contrib/texi2pod.pl

   1 #! /usr/bin/perl -w
   2
   3 # This does trivial (and I mean _trivial_) conversion of Texinfo
   4 # markup to Perl POD format.  It's intended to be used to extract
   5 # something suitable for a manpage from a Texinfo document.
   6
   7 $output = 0;
   8 $skipping = 0;
   9 %sects = ();
  10 $section = "";
  11 @icstack = ();
  12 @endwstack = ();
  13 @skstack = ();
  14 $shift = "";
  15 %defs = ();
  16 $fnno = 1;
  17
  18 while ($_ = shift) {
  19     if (/^-D(.*)$/) {
  20         if ($1 ne "") {
  21             $flag = $1;
  22         } else {
  23             $flag = shift;
  24         }
  25         die "no flag specified for -D\n"
  26             unless $flag ne "";
  27         die "flags may only contain letters, digits, hyphens, and underscores\n"
  28             unless $flag =~ /^[a-zA-Z0-9_-]+$/;
  29         $defs{$flag} = "";
  30     } elsif (/^-/) {
  31         usage();
  32     } else {
  33         $in = $_, next unless defined $in;
  34         $out = $_, next unless defined $out;
  35         usage();
  36     }
  37 }
  38
  39 if (defined $in) {
  40     open(STDIN, $in) or die "opening \"$in\": $!\n";
  41 }
  42 if (defined $out) {
  43     open(STDOUT, ">$out") or die "opening \"$out\": $!\n";
  44 }
  45
  46 while(<STDIN>)
  47 {
  48     # Certain commands are discarded without further processing.
  49     /^\@(?:
  50          [a-z]+index            # @*index: useful only in complete manual
  51          |need                  # @need: useful only in printed manual
  52          |(?:end\s+)?group      # @group .. @end group: ditto
  53          |page                  # @page: ditto
  54          |node                  # @node: useful only in .info file
  55         )\b/x and next;
  56
  57     chomp;
  58
  59     # Look for filename and title markers.
  60     /^\@setfilename\s+([^.]+)/ and $fn = $1, next;
  61     /^\@settitle\s+([^.]+)/ and $tl = $1, next;
  62
  63     # Look for blocks surrounded by @c man begin SECTION ... @c man end.
  64     # This really oughta be @ifman ... @end ifman and the like, but such
  65     # would require rev'ing all other Texinfo translators.
  66     /^\@c man begin ([A-Z]+)/ and $sect = $1, $output = 1, next;
  67     /^\@c man end/ and do {
  68         $sects{$sect} = postprocess($section);
  69         $section = "";
  70         $output = 0;
  71         next;
  72     };
  73     next unless $output;
  74
  75     # Discard comments.  (Can't do it above, because then we'd never see
  76     # @c man lines.)
  77     /^\@c\b/ and next;
  78
  79     # End-block handler goes up here because it needs to operate even
  80     # if we are skipping.
  81     /^\@end\s+([a-z]+)/ and do {
  82         die "\@end $1 without \@$1 at line $.\n" unless defined $endw;
  83         die "\@$endw ended by \@end $1 at line $.\n" unless $1 eq $endw;
  84
  85         $endw = pop @endwstack;
  86
  87         if ($1 =~ /example$/) {
  88             $shift = "";
  89             next;
  90         } elsif ($1 =~ /^(if|ignore|menu)/) {
  91             $skipping = pop @skstack;
  92             next;
  93         } else {
  94             $_ = "\n=back\n";
  95             $ic = pop @icstack;
  96         }
  97     };
  98     next if $skipping;
  99
 100     # Character entities.  First the ones that can be replaced by raw text
 101     # or discarded outright:
 102     s/\@copyright\{\}/(c)/g;
 103     s/\@dots\{\}/.../g;
 104     s/\@enddots\{\}/..../g;
 105     s/\@([.!? ])/$1/g;
 106     s/\@[:-]//g;
 107     s/\@bullet(?:\{\})?/*/g;
 108     s/\@TeX\{\}/TeX/g;
 109     s/\@pounds\{\}/\#/g;
 110     s/\@minus(?:\{\})?/-/g;
 111
 112     # Now the ones that have to be replaced by special escapes
 113     # (which will be turned back into text by unmunge())
 114     s/&/&amp;/g;
 115     s/\@\{/&lbrace;/g;
 116     s/\@\}/&rbrace;/g;
 117     s/\@\@/&at;/g;
 118     # POD doesn't interpret E<> inside a verbatim block.
 119     if ($shift eq "") {
 120         s/</&lt;/g;
 121         s/>/&gt;/g;
 122     } else {
 123         s/</&LT;/g;
 124         s/>/&GT;/g;
 125     }
 126
 127     # Single line command handlers.
 128     /^\@set\s+([a-zA-Z0-9_-]+)\s*(.*)$/ and $defs{$1} = $2, next;
 129     /^\@clear\s+([a-zA-Z0-9_-]+)/ and delete $defs{$1}, next;
 130
 131     /^\@section\s+(.+)$/ and $_ = "\n=head2 $1\n";
 132     /^\@subsection\s+(.+)$/ and $_ = "\n=head3 $1\n";
 133
 134     # Block command handlers:
 135     /^\@ifset\s+([a-zA-Z0-9_-]+)/ and do {
 136         push @endwstack, $endw;
 137         push @skstack, $skipping;
 138         $endw = "ifset";
 139         $skipping = 1 unless exists $defs{$1};
 140         next;
 141     };
 142
 143     /^\@ifclear\s+([a-zA-Z0-9_-]+)/ and do {
 144         push @endwstack, $endw;
 145         push @skstack, $skipping;
 146         $endw = "ifclear";
 147         $skipping = 1 if exists $defs{$1};
 148         next;
 149     };
 150
 151     /^\@(ignore|menu)\b/ and do {
 152         push @endwstack, $endw;
 153         push @skstack, $skipping;
 154         $endw = $1;
 155         $skipping = 1;
 156         next;
 157     };
 158
 159     /^\@itemize\s+(\@[a-z]+|\*|-)/ and do {
 160         push @endwstack, $endw;
 161         push @icstack, $ic;
 162         $ic = $1;
 163         $_ = "\n=over 4\n";
 164         $endw = "itemize";
 165     };
 166
 167     /^\@enumerate(?:\s+([A-Z0-9]+))?/ and do {
 168         push @endwstack, $endw;
 169         push @icstack, $ic;
 170         if (defined $1) {
 171             $ic = $1 . ".";
 172         } else {
 173             $ic = "1.";
 174         }
 175         $_ = "\n=over 4\n";
 176         $endw = "enumerate";
 177     };
 178
 179     /^\@table\s+(\@[a-z]+)/ and do {
 180         push @endwstack, $endw;
 181         push @icstack, $ic;
 182         $ic = $1;
 183         $ic =~ s/\@(?:samp|strong|key)/B/;
 184         $ic =~ s/\@(?:code|kbd)/C/;
 185         $ic =~ s/\@(?:dfn|var|emph|cite|i)/I/;
 186         $ic =~ s/\@(?:file)/F/;
 187         $_ = "\n=over 4\n";
 188         $endw = "table";
 189     };
 190
 191     /^\@((?:small)?example)/ and do {
 192         push @endwstack, $endw;
 193         $endw = $1;
 194         $shift = "\t";
 195         next;
 196     };
 197
 198     /^\@itemx?\s*(.+)?$/ and do {
 199         if (defined $1) {
 200             # Entity escapes prevent munging by the <> processing below.
 201             $_ = "\n=item $ic\&LT;$1\&GT;\n";
 202         } else {
 203             $_ = "\n=item $ic\n";
 204             $ic =~ y/A-Ya-y1-8/B-Zb-z2-9/;
 205         }
 206     };
 207
 208     $section .= $shift.$_."\n";
 209 }
 210
 211 die "No filename or title\n" unless defined $fn && defined $tl;
 212
 213 $sects{NAME} = "$fn \- $tl\n";
 214 $sects{FOOTNOTES} .= "=back\n" if exists $sects{FOOTNOTES};
 215
 216 for $sect (qw(NAME SYNOPSIS DESCRIPTION OPTIONS ENVIRONMENT FILES
 217               BUGS NOTES FOOTNOTES SEEALSO AUTHOR COPYRIGHT)) {
 218     if(exists $sects{$sect}) {
 219         $head = $sect;
 220         $head =~ s/SEEALSO/SEE ALSO/;
 221         print "=head1 $head\n\n";
 222         print scalar unmunge ($sects{$sect});
 223         print "\n";
 224     }
 225 }
 226
 227 sub usage
 228 {
 229     die "usage: $0 [-D toggle...] [infile [outfile]]\n";
 230 }
 231
 232 sub postprocess
 233 {
 234     local $_ = $_[0];
 235
 236     # @value{foo} is replaced by whatever 'foo' is defined as.
 237     s/\@value\{([a-zA-Z0-9_-]+)\}/$defs{$1}/g;
 238
 239     # Formatting commands.
 240     s/\@(?:dfn|var|emph|cite|i)\{([^\}]*)\}/I<$1>/g;
 241     s/\@(?:code|kbd)\{([^\}]*)\}/C<$1>/g;
 242     s/\@(?:samp|strong|key|b)\{([^\}]*)\}/B<$1>/g;
 243     s/\@sc\{([^\}]*)\}/\U$1/g;
 244     s/\@file\{([^\}]*)\}/F<$1>/g;
 245     s/\@w\{([^\}]*)\}/S<$1>/g;
 246     s/\@(?:dmn|math)\{([^\}]*)\}/$1/g;
 247
 248     # Cross references are thrown away, as are @noindent and @refill.
 249     # (@noindent is impossible in .pod, and @refill is unnecessary.)
 250     # @* is also impossible in .pod; we discard it and any newline that
 251     # follows it.
 252
 253     s/\@xref\{(?:[^\}]*)\}[^.]*.//g;
 254     s/\s+\(\@pxref\{(?:[^\}]*)\}\)//g;
 255     s/;\s+\@pxref\{(?:[^\}]*)\}//g;
 256     s/\@noindent\s*//g;
 257     s/\@refill//g;
 258     s/\@\*\s*\n?//g;
 259
 260     # @uref can take one, two, or three arguments, with different
 261     # semantics each time.  @url and @email are just like @uref with
 262     # one argument, for our purposes.
 263     s/\@(?:uref|url|email)\{([^\},]*)\}/&lt;C<$1>&gt;/g;
 264     s/\@uref\{([^\},]*),([^\},]*)\}/$2 (C<$1>)/g;
 265     s/\@uref\{([^\},]*),([^\},]*),([^\},]*)\}/$3/g;
 266
 267     # Turn B<blah I<blah> blah> into B<blah> I<blah> B<blah> to
 268     # match Texinfo semantics of @emph inside @samp.
 269     s/&LT;/</g;
 270     s/&GT;/>/g;
 271     1 while (s/B<([^<>]*)I<([^>]+)>/B<$1>I<$2>B</g);
 272     1 while (s/I<([^<>]*)B<([^>]+)>/I<$1>B<$2>I</g);
 273     s/[BI]<>//g;
 274     s/([BI])<(\s+)([^>]+)>/$2$1<$3>/g;
 275     s/([BI])<([^>]+?)(\s+)>/$1<$2>$3/g;
 276
 277     # Extract footnotes.  This has to be done after all other
 278     # processing because otherwise the regexp will choke on formatting
 279     # inside @footnote.
 280     while (/\@footnote/g) {
 281         s/\@footnote\{([^\}]+)\}/[$fnno]/;
 282         add_footnote($1, $fnno);
 283         $fnno++;
 284     }
 285
 286     return $_;
 287 }
 288
 289 sub unmunge
 290 {
 291     # Replace escaped symbols with their equivalents.
 292     local $_ = $_[0];
 293
 294     s/&lt;/E<lt>/g;
 295     s/&gt;/E<gt>/g;
 296     s/&lbrace;/\{/g;
 297     s/&rbrace;/\}/g;
 298     s/&at;/\@/g;
 299     s/&amp;/&/g;
 300     return $_;
 301 }
 302
 303 sub add_footnote
 304 {
 305     unless (exists $sects{FOOTNOTES}) {
 306         $sects{FOOTNOTES} = "\n=over 4\n\n";
 307     }
 308
 309     $sects{FOOTNOTES} .= "=item $fnno.\n\n"; $fnno++;
 310     $sects{FOOTNOTES} .= $_[0];
 311     $sects{FOOTNOTES} .= "\n\n";
 312 }
 313