}\s*$/;
 237     $idxalias{$1} = $2;
 238     return; # avoid word-by-word code
 239   } elsif (/^\\b/) {
 240     # A bulleted paragraph. Strip off the initial \b and let the
 241     # word-by-word code take care of the rest.
 242     $pflags = "bull";
 243     s/^\\b\s*//;
 244   } else {
 245     # A normal paragraph. Just set $pflags: the word-by-word code does
 246     # the rest.
 247     $pflags = "norm";
 248   }
 249
 250   # The word-by-word code: unless @$pname is already defined (which it
 251   # will be in the case of a code paragraph), split the paragraph up
 252   # into words and push each on @$pname.
 253   #
 254   # Each thing pushed on @$pname should have a two-character type
 255   # code followed by the text.
 256   #
 257   # Type codes are:
 258   # "n " for normal
 259   # "da" for a dash
 260   # "es" for first emphasised word in emphasised bit
 261   # "e " for emphasised in mid-emphasised-bit
 262   # "ee" for last emphasised word in emphasised bit
 263   # "eo" for single (only) emphasised word
 264   # "c " for code
 265   # "k " for cross-ref
 266   # "kK" for capitalised cross-ref
 267   # "w " for Web link
 268   # "wc" for code-type Web link
 269   # "x " for beginning of resolved cross-ref; generates no visible output,
 270   #      and the text is the cross-reference code
 271   # "xe" for end of resolved cross-ref; text is same as for "x ".
 272   # "i " for point to be indexed: the text is the internal index into the
 273   #      index-items arrays
 274   # "sp" for space
 275   while (/\S/) {
 276     s/^\s*//, push @$pname, "sp" if /^\s/;
 277     $indexing = $qindex = 0;
 278     if (/^(\\[iI])?\\c/) {
 279       $qindex = 1 if $1 eq "\\I";
 280       $indexing = 1, s/^\\[iI]// if $1;
 281       s/^\\c//;
 282       die "badly formatted \\c: \\c$_\n" if !/{(([^\\}]|\\.)*)}(.*)$/;
 283       $w = $1;
 284       $_ = $3;
 285       $w =~ s/\\{/{/g;
 286       $w =~ s/\\}/}/g;
 287       $w =~ s/\\-/-/g;
 288       $w =~ s/\\\\/\\/g;
 289       (push @$pname,"i"),$lastp = $#$pname if $indexing;
 290       push @$pname,"c $w" if !$qindex;
 291       $$pname[$lastp] = &addidx($node, $w, "c $w") if $indexing;
 292     } elsif (/^\\[iIe]/) {
 293       /^(\\[iI])?(\\e)?/;
 294       $emph = 0;
 295       $qindex = 1 if $1 eq "\\I";
 296       $indexing = 1, $type = "\\i" if $1;
 297       $emph = 1, $type = "\\e" if $2;
 298       s/^(\\[iI])?(\\e?)//;
 299       die "badly formatted $type: $type$_\n" if !/{(([^\\}]|\\.)*)}(.*)$/;
 300       $w = $1;
 301       $_ = $3;
 302       $w =~ s/\\{/{/g;
 303       $w =~ s/\\}/}/g;
 304       $w =~ s/\\-/-/g;
 305       $w =~ s/\\\\/\\/g;
 306       $t = $emph ? "es" : "n ";
 307       @ientry = ();
 308       (push @$pname,"i"),$lastp = $#$pname if $indexing;
 309       foreach $i (split /\s+/,$w) {  # \e and \i can be multiple words
 310         push @$pname,"$t$i","sp" if !$qindex;
 311         ($ii=$i) =~ tr/A-Z/a-z/, push @ientry,"n $ii","sp" if $indexing;
 312         $t = $emph ? "e " : "n ";
 313       }
 314       $w =~ tr/A-Z/a-z/, pop @ientry if $indexing;
 315       $$pname[$lastp] = &addidx($node, $w, @ientry) if $indexing;
 316       pop @$pname if !$qindex; # remove final space
 317       if (substr($$pname[$#$pname],0,2) eq "es" && !$qindex) {
 318         substr($$pname[$#$pname],0,2) = "eo";
 319       } elsif ($emph && !$qindex) {
 320         substr($$pname[$#$pname],0,2) = "ee";
 321       }
 322     } elsif (/^\\[kK]/) {
 323       $t = "k ";
 324       $t = "kK" if /^\\K/;
 325       s/^\\[kK]//;
 326       die "badly formatted \\k: \\c$_\n" if !/{([^}]*)}(.*)$/;
 327       $_ = $2;
 328       push @$pname,"$t$1";
 329     } elsif (/^\\W/) {
 330       s/^\\W//;
 331       die "badly formatted \\W: \\W$_\n"
 332           if !/{([^}]*)}(\\i)?(\\c)?{(([^\\}]|\\.)*)}(.*)$/;
 333       $l = $1;
 334       $w = $4;
 335       $_ = $6;
 336       $t = "w ";
 337       $t = "wc" if $3 eq "\\c";
 338       $indexing = 1 if $2;
 339       $w =~ s/\\{/{/g;
 340       $w =~ s/\\}/}/g;
 341       $w =~ s/\\-/-/g;
 342       $w =~ s/\\\\/\\/g;
 343       (push @$pname,"i"),$lastp = $#$pname if $indexing;
 344       push @$pname,"$t<$l>$w";
 345       $$pname[$lastp] = &addidx($node, $w, "c $w") if $indexing;
 346     } else {
 347       die "what the hell? $_\n" if !/^(([^\s\\\-]|\\[\\{}\-])*-?)(.*)$/;
 348       die "painful death! $_\n" if !length $1;
 349       $w = $1;
 350       $_ = $3;
 351       $w =~ s/\\{/{/g;
 352       $w =~ s/\\}/}/g;
 353       $w =~ s/\\-/-/g;
 354       $w =~ s/\\\\/\\/g;
 355       if ($w eq "-") {
 356         push @$pname,"da";
 357       } else {
 358         push @$pname,"n $w";
 359       }
 360     }
 361   }
 362   if ($irewrite ne undef) {
 363     &addidx(undef, $irewrite, @$pname);
 364     @$pname = ();
 365   } else {
 366     push @pnames, $pname;
 367     push @pflags, $pflags;
 368     $pname++;
 369   }
 370 }
 371
 372 sub addidx {
 373   my ($node, $text, @ientry) = @_;
 374   $text = $idxalias{$text} || $text;
 375   if ($node eq undef || !$idxmap{$text}) {
 376     @$ientry = @ientry;
 377     $idxmap{$text} = $ientry;
 378     $ientry++;
 379   }
 380   if ($node) {
 381     $idxnodes{$node,$text} = 1;
 382     return "i $text";
 383   }
 384 }
 385
 386 sub indexsort {
 387   my $iitem, $ientry, $i, $piitem, $pcval, $cval, $clrcval;
 388
 389   @itags = map { # get back the original data as the 1st elt of each list
 390              $_->[0]
 391            } sort { # compare auxiliary (non-first) elements of lists
 392              $a->[1] cmp $b->[1] ||
 393              $a->[2] cmp $b->[2] ||
 394              $a->[0] cmp $b->[0]
 395            } map { # transform array into list of 3-element lists
 396              my $ientry = $idxmap{$_};
 397              my $a = substr($$ientry[0],2);
 398              $a =~ tr/A-Za-z//cd;
 399              [$_, uc($a), substr($$ientry[0],0,2)]
 400            } keys %idxmap;
 401
 402   # Having done that, check for comma-hood.
 403   $cval = 0;
 404   foreach $iitem (@itags) {
 405     $ientry = $idxmap{$iitem};
 406     $clrcval = 1;
 407     $pcval = $cval;
 408     FL:for ($i=0; $i <= $#$ientry; $i++) {
 409       if ($$ientry[$i] =~ /^(n .*,)(.*)/) {
 410         $$ientry[$i] = $1;
 411         splice @$ientry,$i+1,0,"n $2" if length $2;
 412         $commapos{$iitem} = $i+1;
 413         $cval = join("\002", @$ientry[0..$i]);
 414         $clrcval = 0;
 415         last FL;
 416       }
 417     }
 418     $cval = undef if $clrcval;
 419     $commanext{$iitem} = $commaafter{$piitem} = 1
 420       if $cval and ($cval eq $pcval);
 421     $piitem = $iitem;
 422   }
 423 }
 424
 425 sub indexdiag {
 426   my $iitem,$ientry,$w,$ww,$foo,$node;
 427   open INDEXDIAG,">index.diag";
 428   foreach $iitem (@itags) {
 429     $ientry = $idxmap{$iitem};
 430     print INDEXDIAG "<$iitem> ";
 431     foreach $w (@$ientry) {
 432       $ww = &word_txt($w);
 433       print INDEXDIAG $ww unless $ww eq "\001";
 434     }
 435     print INDEXDIAG ":";
 436     $foo = " ";
 437     foreach $node (@nodes) {
 438       (print INDEXDIAG $foo,$node), $foo = ", " if $idxnodes{$node,$iitem};
 439     }
 440     print INDEXDIAG "\n";
 441   }
 442   close INDEXDIAG;
 443 }
 444
 445 sub fixup_xrefs {
 446   my $pname, $p, $i, $j, $k, $caps, @repl;
 447
 448   for ($p=0; $p<=$#pnames; $p++) {
 449     next if $pflags[$p] eq "code";
 450     $pname = $pnames[$p];
 451     for ($i=$#$pname; $i >= 0; $i--) {
 452       if ($$pname[$i] =~ /^k/) {
 453         $k = $$pname[$i];
 454         $caps = ($k =~ /^kK/);
 455         $k = substr($k,2);
 456         $repl = $refs{$k};
 457         die "undefined keyword `$k'\n" unless $repl;
 458         substr($repl,0,1) =~ tr/a-z/A-Z/ if $caps;
 459         @repl = ();
 460         push @repl,"x $xrefs{$k}";
 461         foreach $j (split /\s+/,$repl) {
 462           push @repl,"n $j";
 463           push @repl,"sp";
 464         }
 465         pop @repl; # remove final space
 466         push @repl,"xe$xrefs{$k}";
 467         splice @$pname,$i,1,@repl;
 468       }
 469     }
 470   }
 471 }
 472
 473 sub write_txt {
 474   # This is called from the top level, so I won't bother using
 475   # my or local.
 476
 477   # Open file.
 478   print "writing file...";
 479   open TEXT,">nasmdoc.txt";
 480   select TEXT;
 481
 482   # Preamble.
 483   $title = "The Netwide Assembler: NASM";
 484   $spaces = ' ' x ((75-(length $title))/2);
 485   ($underscore = $title) =~ s/./=/g;
 486   print "$spaces$title\n$spaces$underscore\n";
 487
 488   for ($para = 0; $para <= $#pnames; $para++) {
 489     $pname = $pnames[$para];
 490     $pflags = $pflags[$para];
 491     $ptype = substr($pflags,0,4);
 492
 493     print "\n"; # always one of these before a new paragraph
 494
 495     if ($ptype eq "chap") {
 496       # Chapter heading. "Chapter N: Title" followed by a line of
 497       # minus signs.
 498       $pflags =~ /chap (.*) :(.*)/;
 499       $title = "Chapter $1: ";
 500       foreach $i (@$pname) {
 501         $ww = &word_txt($i);
 502         $title .= $ww unless $ww eq "\001";
 503       }
 504       print "$title\n";
 505       $title =~ s/./-/g;
 506       print "$title\n";
 507     } elsif ($ptype eq "appn") {
 508       # Appendix heading. "Appendix N: Title" followed by a line of
 509       # minus signs.
 510       $pflags =~ /appn (.*) :(.*)/;
 511       $title = "Appendix $1: ";
 512       foreach $i (@$pname) {
 513         $ww = &word_txt($i);
 514         $title .= $ww unless $ww eq "\001";
 515       }
 516       print "$title\n";
 517       $title =~ s/./-/g;
 518       print "$title\n";
 519     } elsif ($ptype eq "head" || $ptype eq "subh") {
 520       # Heading or subheading. Just a number and some text.
 521       $pflags =~ /.... (.*) :(.*)/;
 522       $title = sprintf "%6s ", $1;
 523       foreach $i (@$pname) {
 524         $ww = &word_txt($i);
 525         $title .= $ww unless $ww eq "\001";
 526       }
 527       print "$title\n";
 528     } elsif ($ptype eq "code") {
 529       # Code paragraph. Emit each line with a seven character indent.
 530       foreach $i (@$pname) {
 531         warn "code line longer than 68 chars: $i\n" if length $i > 68;
 532         print ' 'x7, $i, "\n";
 533       }
 534     } elsif ($ptype eq "bull" || $ptype eq "norm") {
 535       # Ordinary paragraph, optionally bulleted. We wrap, with ragged
 536       # 75-char right margin and either 7 or 11 char left margin
 537       # depending on bullets.
 538       if ($ptype eq "bull") {
 539         $line = ' 'x7 . '(*) ';
 540         $next = ' 'x11;
 541       } else {
 542         $line = $next = ' 'x7;
 543       }
 544       @a = @$pname;
 545       $wd = $wprev = '';
 546       do {
 547         do { $w = &word_txt(shift @a) } while $w eq "\001"; # nasty hack
 548         $wd .= $wprev;
 549         if ($wprev =~ /-$/ || $w eq ' ' || $w eq '' || $w eq undef) {
 550           if (length ($line . $wd) > 75) {
 551             $line =~ s/\s*$//; # trim trailing spaces
 552             print "$line\n";
 553             $line = $next;
 554             $wd =~ s/^\s*//; # trim leading spaces
 555           }
 556           $line .= $wd;
 557           $wd = '';
 558         }
 559         $wprev = $w;
 560       } while ($w ne '' && $w ne undef);
 561       if ($line =~ /\S/) {
 562         $line =~ s/\s*$//; # trim trailing spaces
 563         print "$line\n";
 564       }
 565     }
 566   }
 567
 568   # Close file.
 569   select STDOUT;
 570   close TEXT;
 571 }
 572
 573 sub word_txt {
 574   my ($w) = @_;
 575   my $wtype, $wmajt;
 576
 577   return undef if $w eq '' || $w eq undef;
 578   $wtype = substr($w,0,2);
 579   $wmajt = substr($wtype,0,1);
 580   $w = substr($w,2);
 581   $w =~ s/<.*>// if $wmajt eq "w"; # remove web links
 582   if ($wmajt eq "n" || $wtype eq "e " || $wtype eq "w ") {
 583     return $w;
 584   } elsif ($wtype eq "sp") {
 585     return ' ';
 586   } elsif ($wtype eq "da") {
 587     return '-';
 588   } elsif ($wmajt eq "c" || $wtype eq "wc") {
 589     return "`${w}'";
 590   } elsif ($wtype eq "es") {
 591     return "_${w}";
 592   } elsif ($wtype eq "ee") {
 593     return "${w}_";
 594   } elsif ($wtype eq "eo") {
 595     return "_${w}_";
 596   } elsif ($wmajt eq "x" || $wmajt eq "i") {
 597     return "\001";
 598   } else {
 599     die "panic in word_txt: $wtype$w\n";
 600   }
 601 }
 602
 603 sub write_html {
 604   # This is called from the top level, so I won't bother using
 605   # my or local.
 606
 607   # Write contents file. Just the preamble, then a menu of links to the
 608   # separate chapter files and the nodes therein.
 609   print "writing contents file...";
 610   open TEXT,">nasmdoc0.html";
 611   select TEXT;
 612   &html_preamble(0);
 613   print "<p>This manual documents NASM, the Netwide Assembler: an assembler\n";
 614   print "targetting the Intel x86 series of processors, with portable source.\n";
 615   print "<p>";
 616   for ($node = $tstruct_next{'Top'}; $node; $node = $tstruct_next{$node}) {
 617     if ($tstruct_level{$node} == 1) {
 618       # Invent a file name.
 619       ($number = lc($xrefnodes{$node})) =~ s/.*-//;
 620       $fname="nasmdocx.html";
 621       substr($fname,8 - length $number, length $number) = $number;
 622       $html_fnames{$node} = $fname;
 623       $link = $fname;
 624       print "<p>";
 625     } else {
 626       # Use the preceding filename plus a marker point.
 627       $link = $fname . "#$xrefnodes{$node}";
 628     }
 629     $title = "$node: ";
 630     $pname = $tstruct_pname{$node};
 631     foreach $i (@$pname) {
 632       $ww = &word_html($i);
 633       $title .= $ww unless $ww eq "\001";
 634     }
 635     print "<a href=\"$link\">$title</a><br>\n";
 636   }
 637   print "<p><a href=\"nasmdoci.html\">Index</a>\n";
 638   print "</body></html>\n";
 639   select STDOUT;
 640   close TEXT;
 641
 642   # Open a null file, to ensure output (eg random &html_jumppoints calls)
 643   # goes _somewhere_.
 644   print "writing chapter files...";
 645   open TEXT,">/dev/null";
 646   select TEXT;
 647   $html_lastf = '';
 648
 649   $in_list = 0;
 650
 651   for ($para = 0; $para <= $#pnames; $para++) {
 652     $pname = $pnames[$para];
 653     $pflags = $pflags[$para];
 654     $ptype = substr($pflags,0,4);
 655
 656     $in_list = 0, print "</ul>\n" if $in_list && $ptype ne "bull";
 657     if ($ptype eq "chap") {
 658       # Chapter heading. Begin a new file.
 659       $pflags =~ /chap (.*) :(.*)/;
 660       $title = "Chapter $1: ";
 661       $xref = $2;
 662       &html_jumppoints; print "</body></html>\n"; select STDOUT; close TEXT;
 663       $html_lastf = $html_fnames{$chapternode};
 664       $chapternode = $nodexrefs{$xref};
 665       $html_nextf = $html_fnames{$tstruct_mnext{$chapternode}};
 666       open TEXT,">$html_fnames{$chapternode}"; select TEXT; &html_preamble(1);
 667       foreach $i (@$pname) {
 668         $ww = &word_html($i);
 669         $title .= $ww unless $ww eq "\001";
 670       }
 671       $h = "<h2><a name=\"$xref\">$title</a></h2>\n";
 672       print $h; print FULL $h;
 673     } elsif ($ptype eq "appn") {
 674       # Appendix heading. Begin a new file.
 675       $pflags =~ /appn (.*) :(.*)/;
 676       $title = "Appendix $1: ";
 677       $xref = $2;
 678       &html_jumppoints; print "</body></html>\n"; select STDOUT; close TEXT;
 679       $html_lastf = $html_fnames{$chapternode};
 680       $chapternode = $nodexrefs{$xref};
 681       $html_nextf = $html_fnames{$tstruct_mnext{$chapternode}};
 682       open TEXT,">$html_fnames{$chapternode}"; select TEXT; &html_preamble(1);
 683       foreach $i (@$pname) {
 684         $ww = &word_html($i);
 685         $title .= $ww unless $ww eq "\001";
 686       }
 687       print "<h2><a name=\"$xref\">$title</a></h2>\n";
 688     } elsif ($ptype eq "head" || $ptype eq "subh") {
 689       # Heading or subheading.
 690       $pflags =~ /.... (.*) :(.*)/;
 691       $hdr = ($ptype eq "subh" ? "h4" : "h3");
 692       $title = $1 . " ";
 693       $xref = $2;
 694       foreach $i (@$pname) {
 695         $ww = &word_html($i);
 696         $title .= $ww unless $ww eq "\001";
 697       }
 698       print "<$hdr><a name=\"$xref\">$title</a></$hdr>\n";
 699     } elsif ($ptype eq "code") {
 700       # Code paragraph.
 701       print "<p><pre>\n";
 702       foreach $i (@$pname) {
 703         $w = $i;
 704         $w =~ s/&/&amp;/g;
 705         $w =~ s/</&lt;/g;
 706         $w =~ s/>/&gt;/g;
 707         print $w, "\n";
 708       }
 709       print "</pre>\n";
 710     } elsif ($ptype eq "bull" || $ptype eq "norm") {
 711       # Ordinary paragraph, optionally bulleted. We wrap, with ragged
 712       # 75-char right margin and either 7 or 11 char left margin
 713       # depending on bullets.
 714       if ($ptype eq "bull") {
 715         $in_list = 1, print "<ul>\n" unless $in_list;
 716         $line = '<li>';
 717       } else {
 718         $line = '<p>';
 719       }
 720       @a = @$pname;
 721       $wd = $wprev = '';
 722       do {
 723         do { $w = &word_html(shift @a) } while $w eq "\001"; # nasty hack
 724         $wd .= $wprev;
 725         if ($w eq ' ' || $w eq '' || $w eq undef) {
 726           if (length ($line . $wd) > 75) {
 727             $line =~ s/\s*$//; # trim trailing spaces
 728             print "$line\n";
 729             $line = '';
 730             $wd =~ s/^\s*//; # trim leading spaces
 731           }
 732           $line .= $wd;
 733           $wd = '';
 734         }
 735         $wprev = $w;
 736       } while ($w ne '' && $w ne undef);
 737       if ($line =~ /\S/) {
 738         $line =~ s/\s*$//; # trim trailing spaces
 739         print "$line\n";
 740       }
 741     }
 742   }
 743
 744   # Close whichever file was open.
 745   &html_jumppoints;
 746   print "</body></html>\n";
 747   select STDOUT;
 748   close TEXT;
 749
 750   print "\n   writing index file...";
 751   open TEXT,">nasmdoci.html";
 752   select TEXT;
 753   &html_preamble(0);
 754   print "<p align=center><a href=\"nasmdoc0.html\">Contents</a>\n";
 755   print "<p>";
 756   &html_index;
 757   print "<p align=center><a href=\"nasmdoc0.html\">Contents</a>\n";
 758   print "</body></html>\n";
 759   select STDOUT;
 760   close TEXT;
 761 }
 762
 763 sub html_preamble {
 764   print "<html><head><title>NASM Manual</title></head>\n";
 765   print "<body><h1 align=center>The Netwide Assembler: NASM</h1>\n\n";
 766   &html_jumppoints if $_[0];
 767 }
 768
 769 sub html_jumppoints {
 770   print "<p align=center>";
 771   print "<a href=\"$html_nextf\">Next Chapter</a> |\n" if $html_nextf;
 772   print "<a href=\"$html_lastf\">Previous Chapter</a> |\n" if $html_lastf;
 773   print "<a href=\"nasmdoc0.html\">Contents</a> |\n";
 774   print "<a href=\"nasmdoci.html\">Index</a>\n";
 775 }
 776
 777 sub html_index {
 778   my $itag, $a, @ientry, $sep, $w, $wd, $wprev, $line;
 779
 780   $chapternode = '';
 781   foreach $itag (@itags) {
 782     $ientry = $idxmap{$itag};
 783     @a = @$ientry;
 784     push @a, "n :";
 785     $sep = 0;
 786     foreach $node (@nodes) {
 787       next if !$idxnodes{$node,$itag};
 788       push @a, "n ," if $sep;
 789       push @a, "sp", "x $xrefnodes{$node}", "n $node", "xe$xrefnodes{$node}";
 790       $sep = 1;
 791     }
 792     $line = '';
 793     do {
 794       do { $w = &word_html(shift @a) } while $w eq "\001"; # nasty hack
 795       $wd .= $wprev;
 796       if ($w eq ' ' || $w eq '' || $w eq undef) {
 797         if (length ($line . $wd) > 75) {
 798           $line =~ s/\s*$//; # trim trailing spaces
 799           print "$line\n";
 800           $line = '';
 801           $wd =~ s/^\s*//; # trim leading spaces
 802         }
 803         $line .= $wd;
 804         $wd = '';
 805       }
 806       $wprev = $w;
 807     } while ($w ne '' && $w ne undef);
 808     if ($line =~ /\S/) {
 809       $line =~ s/\s*$//; # trim trailing spaces
 810       print "$line\n";
 811     }
 812     print "<br>\n";
 813   }
 814 }
 815
 816 sub word_html {
 817   my ($w) = @_;
 818   my $wtype, $wmajt, $pfx, $sfx;
 819
 820   return undef if $w eq '' || $w eq undef;
 821
 822   $wtype = substr($w,0,2);
 823   $wmajt = substr($wtype,0,1);
 824   $w = substr($w,2);
 825   $pfx = $sfx = '';
 826   $pfx = "<a href=\"$1\">", $sfx = "</a>", $w = $2
 827     if $wmajt eq "w" && $w =~ /^<(.*)>(.*)$/;
 828   $w =~ s/&/&amp;/g;
 829   $w =~ s/</&lt;/g;
 830   $w =~ s/>/&gt;/g;
 831   if ($wmajt eq "n" || $wtype eq "e " || $wtype eq "w ") {
 832     return $pfx . $w . $sfx;
 833   } elsif ($wtype eq "sp") {
 834     return ' ';
 835   } elsif ($wtype eq "da") {
 836     return '-'; # sadly, en-dashes are non-standard in HTML
 837   } elsif ($wmajt eq "c" || $wtype eq "wc") {
 838     return $pfx . "<code><nobr>${w}</nobr></code>" . $sfx;
 839   } elsif ($wtype eq "es") {
 840     return "<em>${w}";
 841   } elsif ($wtype eq "ee") {
 842     return "${w}</em>";
 843   } elsif ($wtype eq "eo") {
 844     return "<em>${w}</em>";
 845   } elsif ($wtype eq "x ") {
 846     # Magic: we must resolve the cross reference into file and marker
 847     # parts, then dispose of the file part if it's us, and dispose of
 848     # the marker part if the cross reference describes the top node of
 849     # another file.
 850     my $node = $nodexrefs{$w}; # find the node we're aiming at
 851     my $level = $tstruct_level{$node}; # and its level
 852     my $up = $node, $uplev = $level-1;
 853     $up = $tstruct_up{$up} while $uplev--; # get top node of containing file
 854     my $file = ($up ne $chapternode) ? $html_fnames{$up} : "";
 855     my $marker = ($level == 1 and $file) ? "" : "#$w";
 856     return "<a href=\"$file$marker\">";
 857   } elsif ($wtype eq "xe") {
 858     return "</a>";
 859   } elsif ($wmajt eq "i") {
 860     return "\001";
 861   } else {
 862     die "panic in word_html: $wtype$w\n";
 863   }
 864 }
 865
 866 sub write_ps {
 867   # This is called from the top level, so I won't bother using
 868   # my or local.
 869
 870   # First, set up the font metric arrays.
 871   &font_metrics;
 872
 873   # First stage: reprocess the source arrays into a list of
 874   # lines, each of which is a list of word-strings, each of
 875   # which has a single-letter font code followed by text.
 876   # Each line also has an associated type, which will be
 877   # used for final alignment and font selection and things.
 878   #
 879   # Font codes are:
 880   #   n == Normal
 881   #   e == Emphasised
 882   #   c == Code
 883   #  ' ' == space (no following text required)
 884   #  '-' == dash (no following text required)
 885   #
 886   # Line types are:
 887   #   chap == Chapter or appendix heading.
 888   #   head == Major heading.
 889   #   subh == Sub-heading.
 890   #   Ccha == Contents entry for a chapter.
 891   #   Chea == Contents entry for a heading.
 892   #   Csub == Contents entry for a subheading.
 893   #   cone == Code paragraph with just this one line on it.
 894   #   cbeg == First line of multi-line code paragraph.
 895   #   cbdy == Interior line of multi-line code paragraph.
 896   #   cend == Final line of multi-line code paragraph.
 897   #   none == Normal paragraph with just this one line on it.
 898   #   nbeg == First line of multi-line normal paragraph.
 899   #   nbdy == Interior line of multi-line normal paragraph.
 900   #   nend == Final line of multi-line normal paragraph.
 901   #   bone == Bulleted paragraph with just this one line on it.
 902   #   bbeg == First line of multi-line bulleted paragraph.
 903   #   bbdy == Interior line of multi-line bulleted paragraph.
 904   #   bend == Final line of multi-line bulleted paragraph.
 905   print "line-breaks...";
 906   $lname = "psline000000";
 907   $lnamei = "idx" . $lname;
 908   @lnames = @ltypes = ();
 909
 910   for ($para = 0; $para <= $#pnames; $para++) {
 911     $pname = $pnames[$para];
 912     $pflags = $pflags[$para];
 913     $ptype = substr($pflags,0,4);
 914
 915     # New paragraph _ergo_ new line.
 916     @line = ();
 917     @lindex = (); # list of index tags referenced to this line
 918
 919     if ($ptype eq "chap") {
 920       # Chapter heading. "Chapter N: Title" followed by a line of
 921       # minus signs.
 922       $pflags =~ /chap (.*) :(.*)/;
 923       push @line, "nChapter", " ", "n$1:", " ";
 924       foreach $i (@$pname) {
 925         $ww = &word_ps($i);
 926         push @line, $ww unless $ww eq "x";
 927       }
 928       @$lname = @line; @$lnamei = @lindex;
 929       push @lnames, $lname++;
 930       $lnamei = "idx" . $lname;
 931       push @ltypes, "chap";
 932     } elsif ($ptype eq "appn") {
 933       # Appendix heading. "Appendix N: Title" followed by a line of
 934       # minus signs.
 935       $pflags =~ /appn (.*) :(.*)/;
 936       push @line, "nAppendix", " ", "n$1:", " ";
 937       foreach $i (@$pname) {
 938         $ww = &word_ps($i);
 939         push @line, $ww unless $ww eq "x";
 940       }
 941       @$lname = @line; @$lnamei = @lindex;
 942       push @lnames, $lname++;
 943       $lnamei = "idx" . $lname;
 944       push @ltypes, "chap";
 945     } elsif ($ptype eq "head") {
 946       # Heading. Just a number and some text.
 947       $pflags =~ /.... (.*) :(.*)/;
 948       push @line, "n$1";
 949       foreach $i (@$pname) {
 950         $ww = &word_ps($i);
 951         push @line, $ww unless $ww eq "x";
 952       }
 953       @$lname = @line; @$lnamei = @lindex;
 954       push @lnames, $lname++;
 955       $lnamei = "idx" . $lname;
 956       push @ltypes, $ptype;
 957     } elsif ($ptype eq "subh") {
 958       # Subheading. Just a number and some text.
 959       $pflags =~ /subh (.*) :(.*)/;
 960       push @line, "n$1";
 961       foreach $i (@$pname) {
 962         push @line, &word_ps($i);
 963       }
 964       @$lname = @line; @$lnamei = @lindex;
 965       push @lnames, $lname++;
 966       $lnamei = "idx" . $lname;
 967       push @ltypes, "subh";
 968     } elsif ($ptype eq "code") {
 969       # Code paragraph. Emit lines one at a time.
 970       $type = "cbeg";
 971       foreach $i (@$pname) {
 972         @$lname = ("c$i");
 973         push @lnames, $lname++;
 974         $lnamei = "idx" . $lname;
 975         push @ltypes, $type;
 976         $type = "cbdy";
 977       }
 978       $ltypes[$#ltypes] = ($ltypes[$#ltypes] eq "cbeg" ? "cone" : "cend");
 979     } elsif ($ptype eq "bull" || $ptype eq "norm") {
 980       # Ordinary paragraph, optionally bulleted. We wrap, with ragged
 981       # 75-char right margin and either 7 or 11 char left margin
 982       # depending on bullets.
 983       if ($ptype eq "bull") {
 984         $width = 456; # leave 12-pt left indent for the bullet
 985         $type = $begtype = "bbeg";
 986         $bodytype = "bbdy";
 987         $onetype = "bone";
 988         $endtype = "bend";
 989       } else {
 990         $width = 468;
 991         $type = $begtype = "nbeg";
 992         $bodytype = "nbdy";
 993         $onetype = "none";
 994         $endtype = "nend";
 995       }
 996       @a = @$pname;
 997       @line = @wd = ();
 998       $linelen = 0;
 999       $wprev = undef;
1000       do {
1001         do { $w = &word_ps(shift @a) } while ($w eq "x");
1002         push @wd, $wprev if $wprev;
1003         if ($wprev =~ /^n.*-$/ || $w eq ' ' || $w eq '' || $w eq undef) {
1004           $wdlen = &len_ps(@wd);
1005           if ($linelen + $wdlen > $width) {
1006             pop @line while $line[$#line] eq ' '; # trim trailing spaces
1007             @$lname = @line; @$lnamei = @lindex;
1008             push @lnames, $lname++;
1009             $lnamei = "idx" . $lname;
1010             push @ltypes, $type;
1011             $type = $bodytype;
1012             @line = @lindex = ();
1013             $linelen = 0;
1014             shift @wd while $wd[0] eq ' '; # trim leading spaces
1015           }
1016           push @line, @wd;
1017           $linelen += $wdlen;
1018           @wd = ();
1019         }
1020         $wprev = $w;
1021       } while ($w ne '' && $w ne undef);
1022       if (@line) {
1023         pop @line while $line[$#line] eq ' '; # trim trailing spaces
1024         @$lname = @line; @$lnamei = @lindex;
1025         push @lnames, $lname++;
1026         $lnamei = "idx" . $lname;
1027         push @ltypes, $type;
1028         $type = $bodytype;
1029       }
1030       $ltypes[$#ltypes] =
1031         ($ltypes[$#ltypes] eq $begtype ? $onetype : $endtype);
1032     }
1033   }
1034
1035   # We've now processed the document source into lines. Before we
1036   # go on and do the page breaking, we'll fabricate a table of contents,
1037   # line by line, and then after doing page breaks we'll go back and
1038   # insert the page numbers into the contents entries.
1039   print "building contents...";
1040   @clnames = @cltypes = ();
1041   $clname = "pscont000000";
1042   @$clname = ("nContents"); # "chapter heading" for TOC
1043   push @clnames,$clname++;
1044   push @cltypes,"chap";
1045   for ($i=0; $i<=$#lnames; $i++) {
1046     $lname = $lnames[$i];
1047     if ($ltypes[$i] =~ /^(chap|head|subh)/) {
1048       @$clname = @$lname;
1049       splice @$clname,1,0," " if ($ltypes[$i] !~ /chap/);
1050       push @$clname,$i; # placeholder for page number
1051       push @clnames,$clname++;
1052       push @cltypes,"C" . substr($ltypes[$i],0,3);
1053     }
1054   }
1055   @$clname = ("nIndex"); # contents entry for Index
1056   push @$clname,$i;      # placeholder for page number
1057   $idx_clname = $clname;
1058   push @clnames,$clname++;
1059   push @cltypes,"Ccha";
1060   $contlen = $#clnames + 1;
1061   unshift @lnames,@clnames;
1062   unshift @ltypes,@cltypes;
1063
1064   # Second stage: now we have a list of lines, break them into pages.
1065   # We do this by means of adding a third array in parallel with
1066   # @lnames and @ltypes, called @lpages, in which we store the page
1067   # number that each line resides on. We also add @ycoord which
1068   # stores the vertical position of each line on the page.
1069   #
1070   # Page breaks may not come after line-types:
1071   #   chap head subh cbeg nbeg bbeg
1072   # and may not come before line-types:
1073   #   cend nend bend
1074   # They are forced before line-types:
1075   #   chap
1076   print "page-breaks...";
1077   $pmax = 600; # ADJUSTABLE: maximum length of a page in points
1078   $textht = 11; # ADJUSTABLE: height of a normal line in points
1079   $spacing = 6; # ADJUSTABLE: space between paragraphs, in points
1080   $headht = 14; # ADJUSTABLE: height of a major heading in points
1081   $subht = 12; # ADJUSTABLE: height of a sub-heading in points
1082   $pstart = 0; # start line of current page
1083   $plen = 0; # current length of current page
1084   $pnum = 1; # number of current page
1085   $bpt = -1; # last feasible break point
1086   $i = 0; # line number
1087   while ($i <= $#lnames) {
1088     $lname = $lnames[$i];
1089     # Add the height of this line (computed the last time we went round
1090     # the loop, unless we're a chapter heading in which case we do it
1091     # now) to the length of the current page. Also, _put_ this line on
1092     # the current page, and allocate it a y-coordinate.
1093     if ($ltypes[$i] =~ /^chap$/) {
1094       $plen = 100; # ADJUSTABLE: space taken up by a chapter heading
1095       $ycoord[$i] = 0; # chapter heading: y-coord doesn't matter
1096     } else {
1097       $ycoord[$i] = $plen + $space;
1098       $plen += $space + $ht;
1099     }
1100     # See if we can break after this line.
1101     $bpt = $i if $ltypes[$i] !~ /^chap|head|subh|cbeg|nbeg|bbeg$/ &&
1102                  $ltypes[$i+1] !~ /^cend|nend|bend$/;
1103     # Assume, to start with, that we don't break after this line.
1104     $break = 0;
1105     # See if a break is forced.
1106     $break = 1, $bpt = $i if $ltypes[$i+1] eq "chap" || !$ltypes[$i+1];
1107     # Otherwise, compute the height of the next line, and break if
1108     # it would make this page too long.
1109     $ht = $textht, $space = 0 if $ltypes[$i+1] =~ /^[nbc](bdy|end)$/;
1110     $ht = $textht, $space = $spacing if $ltypes[$i+1] =~ /^[nbc](one|beg)$/;
1111     $ht = $textht, $space = $spacing if $ltypes[$i+1] =~ /^C/;
1112     $ht = $subht, $space = $spacing if $ltypes[$i+1] eq "subh";
1113     $ht = $headht, $space = $spacing if $ltypes[$i+1] eq "head";
1114     $break = 1 if $plen + $space + $ht > $pmax;
1115     # Now, if we're breaking, assign page number $pnum to all lines up
1116     # to $bpt, set $i == $bpt+1, and zero $space since we are at the
1117     # start of a new page and don't want leading space.
1118     if ($break) {
1119       die "no feasible break point at all on page $pnum\n" if $bpt == -1;
1120       for ($j = $pstart; $j <= $bpt; $j++) {
1121         $lnamei = "idx" . $lnames[$j];
1122         foreach $k (@$lnamei) {
1123           ${$psidxpp{$k}}{$pnum} = 1;
1124         }
1125         $lpages[$j] = $pnum;
1126       }
1127       $pnum++;
1128       $i = $bpt;
1129       $bpt = -1;
1130       $pstart = $i+1;
1131       $plen = 0;
1132       $space = 0;
1133     }
1134     $i++;
1135   }
1136
1137   # Now fix up the TOC with page numbers.
1138   print "\n   fixing up contents...";
1139   for ($i=0; $i<=$#lnames; $i++) {
1140     $lname = $lnames[$i];
1141     if ($ltypes[$i] =~ /^C/) {
1142       $j = pop @$lname;
1143       push @$lname, "n" . $lpages[$j+$contlen];
1144     }
1145   }
1146
1147   # Having got page numbers for most stuff, generate an index.
1148   print "building index...";
1149   $iwid = 222;
1150   $sep = 12;
1151   $commaindent = 32;
1152   foreach $k (@itags) {
1153     @line = ();
1154     $cmd = "index";
1155     @idxentry = @{$idxmap{$k}};
1156     if ($commaafter{$k} and !$commanext{$k}) {
1157       # This line is a null line beginning a multiple entry. We must
1158       # output the prefix on a line by itself.
1159
1160       @idxhead = splice @idxentry,0,$commapos{$k};
1161       @line = ();
1162       foreach $i (@idxhead) {
1163         $ww = &word_ps($i);
1164         push @line, $ww unless $ww eq "x";
1165       }
1166       &ps_idxout("index",\@line,[]);
1167       $cmd = "iindex";
1168       @line = ();
1169     }
1170     $cmd = "iindex", splice @idxentry,0,$commapos{$k} if $commanext{$k};
1171     foreach $i (@idxentry) {
1172       $ww = &word_ps($i);
1173       push @line, $ww unless $ww eq "x";
1174     }
1175     $len = $iwid - $sep - &len_ps(@line);
1176     warn "text for index tag `%s' is longer than one index line!\n"
1177       if $len < -$sep;
1178     @pp = ();
1179     $inums = join(',',sort { $a <=> $b } keys %{$psidxpp{$k}});
1180     while (length $inums) {
1181       $inums =~ /^([^,]+,?)(.*)$/;
1182       $inums = $2, $inum = $1;
1183       @pnum = (" ", "n$inum");
1184       $pnumlen = &len_ps(@pnum);
1185       if ($pnumlen > $len) {
1186         &ps_idxout($cmd,\@line,\@pp);
1187         @pp = ();
1188         @line = ();
1189         $cmd = "index";
1190         $len = $iwid - $sep;
1191       }
1192       push @pp, @pnum;
1193       $len -= $pnumlen;
1194     }
1195     &ps_idxout($cmd,\@line,\@pp) if (length @pp);
1196     $l1 = &len_ps(@line);
1197     $l2 = &len_ps($pp);
1198   }
1199   $$idx_clname[$#$idx_clname] = "n" . $pnum; # fix up TOC entry for index
1200
1201   print "writing file...";
1202   open PS,">nasmdoc.ps";
1203   select PS;
1204   $page = $lpages[0];
1205   &ps_header;
1206   for ($i=0; $i<=$#lnames; $i++) {
1207     &ps_throw_pg($page,$lpages[$i]) if $page != $lpages[$i];
1208     $page = $lpages[$i];
1209     &ps_out_line($ycoord[$i],$ltypes[$i],$lnames[$i]);
1210   }
1211   $i = 0;
1212   while ($i <= $#psindex) {
1213     &ps_throw_pg($page, $pnum) if $page != $pnum;
1214     $page = $pnum++;
1215     $ypos = 0;
1216     $ypos = 100, &ps_out_line(0, "chap", ["nIndex"]) if !$i;
1217     $lines = ($pmax - $ypos) / $textht;
1218     my $col; # ps_out_line hits this variable
1219     PAGE:for ($col = 1; $col <= 2; $col++) {
1220       $y = $ypos; $l = $lines;
1221       COL: while ($l > 0) {
1222         $j = $i+1;
1223         $j++ while $psindex[$j] and ($psindex[$j][3] == 0); # find next break
1224         last COL if $j-$i > $l or $i > $#psindex;
1225         while ($i < $j) {
1226           &ps_out_line($y, $psindex[$i][0] eq "index" ? "idl$col" : "ldl$col",
1227                        $psindex[$i][1]);
1228           &ps_out_line($y,"idr$col",$psindex[$i][2]);
1229           $i++;
1230           $y += $textht;
1231           $l--;
1232         }
1233       }
1234       last PAGE if $i > $#psindex;
1235     }
1236   }
1237   &ps_trailer;
1238   close PS;
1239   select STDOUT;
1240 }
1241
1242 sub ps_idxout {
1243   my ($cmd, $left, $right) = @_;
1244   my $break = 1;
1245   $break = 0
1246       if ($#psindex >= 0) and ( ($#$left < 0) or ($cmd eq "iindex") );
1247   push @psindex,[$cmd,[@$left],[@$right],$break];
1248 }
1249
1250 sub ps_header {
1251   @pshdr = (
1252     '/sp (n ) def', # here it's sure not to get wrapped inside ()
1253     '/nf /Times-Roman findfont 11 scalefont def',
1254     '/ef /Times-Italic findfont 11 scalefont def',
1255     '/cf /Courier findfont 11 scalefont def',
1256     '/nc /Helvetica-Bold findfont 18 scalefont def',
1257     '/ec /Helvetica-Oblique findfont 18 scalefont def',
1258     '/cc /Courier-Bold findfont 18 scalefont def',
1259     '/nh /Helvetica-Bold findfont 14 scalefont def',
1260     '/eh /Helvetica-Oblique findfont 14 scalefont def',
1261     '/ch /Courier-Bold findfont 14 scalefont def',
1262     '/ns /Helvetica-Bold findfont 12 scalefont def',
1263     '/es /Helvetica-Oblique findfont 12 scalefont def',
1264     '/cs /Courier-Bold findfont 12 scalefont def',
1265     '/n 16#6E def /e 16#65 def /c 16#63 def',
1266     '/chapter {',
1267     '  100 620 moveto',
1268     '  {',
1269     '    dup 0 get',
1270     '    dup n eq {pop nc setfont} {',
1271     '      e eq {ec setfont} {cc setfont} ifelse',
1272     '    } ifelse',
1273     '    dup length 1 sub 1 exch getinterval show',
1274     '  } forall',
1275     '  0 setlinecap 3 setlinewidth',
1276     '  newpath 100 610 moveto 468 0 rlineto stroke',
1277     '} def',
1278     '/heading {',
1279     '  686 exch sub /y exch def /a exch def',
1280     '  90 y moveto a 0 get dup length 1 sub 1 exch getinterval',
1281     '  nh setfont dup stringwidth pop neg 0 rmoveto show',
1282     '  100 y moveto',
1283     '  a dup length 1 sub 1 exch getinterval {',
1284     '    /s exch def',
1285     '    s 0 get',
1286     '    dup n eq {pop nh setfont} {',
1287     '      e eq {eh setfont} {ch setfont} ifelse',
1288     '    } ifelse',
1289     '    s s length 1 sub 1 exch getinterval show',
1290     '  } forall',
1291     '} def',
1292     '/subhead {',
1293     '  688 exch sub /y exch def /a exch def',
1294     '  90 y moveto a 0 get dup length 1 sub 1 exch getinterval',
1295     '  ns setfont dup stringwidth pop neg 0 rmoveto show',
1296     '  100 y moveto',
1297     '  a dup length 1 sub 1 exch getinterval {',
1298     '    /s exch def',
1299     '    s 0 get',
1300     '    dup n eq {pop ns setfont} {',
1301     '      e eq {es setfont} {cs setfont} ifelse',
1302     '    } ifelse',
1303     '    s s length 1 sub 1 exch getinterval show',
1304     '  } forall',
1305     '} def',
1306     '/disp { /j exch def',
1307     '  568 exch sub exch 689 exch sub moveto',
1308     '  {',
1309     '    /s exch def',
1310     '    s 0 get',
1311     '    dup n eq {pop nf setfont} {',
1312     '      e eq {ef setfont} {cf setfont} ifelse',
1313     '    } ifelse',
1314     '    s s length 1 sub 1 exch getinterval show',
1315     '    s sp eq {j 0 rmoveto} if',
1316     '  } forall',
1317     '} def',
1318     '/contents { /w exch def /y exch def /a exch def',
1319     '  /yy 689 y sub def',
1320     '  a a length 1 sub get dup length 1 sub 1 exch getinterval /s exch def',
1321     '  nf setfont 568 s stringwidth pop sub /ex exch def',
1322     '  ex yy moveto s show',
1323     '  a 0 a length 1 sub getinterval y w 0 disp',
1324     '  /sx currentpoint pop def nf setfont',
1325     '  100 10 568 { /i exch def',
1326     '    i 5 sub sx gt i 5 add ex lt and {',
1327     '      i yy moveto (.) show',
1328     '    } if',
1329     '  } for',
1330     '} def',
1331     '/just { /w exch def /y exch def /a exch def',
1332     '  /jj w def /spaces 0 def',
1333     '  a {',
1334     '    /s exch def',
1335     '    s 0 get',
1336     '    dup n eq {pop nf setfont} {',
1337     '      e eq {ef setfont} {cf setfont} ifelse',
1338     '    } ifelse',
1339     '    s s length 1 sub 1 exch getinterval stringwidth pop',
1340     '    jj exch sub /jj exch def',
1341     '    s sp eq {/spaces spaces 1 add def} if',
1342     '  } forall',
1343     '  a y w jj spaces spaces 0 eq {pop pop 0} {div} ifelse disp',
1344     '} def',
1345     '/idl { 468 exch sub 0 disp } def',
1346     '/ldl { 436 exch sub 0 disp } def',
1347     '/idr { 222 add 468 exch sub /x exch def /y exch def /a exch def',
1348     '  a {',
1349     '    /s exch def',
1350     '    s 0 get',
1351     '    dup n eq {pop nf setfont} {',
1352     '      e eq {ef setfont} {cf setfont} ifelse',
1353     '    } ifelse',
1354     '    s s length 1 sub 1 exch getinterval stringwidth pop',
1355     '    x add /x exch def',
1356     '  } forall',
1357     '  a y x 0 disp',
1358     '} def',
1359     '/left {0 disp} def',
1360     '/bullet {',
1361     '  nf setfont dup 100 exch 689 exch sub moveto (\267) show',
1362     '} def'
1363   );
1364   print "%!PS-Adobe-3.0\n";
1365   print "%%BoundingBox: 95 95 590 705\n";
1366   print "%%Creator: a nasty Perl script\n";
1367   print "%%DocumentData: Clean7Bit\n";
1368   print "%%Orientation: Portrait\n";
1369   print "%%Pages: $lpages[$#lpages]\n";
1370   print "%%DocumentNeededResources: font Times-Roman Times-Italic\n";
1371   print "%%+ font Helvetica-Bold Courier Courier-Bold\n";
1372   print "%%EndComments\n%%BeginProlog\n%%EndProlog\n%%BeginSetup\nsave\n";
1373   $pshdr = join(' ',@pshdr);
1374   $pshdr =~ s/\s+/ /g;
1375   while ($pshdr =~ /\S/) {
1376     last if length($pshdr) < 72 || $pshdr !~ /^(.{0,72}\S)\s(.*)$/;
1377     $pshdr = $2;
1378     print "$1\n";
1379   }
1380   print "$pshdr\n" if $pshdr =~ /\S/;
1381   print "%%EndSetup\n";
1382   &ps_initpg($lpages[0]);
1383 }
1384
1385 sub ps_trailer {
1386   &ps_donepg;
1387   print "%%Trailer\nrestore\n%%EOF\n";
1388 }
1389
1390 sub ps_throw_pg {
1391   my ($oldpg, $newpg) = @_;
1392   &ps_donepg;
1393   &ps_initpg($newpg);
1394 }
1395
1396 sub ps_initpg {
1397   my ($pgnum) = @_;
1398   print "%%Page: $pgnum $pgnum\n";
1399   print "%%BeginPageSetup\nsave\n%%EndPageSetup\n";
1400 }
1401
1402 sub ps_donepg {
1403   print "%%PageTrailer\nrestore showpage\n";
1404 }
1405
1406 sub ps_out_line {
1407   my ($ypos,$ltype,$lname) = @_;
1408   my $c,$d,$wid;
1409
1410   print "[";
1411   $col = 1;
1412   foreach $c (@$lname) {#
1413     $c= "n " if $c eq " ";
1414     $c = "n\261" if $c eq "-";
1415     $d = '';
1416     while (length $c) {
1417       $d .= $1, $c = $2 while $c =~ /^([ -'\*-\[\]-~]+)(.*)$/;
1418       while (1) {
1419         $d .= "\\$1", $c = $2, next if $c =~ /^([\\\(\)])(.*)$/;
1420         ($d .= sprintf "\\%3o",unpack("C",$1)), $c = $2, next
1421           if $c =~ /^([^ -~])(.*)$/;
1422         last;
1423       }
1424     }
1425     $d = "($d)";
1426     $col = 0, print "\n" if $col>0 && $col+length $d > 77;
1427     print $d;
1428     $col += length $d;
1429   }
1430   print "\n" if $col > 60;
1431   print "]";
1432   if ($ltype =~ /^[nb](beg|bdy)$/) {
1433     printf "%d %s%d just\n",
1434       $ypos, ($ltype eq "bbeg" ? "bullet " : ""),
1435       ($ltype =~ /^b/ ? 456 : 468);
1436   } elsif ($ltype =~ /^[nb](one|end)$/) {
1437     printf "%d %s%d left\n",
1438       $ypos, ($ltype eq "bone" ? "bullet " : ""),
1439       ($ltype =~ /^b/ ? 456 : 468);
1440   } elsif ($ltype =~ /^c(one|beg|bdy|end)$/) {
1441     printf "$ypos 468 left\n";
1442   } elsif ($ltype =~ /^C/) {
1443     $wid = 468;
1444     $wid = 456 if $ltype eq "Chea";
1445     $wid = 444 if $ltype eq "Csub";
1446     printf "$ypos $wid contents\n";
1447   } elsif ($ltype eq "chap") {
1448     printf "chapter\n";
1449   } elsif ($ltype eq "head") {
1450     printf "$ypos heading\n";
1451   } elsif ($ltype eq "subh") {
1452     printf "$ypos subhead\n";
1453   } elsif ($ltype =~ /([il]d[lr])([12])/) {
1454     $left = ($2 eq "2" ? 468-222 : 0);
1455     printf "$ypos $left $1\n";
1456   }
1457 }
1458
1459 sub word_ps {
1460   my ($w) = @_;
1461   my $wtype, $wmajt;
1462
1463   return undef if $w eq '' || $w eq undef;
1464
1465   $wtype = substr($w,0,2);
1466   $wmajt = substr($wtype,0,1);
1467   $w = substr($w,2);
1468   $w =~ s/<.*>// if $wmajt eq "w"; # remove web links
1469   if ($wmajt eq "n" || $wtype eq "w ") {
1470     return "n$w";
1471   } elsif ($wtype eq "sp") {
1472     return ' ';
1473   } elsif ($wtype eq "da") {
1474     return '-';
1475   } elsif ($wmajt eq "c" || $wtype eq "wc") {
1476     return "c$w";
1477   } elsif ($wmajt eq "e") {
1478     return "e$w";
1479   } elsif ($wmajt eq "x") {
1480     return "x";
1481   } elsif ($wtype eq "i ") {
1482     push @lindex, $w;
1483     return "x";
1484   } else {
1485     die "panic in word_ps: $wtype$w\n";
1486   }
1487 }
1488
1489 sub len_ps {
1490   my (@line) = @_;
1491   my $l = 0;
1492   my $w, $size;
1493
1494   $size = 11/1000; # used only for length calculations
1495   while ($w = shift @line) {
1496     $w = "n " if $w eq " ";
1497     $w = "n\261" if $w eq "-";
1498     $f = substr($w,0,1);
1499     $f = "timesr" if $f eq "n";
1500     $f = "timesi" if $f eq "e";
1501     $f = "courr" if $f eq "c";
1502     foreach $c (unpack 'C*',substr($w,1)) {
1503       $l += $size * $$f[$c];
1504     }
1505   }
1506   return $l;
1507 }
1508
1509 sub write_texi {
1510   # This is called from the top level, so I won't bother using
1511   # my or local.
1512
1513   # Open file.
1514   print "writing file...";
1515   open TEXT,">nasmdoc.texi";
1516   select TEXT;
1517
1518   # Preamble.
1519   print "\input texinfo   \@c -*-texinfo-*-\n";
1520   print "\@c \%**start of header\n";
1521   print "\@setfilename nasm.info\n";
1522   print "\@settitle NASM: The Netwide Assembler\n";
1523   print "\@setchapternewpage odd\n";
1524   print "\@c \%**end of header\n";
1525   print "\n";
1526   print "\@ifinfo\n";
1527   print "This file documents NASM, the Netwide Assembler: an assembler\n";
1528   print "targetting the Intel x86 series of processors, with portable source.\n";
1529   print "\n";
1530   print "Copyright 1997 Simon Tatham\n";
1531   print "\n";
1532   print "All rights reserved. This document is redistributable under the\n";
1533   print "licence given in the file \"Licence\" distributed in the NASM archive.\n";
1534   print "\@end ifinfo\n";
1535   print "\n";
1536   print "\@titlepage\n";
1537   print "\@title NASM: The Netwide Assembler\n";
1538   print "\@author Simon Tatham\n";
1539   print "\n";
1540   print "\@page\n";
1541   print "\@vskip 0pt plus 1filll\n";
1542   print "Copyright \@copyright{} 1997 Simon Tatham\n";
1543   print "\n";
1544   print "All rights reserved. This document is redistributable under the\n";
1545   print "licence given in the file \"Licence\" distributed in the NASM archive.\n";
1546   print "\@end titlepage\n";
1547   print "\n";
1548   print "\@node Top, $tstruct_next{'Top'}, (dir), (dir)\n";
1549   print "\@top\n";
1550   print "\n";
1551   print "\@ifinfo\n";
1552   print "This file documents NASM, the Netwide Assembler: an assembler\n";
1553   print "targetting the Intel x86 series of processors, with portable source.\n";
1554   print "\@end ifinfo\n";
1555
1556   $node = "Top";
1557
1558   $bulleting = 0;
1559   for ($para = 0; $para <= $#pnames; $para++) {
1560     $pname = $pnames[$para];
1561     $pflags = $pflags[$para];
1562     $ptype = substr($pflags,0,4);
1563
1564     $bulleting = 0, print "\@end itemize\n" if $bulleting && $ptype ne "bull";
1565     print "\n"; # always one of these before a new paragraph
1566
1567     if ($ptype eq "chap") {
1568       # Chapter heading. Begin a new node.
1569       &texi_menu($node)
1570         if $tstruct_level{$tstruct_next{$node}} > $tstruct_level{$node};
1571       $pflags =~ /chap (.*) :(.*)/;
1572       $node = "Chapter $1";
1573       $title = "Chapter $1: ";
1574       foreach $i (@$pname) {
1575         $ww = &word_texi($i);
1576         $title .= $ww unless $ww eq "\001";
1577       }
1578       print "\@node $node, $tstruct_next{$node}, $tstruct_prev{$node},";
1579       print " $tstruct_up{$node}\n\@unnumbered $title\n";
1580     } elsif ($ptype eq "appn") {
1581       # Appendix heading. Begin a new node.
1582       &texi_menu($node)
1583         if $tstruct_level{$tstruct_next{$node}} > $tstruct_level{$node};
1584       $pflags =~ /appn (.*) :(.*)/;
1585       $node = "Appendix $1";
1586       $title = "Appendix $1: ";
1587       foreach $i (@$pname) {
1588         $ww = &word_texi($i);
1589         $title .= $ww unless $ww eq "\001";
1590       }
1591       print "\@node $node, $tstruct_next{$node}, $tstruct_prev{$node},";
1592       print " $tstruct_up{$node}\n\@unnumbered $title\n";
1593     } elsif ($ptype eq "head" || $ptype eq "subh") {
1594       # Heading or subheading. Begin a new node.
1595       &texi_menu($node)
1596         if $tstruct_level{$tstruct_next{$node}} > $tstruct_level{$node};
1597       $pflags =~ /.... (.*) :(.*)/;
1598       $node = "Section $1";
1599       $title = "$1. ";
1600       foreach $i (@$pname) {
1601         $ww = &word_texi($i);
1602         $title .= $ww unless $ww eq "\001";
1603       }
1604       print "\@node $node, $tstruct_next{$node}, $tstruct_prev{$node},";
1605       print " $tstruct_up{$node}\n\@unnumbered $title\n";
1606     } elsif ($ptype eq "code") {
1607       # Code paragraph. Surround with @example / @end example.
1608       print "\@example\n";
1609       foreach $i (@$pname) {
1610         warn "code line longer than 68 chars: $i\n" if length $i > 68;
1611         $i =~ s/\@/\@\@/g;
1612         $i =~ s/\{/\@\{/g;
1613         $i =~ s/\}/\@\}/g;
1614         print "$i\n";
1615       }
1616       print "\@end example\n";
1617     } elsif ($ptype eq "bull" || $ptype eq "norm") {
1618       # Ordinary paragraph, optionally bulleted. We wrap, FWIW.
1619       if ($ptype eq "bull") {
1620         $bulleting = 1, print "\@itemize \@bullet\n" if !$bulleting;
1621         print "\@item\n";
1622       }
1623       $line = '';
1624       @a = @$pname;
1625       $wd = $wprev = '';
1626       do {
1627         do { $w = &word_texi(shift @a); } while $w eq "\001"; # hack
1628         $wd .= $wprev;
1629         if ($wprev =~ /-$/ || $w eq ' ' || $w eq '' || $w eq undef) {
1630           if (length ($line . $wd) > 75) {
1631             $line =~ s/\s*$//; # trim trailing spaces
1632             print "$line\n";
1633             $line = '';
1634             $wd =~ s/^\s*//; # trim leading spaces
1635           }
1636           $line .= $wd;
1637           $wd = '';
1638         }
1639         $wprev = $w;
1640       } while ($w ne '' && $w ne undef);
1641       if ($line =~ /\S/) {
1642         $line =~ s/\s*$//; # trim trailing spaces
1643         print "$line\n";
1644       }
1645     }
1646   }
1647
1648   # Write index.
1649   &texi_index;
1650
1651   # Close file.
1652   print "\n\@contents\n\@bye\n";
1653   select STDOUT;
1654   close TEXT;
1655 }
1656
1657 # Side effect of this procedure: update global `texiwdlen' to be the length
1658 # in chars of the formatted version of the word.
1659 sub word_texi {
1660   my ($w) = @_;
1661   my $wtype, $wmajt;
1662
1663   return undef if $w eq '' || $w eq undef;
1664   $wtype = substr($w,0,2);
1665   $wmajt = substr($wtype,0,1);
1666   $w = substr($w,2);
1667   $wlen = length $w;
1668   $w =~ s/\@/\@\@/g;
1669   $w =~ s/\{/\@\{/g;
1670   $w =~ s/\}/\@\}/g;
1671   $w =~ s/<.*>// if $wmajt eq "w"; # remove web links
1672   substr($w,0,1) =~ tr/a-z/A-Z/, $capital = 0 if $capital;
1673   if ($wmajt eq "n" || $wtype eq "e " || $wtype eq "w ") {
1674     $texiwdlen = $wlen;
1675     return $w;
1676   } elsif ($wtype eq "sp") {
1677     $texiwdlen = 1;
1678     return ' ';
1679   } elsif ($wtype eq "da") {
1680     $texiwdlen = 2;
1681     return '--';
1682   } elsif ($wmajt eq "c" || $wtype eq "wc") {
1683     $texiwdlen = 2 + $wlen;
1684     return "\@code\{$w\}";
1685   } elsif ($wtype eq "es") {
1686     $texiwdlen = 1 + $wlen;
1687     return "\@emph\{${w}";
1688   } elsif ($wtype eq "ee") {
1689     $texiwdlen = 1 + $wlen;
1690     return "${w}\}";
1691   } elsif ($wtype eq "eo") {
1692     $texiwdlen = 2 + $wlen;
1693     return "\@emph\{${w}\}";
1694   } elsif ($wtype eq "x ") {
1695     $texiwdlen = 0; # we don't need it in this case
1696     $capital = 1; # hack
1697     return "\@ref\{";
1698   } elsif ($wtype eq "xe") {
1699     $texiwdlen = 0; # we don't need it in this case
1700     return "\}";
1701   } elsif ($wmajt eq "i") {
1702     $texiwdlen = 0; # we don't need it in this case
1703     return "\001";
1704   } else {
1705     die "panic in word_texi: $wtype$w\n";
1706   }
1707 }
1708
1709 sub texi_menu {
1710   my ($topitem) = @_;
1711   my $item, $i, $mpname, $title, $wd;
1712
1713   $item = $tstruct_next{$topitem};
1714   print "\@menu\n";
1715   while ($item) {
1716     $title = "";
1717     $mpname = $tstruct_pname{$item};
1718     foreach $i (@$mpname) {
1719       $wd = &word_texi($i);
1720       $title .= $wd unless $wd eq "\001";
1721     }
1722     print "* ${item}:: $title\n";
1723     $item = $tstruct_mnext{$item};
1724   }
1725   print "* Index::\n" if $topitem eq "Top";
1726   print "\@end menu\n";
1727 }
1728
1729 sub texi_index {
1730   my $itag, $ientry, @a, $wd, $item, $len;
1731   my $subnums = "123456789ABCDEFGHIJKLMNOPQRSTU" .
1732                 "VWXYZabcdefghijklmnopqrstuvwxyz";
1733
1734   print "\@ifinfo\n\@node Index, , $FIXMElastnode, Top\n";
1735   print "\@unnumbered Index\n\n\@menu\n";
1736
1737   foreach $itag (@itags) {
1738     $ientry = $idxmap{$itag};
1739     @a = @$ientry;
1740     $item = '';
1741     $len = 0;
1742     foreach $i (@a) {
1743       $wd = &word_texi($i);
1744       $item .= $wd, $len += $texiwdlen unless $wd eq "\001";
1745     }
1746     $i = 0;
1747     foreach $node (@nodes) {
1748       next if !$idxnodes{$node,$itag};
1749       printf "* %s%s (%s): %s.\n",
1750           $item, " " x (40-$len), substr($subnums,$i++,1), $node;
1751     }
1752   }
1753   print "\@end menu\n\@end ifinfo\n";
1754 }
1755
1756 sub write_hlp {
1757   # This is called from the top level, so I won't bother using
1758   # my or local.
1759
1760   # Build the index-tag text forms.
1761   print "building index entries...";
1762   @hlp_index = map {
1763                  my $i,$ww;
1764                  my $ientry = $idxmap{$_};
1765                  my $title = "";
1766                  foreach $i (@$ientry) {
1767                    $ww = &word_hlp($i,0);
1768                    $title .= $ww unless $ww eq "\001";
1769                  }
1770                  $title;
1771                } @itags;
1772
1773   # Write the HPJ project-description file.
1774   print "writing .hpj file...";
1775   open HPJ,">nasmdoc.hpj";
1776   print HPJ "[OPTIONS]\ncompress=true\n";
1777   print HPJ "title=NASM: The Netwide Assembler\noldkeyphrase=no\n\n";
1778   print HPJ "[FILES]\nnasmdoc.rtf\n\n";
1779   print HPJ "[CONFIG]\n";
1780   print HPJ 'CreateButton("btn_up", "&Up",'.
1781             ' "JumpContents(`nasmdoc.hlp'."'".')")';
1782   print HPJ "\nBrowseButtons()\n";
1783   close HPJ;
1784
1785   # Open file.
1786   print "\n   writing .rtf file...";
1787   open TEXT,">nasmdoc.rtf";
1788   select TEXT;
1789
1790   # Preamble.
1791   print "{\\rtf1\\ansi{\\fonttbl\n";
1792   print "\\f0\\froman Times New Roman;\\f1\\fmodern Courier New;\n";
1793   print "\\f2\\fswiss Arial;\\f3\\ftech Wingdings}\\deff0\n";
1794   print "#{\\footnote Top}\n";
1795   print "\${\\footnote Contents}\n";
1796   print "+{\\footnote browse:00000}\n";
1797   print "!{\\footnote DisableButton(\"btn_up\")}\n";
1798   print "\\keepn\\f2\\b\\fs30\\sb0\n";
1799   print "NASM: The Netwide Assembler\n";
1800   print "\\par\\pard\\plain\\sb120\n";
1801   print "This file documents NASM, the Netwide Assembler: an assembler \n";
1802   print "targetting the Intel x86 series of processors, with portable source.\n";
1803
1804   $node = "Top";
1805   $browse = 0;
1806
1807   $newpar = "\\par\\sb120\n";
1808   for ($para = 0; $para <= $#pnames; $para++) {
1809     $pname = $pnames[$para];
1810     $pflags = $pflags[$para];
1811     $ptype = substr($pflags,0,4);
1812
1813     print $newpar;
1814     $newpar = "\\par\\sb120\n";
1815
1816     if ($ptype eq "chap") {
1817       # Chapter heading. Begin a new node.
1818       &hlp_menu($node)
1819         if $tstruct_level{$tstruct_next{$node}} > $tstruct_level{$node};
1820       $pflags =~ /chap (.*) :(.*)/;
1821       $node = "Chapter $1";
1822       $title = $footnotetitle = "Chapter $1: ";
1823       foreach $i (@$pname) {
1824         $ww = &word_hlp($i,1);
1825         $title .= $ww, $footnotetitle .= &word_hlp($i,0) unless $ww eq "\001";
1826       }
1827       print "\\page\n";
1828       printf "#{\\footnote %s}\n", &hlp_sectkw($node);
1829       print "\${\\footnote $footnotetitle}\n";
1830       printf "+{\\footnote browse:%05d}\n", ++$browse;
1831       printf "!{\\footnote ChangeButtonBinding(\"btn_up\"," .
1832              "\"JumpId(\`nasmdoc.hlp',\`%s')\");\n",
1833              &hlp_sectkw($tstruct_up{$node});
1834       print "EnableButton(\"btn_up\")}\n";
1835       &hlp_keywords($node);
1836       print "\\keepn\\f2\\b\\fs30\\sb60\\sa60\n";
1837       print "$title\n";
1838       $newpar = "\\par\\pard\\plain\\sb120\n";
1839     } elsif ($ptype eq "appn") {
1840       # Appendix heading. Begin a new node.
1841       &hlp_menu($node)
1842         if $tstruct_level{$tstruct_next{$node}} > $tstruct_level{$node};
1843       $pflags =~ /appn (.*) :(.*)/;
1844       $node = "Appendix $1";
1845       $title = $footnotetitle = "Appendix $1: ";
1846       foreach $i (@$pname) {
1847         $ww = &word_hlp($i,1);
1848         $title .= $ww, $footnotetitle .= &word_hlp($i,0) unless $ww eq "\001";
1849       }
1850       print "\\page\n";
1851       printf "#{\\footnote %s}\n", &hlp_sectkw($node);
1852       print "\${\\footnote $footnotetitle}\n";
1853       printf "+{\\footnote browse:%05d}\n", ++$browse;
1854       printf "!{\\footnote ChangeButtonBinding(\"btn_up\"," .
1855              "\"JumpId(\`nasmdoc.hlp',\`%s')\");\n",
1856              &hlp_sectkw($tstruct_up{$node});
1857       print "EnableButton(\"btn_up\")}\n";
1858       &hlp_keywords($node);
1859       print "\\keepn\\f2\\b\\fs30\\sb60\\sa60\n";
1860       print "$title\n";
1861       $newpar = "\\par\\pard\\plain\\sb120\n";
1862     } elsif ($ptype eq "head" || $ptype eq "subh") {
1863       # Heading or subheading. Begin a new node.
1864       &hlp_menu($node)
1865         if $tstruct_level{$tstruct_next{$node}} > $tstruct_level{$node};
1866       $pflags =~ /.... (.*) :(.*)/;
1867       $node = "Section $1";
1868       $title = $footnotetitle = "$1. ";
1869       foreach $i (@$pname) {
1870         $ww = &word_hlp($i,1);
1871         $title .= $ww, $footnotetitle .= &word_hlp($i,0) unless $ww eq "\001";
1872       }
1873       print "\\page\n";
1874       printf "#{\\footnote %s}\n", &hlp_sectkw($node);
1875       print "\${\\footnote $footnotetitle}\n";
1876       printf "+{\\footnote browse:%05d}\n", ++$browse;
1877       printf "!{\\footnote ChangeButtonBinding(\"btn_up\"," .
1878              "\"JumpId(\`nasmdoc.hlp',\`%s')\");\n",
1879              &hlp_sectkw($tstruct_up{$node});
1880       print "EnableButton(\"btn_up\")}\n";
1881       &hlp_keywords($node);
1882       print "\\keepn\\f2\\b\\fs30\\sb60\\sa60\n";
1883       print "$title\n";
1884       $newpar = "\\par\\pard\\plain\\sb120\n";
1885     } elsif ($ptype eq "code") {
1886       # Code paragraph.
1887       print "\\keep\\f1\\sb120\n";
1888       foreach $i (@$pname) {
1889         warn "code line longer than 68 chars: $i\n" if length $i > 68;
1890         $i =~ s/\\/\\\\/g;
1891         $i =~ s/\{/\\\{/g;
1892         $i =~ s/\}/\\\}/g;
1893         print "$i\\par\\sb0\n";
1894       }
1895       $newpar = "\\pard\\f0\\sb120\n";
1896     } elsif ($ptype eq "bull" || $ptype eq "norm") {
1897       # Ordinary paragraph, optionally bulleted. We wrap, FWIW.
1898       if ($ptype eq "bull") {
1899         print "\\tx360\\li360\\fi-360{\\f3\\'9F}\\tab\n";
1900         $newpar = "\\par\\pard\\sb120\n";
1901       } else {
1902         $newpar = "\\par\\sb120\n";
1903       }
1904       $line = '';
1905       @a = @$pname;
1906       $wd = $wprev = '';
1907       do {
1908         do { $w = &word_hlp((shift @a),1); } while $w eq "\001"; # hack
1909         $wd .= $wprev;
1910         if ($w eq ' ' || $w eq '' || $w eq undef) {
1911           if (length ($line . $wd) > 75) {
1912             $line =~ s/\s*$//; # trim trailing spaces
1913             print "$line \n"; # and put one back
1914             $line = '';
1915             $wd =~ s/^\s*//; # trim leading spaces
1916           }
1917           $line .= $wd;
1918           $wd = '';
1919         }
1920         $wprev = $w;
1921       } while ($w ne '' && $w ne undef);
1922       if ($line =~ /\S/) {
1923         $line =~ s/\s*$//; # trim trailing spaces
1924         print "$line\n";
1925       }
1926     }
1927   }
1928
1929   # Close file.
1930   print "\\page}\n";
1931   select STDOUT;
1932   close TEXT;
1933 }
1934
1935 sub word_hlp {
1936   my ($w, $docode) = @_;
1937   my $wtype, $wmajt;
1938
1939   return undef if $w eq '' || $w eq undef;
1940   $wtype = substr($w,0,2);
1941   $wmajt = substr($wtype,0,1);
1942   $w = substr($w,2);
1943   $w =~ s/\\/\\\\/g;
1944   $w =~ s/\{/\\\{/g;
1945   $w =~ s/\}/\\\}/g;
1946   $w =~ s/<.*>// if $wmajt eq "w"; # remove web links
1947   substr($w,0,length($w)-1) =~ s/-/\\'AD/g if $wmajt ne "x"; #nonbreakhyphens
1948   if ($wmajt eq "n" || $wtype eq "e " || $wtype eq "w ") {
1949     return $w;
1950   } elsif ($wtype eq "sp") {
1951     return ' ';
1952   } elsif ($wtype eq "da") {
1953     return "\\'96";
1954   } elsif ($wmajt eq "c" || $wtype eq "wc") {
1955     $w =~ s/ /\\'A0/g; # make spaces non-breaking
1956     return $docode ? "{\\f1 ${w}}" : $w;
1957   } elsif ($wtype eq "es") {
1958     return "{\\i ${w}";
1959   } elsif ($wtype eq "ee") {
1960     return "${w}}";
1961   } elsif ($wtype eq "eo") {
1962     return "{\\i ${w}}";
1963   } elsif ($wtype eq "x ") {
1964     return "{\\uldb ";
1965   } elsif ($wtype eq "xe") {
1966     $w = &hlp_sectkw($w);
1967     return "}{\\v ${w}}";
1968   } elsif ($wmajt eq "i") {
1969     return "\001";
1970   } else {
1971     die "panic in word_hlp: $wtype$w\n";
1972   }
1973 }
1974
1975 sub hlp_menu {
1976   my ($topitem) = @_;
1977   my $item, $kword, $i, $mpname, $title;
1978
1979   $item = $tstruct_next{$topitem};
1980   print "\\li360\\fi-360\n";
1981   while ($item) {
1982     $title = "";
1983     $mpname = $tstruct_pname{$item};
1984     foreach $i (@$mpname) {
1985       $ww = &word_hlp($i, 0);
1986       $title .= $ww unless $ww eq "\001";
1987     }
1988     $kword = &hlp_sectkw($item);
1989     print "{\\uldb ${item}: $title}{\\v $kword}\\par\\sb0\n";
1990     $item = $tstruct_mnext{$item};
1991   }
1992   print "\\pard\\sb120\n";
1993 }
1994
1995 sub hlp_sectkw {
1996   my ($node) = @_;
1997   $node =~ tr/A-Z/a-z/;
1998   $node =~ tr/- ./___/;
1999   $node;
2000 }
2001
2002 sub hlp_keywords {
2003   my ($node) = @_;
2004   my $pfx = "K{\\footnote ";
2005   my $done = 0;
2006   foreach $i (0..$#itags) {
2007     (print $pfx,$hlp_index[$i]), $pfx = ";\n", $done++
2008         if $idxnodes{$node,$itags[$i]};
2009   }
2010   print "}\n" if $done;
2011 }
2012
2013 # Make tree structures. $tstruct_* is top-level and global.
2014 sub add_item {
2015   my ($item, $level) = @_;
2016   my $i;
2017
2018   $tstruct_pname{$item} = $pname;
2019   $tstruct_next{$tstruct_previtem} = $item;
2020   $tstruct_prev{$item} = $tstruct_previtem;
2021   $tstruct_level{$item} = $level;
2022   $tstruct_up{$item} = $tstruct_last[$level-1];
2023   $tstruct_mnext{$tstruct_last[$level]} = $item;
2024   $tstruct_last[$level] = $item;
2025   for ($i=$level+1; $i<$MAXLEVEL; $i++) { $tstruct_last[$i] = undef; }
2026   $tstruct_previtem = $item;
2027   push @nodes, $item;
2028 }
2029
2030 # PostScript font metric data. Used for line breaking.
2031 sub font_metrics {
2032   @timesr = (
2033      250,   0,   0,   0,   0,   0,   0,   0,
2034        0,   0,   0,   0,   0,   0,   0,   0,
2035        0,   0,   0,   0,   0,   0,   0,   0,
2036        0,   0,   0,   0,   0,   0,   0,   0,
2037      250, 333, 408, 500, 500, 833, 778, 333,
2038      333, 333, 500, 564, 250, 333, 250, 278,
2039      500, 500, 500, 500, 500, 500, 500, 500,
2040      500, 500, 278, 278, 564, 564, 564, 444,
2041      921, 722, 667, 667, 722, 611, 556, 722,
2042      722, 333, 389, 722, 611, 889, 722, 722,
2043      556, 722, 667, 556, 611, 722, 722, 944,
2044      722, 722, 611, 333, 278, 333, 469, 500,
2045      333, 444, 500, 444, 500, 444, 333, 500,
2046      500, 278, 278, 500, 278, 778, 500, 500,
2047      500, 500, 333, 389, 278, 500, 500, 722,
2048      500, 500, 444, 480, 200, 480, 541,   0,
2049        0,   0,   0,   0,   0,   0,   0,   0,
2050        0,   0,   0,   0,   0,   0,   0,   0,
2051        0,   0,   0,   0,   0,   0,   0,   0,
2052        0,   0,   0,   0,   0,   0,   0,   0,
2053        0, 333, 500, 500, 167, 500, 500, 500,
2054      500, 180, 444, 500, 333, 333, 556, 556,
2055        0, 500, 500, 500, 250,   0, 453, 350,
2056      333, 444, 444, 500,1000,1000,   0, 444,
2057        0, 333, 333, 333, 333, 333, 333, 333,
2058      333,   0, 333, 333,   0, 333, 333, 333,
2059     1000,   0,   0,   0,   0,   0,   0,   0,
2060        0,   0,   0,   0,   0,   0,   0,   0,
2061        0, 889,   0, 276,   0,   0,   0,   0,
2062      611, 722, 889, 310,   0,   0,   0,   0,
2063        0, 667,   0,   0,   0, 278,   0,   0,
2064      278, 500, 722, 500,   0,   0,   0,   0
2065   );
2066   @timesi = (
2067      250,   0,   0,   0,   0,   0,   0,   0,
2068        0,   0,   0,   0,   0,   0,   0,   0,
2069        0,   0,   0,   0,   0,   0,   0,   0,
2070        0,   0,   0,   0,   0,   0,   0,   0,
2071      250, 333, 420, 500, 500, 833, 778, 333,
2072      333, 333, 500, 675, 250, 333, 250, 278,
2073      500, 500, 500, 500, 500, 500, 500, 500,
2074      500, 500, 333, 333, 675, 675, 675, 500,
2075      920, 611, 611, 667, 722, 611, 611, 722,
2076      722, 333, 444, 667, 556, 833, 667, 722,
2077      611, 722, 611, 500, 556, 722, 611, 833,
2078      611, 556, 556, 389, 278, 389, 422, 500,
2079      333, 500, 500, 444, 500, 444, 278, 500,
2080      500, 278, 278, 444, 278, 722, 500, 500,
2081      500, 500, 389, 389, 278, 500, 444, 667,
2082      444, 444, 389, 400, 275, 400, 541,   0,
2083        0,   0,   0,   0,   0,   0,   0,   0,
2084        0,   0,   0,   0,   0,   0,   0,   0,
2085        0,   0,   0,   0,   0,   0,   0,   0,
2086        0,   0,   0,   0,   0,   0,   0,   0,
2087        0, 389, 500, 500, 167, 500, 500, 500,
2088      500, 214, 556, 500, 333, 333, 500, 500,
2089        0, 500, 500, 500, 250,   0, 523, 350,
2090      333, 556, 556, 500, 889,1000,   0, 500,
2091        0, 333, 333, 333, 333, 333, 333, 333,
2092      333,   0, 333, 333,   0, 333, 333, 333,
2093      889,   0,   0,   0,   0,   0,   0,   0,
2094        0,   0,   0,   0,   0,   0,   0,   0,
2095        0, 889,   0, 276,   0,   0,   0,   0,
2096      556, 722, 944, 310,   0,   0,   0,   0,
2097        0, 667,   0,   0,   0, 278,   0,   0,
2098      278, 500, 667, 500,   0,   0,   0,   0
2099   );
2100   @courr = (
2101      600,   0,   0,   0,   0,   0,   0,   0,
2102        0,   0,   0,   0,   0,   0,   0,   0,
2103        0,   0,   0,   0,   0,   0,   0,   0,
2104        0,   0,   0,   0,   0,   0,   0,   0,
2105      600, 600, 600, 600, 600, 600, 600, 600,
2106      600, 600, 600, 600, 600, 600, 600, 600,
2107      600, 600, 600, 600, 600, 600, 600, 600,
2108      600, 600, 600, 600, 600, 600, 600, 600,
2109      600, 600, 600, 600, 600, 600, 600, 600,
2110      600, 600, 600, 600, 600, 600, 600, 600,
2111      600, 600, 600, 600, 600, 600, 600, 600,
2112      600, 600, 600, 600, 600, 600, 600, 600,
2113      600, 600, 600, 600, 600, 600, 600, 600,
2114      600, 600, 600, 600, 600, 600, 600, 600,
2115      600, 600, 600, 600, 600, 600, 600, 600,
2116      600, 600, 600, 600, 600, 600, 600,   0,
2117        0,   0,   0,   0,   0,   0,   0,   0,
2118        0,   0,   0,   0,   0,   0,   0,   0,
2119        0,   0,   0,   0,   0,   0,   0,   0,
2120        0,   0,   0,   0,   0,   0,   0,   0,
2121        0, 600, 600, 600, 600, 600, 600, 600,
2122      600, 600, 600, 600, 600, 600, 600, 600,
2123        0, 600, 600, 600, 600,   0, 600, 600,
2124      600, 600, 600, 600, 600, 600,   0, 600,
2125        0, 600, 600, 600, 600, 600, 600, 600,
2126      600,   0, 600, 600,   0, 600, 600, 600,
2127      600,   0,   0,   0,   0,   0,   0,   0,
2128        0,   0,   0,   0,   0,   0,   0,   0,
2129        0, 600,   0, 600,   0,   0,   0,   0,
2130      600, 600, 600, 600,   0,   0,   0,   0,
2131        0, 600,   0,   0,   0, 600,   0,   0,
2132      600, 600, 600, 600,   0,   0,   0,   0
2133   );
2134 }