*EXTREMELY* ugly NTP-specific html to mdoc converter. Not perfect, but

does all the manual stuff. Suggestions welcome, criticisms accepted, flames cheerfully /dev/null'd...
2000-03-29 13:52:35 +00:00 · 2000-03-29 13:52:35 +00:00 · c428669622
commit c428669622
parent ccb2265fd2
1 changed files with 145 additions and 0 deletions
--- a/dist/ntp/ntp-html2mdoc.pl
+++ b/dist/ntp/ntp-html2mdoc.pl
@ -0,0 +1,145 @@
+#!/usr/pkg/bin/perl
+
+require "ctime.pl";
+
+$sec = 8;				# XXX
+
+sub swallow {
+	my ($what) = @_;
+
+	while (<>) {
+		chomp;
+		return if /$what/i;
+	}
+}
+
+sub chew {
+	my ($match, $reset) = @_;
+
+	$_ = undef if $reset;
+	while (!(/$match/i)) {
+		chomp($_ .= " " . <>);
+	}
+}
+
+sub detag {
+	s/<[^>]+>//g;
+}
+
+sub dehtmlchar {
+	s/&nbsp;/ /g;
+	s/&lt;/</g;
+	s/&gt;/>/g;
+}
+
+($mday, $mon, $year) = (localtime(time))[3 .. 5];
+$date = sprintf "%s %02d, %d", $ctime'MoY[$mon], $mday, $year + 1900;
+
+swallow("<H3>");
+chomp($_ = <>);
+detag;
+($name, $descr) = split(' ', $_, 2);
+$descr =~ s/^\s*-\s*//;
+($NAME = $name) =~ tr/a-z/A-Z/;
+
+print <<EOF;
+.\\"	\$NetBSD\$
+.\\" Converted from HTML to mandoc by html-to-mdoc.pl
+.\\"
+.Dd $date
+.Dt $NAME $sec
+.Os
+.Sh NAME
+.Nm $name
+.Nd $descr
+.Sh SYNOPSIS
+.Nm
+EOF
+
+swallow("Synopsis");
+chew("</TT>", 1);
+detag;
+s/^\s*$name\s*//;
+@args = split;
+while ($_ = shift @args) {
+	next if /\[/;
+	s/^-//;
+	print ".Op Fl $_";
+	if (defined($args[0]) && $args[0] ne "]") {
+		print " Ar ", shift @args;
+	}
+	print "\n";
+	shift @args;	# remove the "]"
+}
+
+while (<>) {
+	chomp;
+	next if length == 0;
+	if (/^<H4>/i) {
+		chew("</H4>");
+		detag;
+		tr/a-z/A-Z/;
+		s/^\s+//;
+		dehtmlchar;
+		print ".Sh $_\n";
+		next;
+	}
+	if (/<ADDRESS>/i) {
+		chew("</ADDRESS>");
+		detag;
+		s/^\s+//;
+		dehtmlchar;
+		print ".Sh AUTHOR\n$_\n";
+		next;
+	}
+	if (/<DL>/i) {
+		print ".Bl -tag -width indent\n";
+		next;
+	}
+	if (m#</DL>#i) {
+		print ".El\n";
+		next;
+	}
+	if (/<DT>/i) {
+		chew("</DT>");
+		detag;
+		s/^\s*-//;
+		s/ / Ar /;
+		dehtmlchar;
+		print ".It Fl $_\n";
+		next;
+	}
+	if (/<TT>-/) {
+		# command line option
+		chew("</TT>");
+		s#<TT>-([^<]*)#\n.Fl $1\n#ig;
+		s#</TT>##ig;
+		s#<I>([^<]+)</I>#\n.Ar $1\n#ig;
+	}
+	if (/<A HREF/) {
+		# html reference to another ntp page
+		chew("</A>");
+		chomp($_ .= " " . <>) if (/<A$/);	# another reference on the next line
+		s#<TT><A HREF="(.*).htm">\1</A></TT>#\n.Xr \1 $sec\n#ig;
+		s#<A HREF="([^"]*)">(.*?) +</A> *page#\n.%T "$2"\npage in\n.Pa /usr/share/doc/ntp/$1\n#ig;
+		s#^<BR>##g;
+		s#<A HREF="([^"]*)">(.*?)</A>#For\n.%T "$2"\n, refer to\n.Pa /usr/share/doc/ntp/$1 .\n.Pp\n#ig;
+		detag;
+	}
+	s#<TT>$name</TT>#\n.Nm\n#ig;
+	s#<TT>([^<]*)</TT>#\n.Pa $1\n#ig;
+	s#<TT><A HREF=[^>]*>([^<]*)</A></TT>#\n.Pa $1\n#ig;
+	s#<PRE>#.Pp\n.nf\n#ig;
+	s#</PRE>#\n.fi\n.Pp\n#ig;
+	s#<P>#\n.Pp\n#ig;
+	s#<(HR|DD)>##ig;
+	s#</(BODY|HTML|DD)>##ig;
+	s#<I>([^<]+)</I>#\n.Ar $1\n#ig;
+	dehtmlchar;
+	s/^\s+//;
+	s/\n+\s*/\n/g;
+	s/\n.Nm\n([,\.:]) /\n.Nm "" $1\n/g;
+	s/\n\.(Pa|%T) (.*)\n([,\.:]) /\n.$1 $2 $3\n/g;
+	s/\n$//;
+	print $_, "\n" if length > 0;
+}