#!/usr/bin/perl
#
# info2www - Gateway between GNU Info nodes and WWW
$id = '$Id: info2www,v 1.2.2.9 1996/07/02 08:44:12 lmdrsm Exp $ ';
#
# This is a script conforming to the CGI - Common Gateway Interface
#
# Author:	Roar Smith (lmdrsm@lmd.ericsson.se)
#
# Copyright:	This program is in the Public Domain.
#
# The original code (most of &info2html) was written by 
# Eelco van Asperen (evas@cs.few.eur.nl).
#
# TODO:
# -----
# * Present a list of choices when there is no exact match for the requested
#   Info file but multiple non-exact matches exist.
# 
# * Use Tag Table to find possible file and offset.
#
#

#----------------- CONFIGURATION -----------------------------------------------
#
#
# DEBUG should be set if you want to debug what's happening.
# 
$DEBUG = 0;
#
# DEBUG_PREFIX is prepended to each debug string.
# DEBUG_POSTFIX is appended to each debug string.
# DEBUG_HTMLIFY should be set if you want to HTML'ify the debug output,
# this shouldn't be necessary within comments, but your mileage may vary.
#
$DEBUG_PREFIX  = "<!--\n";	# Alternative suggestion: "<BR><I>"
$DEBUG_POSTFIX = " -->";	# Alternative suggestion: "</I>\n"
$DEBUG_HTMLIFY = 0;		# Alternative suggestion: 1

#
# INFOPATH is the path of direcories in which to search for Info node files.
#
@INFOPATH =				
    (					
     "/usr/info",
     "/usr/local/info"
     );

#
# ALLOWPATH specifies whether info files may be specified with path-names
# outside of those directories included in INFOPATH .
# It is a possible security hole to set this variable to a true value,
# because *any* file on the system could then be accessed through this gateway.
$ALLOWPATH = 0;

#
# ALLOWRELPATH specifies whether info files may be specified with 
# relativ path-names below the directories included in INFOPATH .
$ALLOWRELPATH = 1;

#
# ALTERNATIVE is a map of alternatives - look for the alternative if the node
# itself isn't found.
# The key (first entry) is the node filename, the value (second entry) is the
# alternative. Both are basenames (i.e. no path!) with no capital letters.
# Note that the keys *must* be unique!
#
%ALTERNATIVE =
    (
     'emacs',		'lemacs',
     'g++',		'gcc',
     'c++',		'gcc',
     'gunzip',		'gzip',
     'zcat' ,		'gzip',
     'elisp',		'lispref',
     'features',	'bash'	# Really easy to guess this huh!
     );

#
# Set the PATH so that the ZCAT and GZCAT programs can be found
#
$ENV{'PATH'} .= "/bin:/usr/bin";

#
# ZCAT is the program to use for reading compressed files (*.Z)
# GZCAT is the program to use for reading gzip'ped files (*.gz)
# Both are arrays to be used in an exec() call, with the first element
# being the program (absolute path, or something to be found in PATH)
# and any additional elements being options.
#
# Set either of these to () if you don't want it used.
#
@ZCAT = ("zcat");
@GZCAT = ("gunzip", "-c");

#
# URL of the icons used for indicating references and stuff:
# $INFO_ICON	- Icon at the top left of each document
# $UP_ICON	- Icon used in an "Up:"   hyperlink at the top
# $NEXT_ICON	- Icon used in a  "Next:" hyperlink at the top
# $PREV_ICON	- Icon used in a  "Prev:" hyperlink at the top
# $MENU_ICON	- Icon used in front of each menu label
# $ALIGN	- How to aling the icons
#
# Set these to "" if you don't want them used.
#
$INFO_ICON =	"/doc/info2www/infodoc.jpg";
$UP_ICON =	"/doc/info2www/up.gif";
$NEXT_ICON =	"/doc/info2www/next.gif";
$PREV_ICON =	"/doc/info2www/prev.gif";
$MENU_ICON =	"/doc/info2www/menu.gif";
$ALIGN =	"BOTTOM";

#
# URL for documentation on info2www
#
# Set this to "" if you don't want it used.
#
$DOCREF = "/doc/info2www/info2www.html";

#
# $INPUTFORM specifies whether to have an input form for going to an Info node.
#
# Set this to 0 if you don't want it used.
#
$INPUTFORM = 1;

#
# CACHE is the dbm(3) or ndbm(3) file for caching lookup information.
# Set this to "" if you don't want it used.
# The effective user of this script should have write permissions to
# the directory in which the dbm files reside, or at least to the files
# $CACHE.dir , $CACHE.pag and $CACHE.lock.
#
$CACHE = "/var/tmp/info2www_cache";

#
# Set this to true if you want to lock the lookup-cache dbm(3) files
# while updating lookup information. If flock(2) doesn't work on your
# system, then set this to false.
# You can get a tiny performance increase by unsetting this variable,
# but at the cost of risking damage to the dbm files, which could happen
# if you get simultaneous update attempts since there is no builtin locking
# in dbm - at least not in SunOS 4.x !
#
$CACHE_LOCKING = 1;

#
# These are the defines for file-locking with flock(2)
#
$LOCK_SH = 1; $LOCK_EX = 2; $LOCK_NB = 4; $LOCK_UN = 8;

#
#----------------- CONFIGURATION END -------------------------------------------

#----------------- MAIN --------------------------------------------------------
#
print "Content-type: text/html\n\n"; # Mime header for NCSA httpd
$DEBUG = 1 if (defined $ENV{'DEBUG'});
$DEBUG && &Debug($id);
$pg = $0; $pg =~ s,^.*/([^/]*)$,$1,;
($version, $date) = ($id =~ m@,v\s+([0-9.]+)\s+([0-9/]+)@);
%CACHE = ();
%INPUT = ();
$CACHE_OPENED = 0;
$NFILES = 0;
@INFOPATH = grep(-d, @INFOPATH); # Only search existing directories

$SCRIPT_NAME = $ENV{'SCRIPT_NAME'};
$SERVER_NAME = $ENV{'SERVER_NAME'};
$QUERY_STRING = $ENV{'QUERY_STRING'};
$REQUEST_METHOD = $ENV{'REQUEST_METHOD'};
$PREFIX = $SCRIPT_NAME . "?";	# prefix for HREF= entries

$DEBUG && &Debug("QUERY_STRING: $QUERY_STRING") if (defined $QUERY_STRING);
$DEBUG && &Debug('ARGV: "', join('", "', @ARGV), '"') if @ARGV;

if ($ENV{'REQUEST_METHOD'} eq "POST") {
    read(STDIN, $request, $ENV{'CONTENT_LENGTH'});
    $DEBUG && &Debug("POST: $request");
} elsif ($ENV{'REQUEST_METHOD'} eq "GET" ) {
    if ($QUERY_STRING) {
	$request = $QUERY_STRING;
    }
}
if ($request) {
    # The argument string is encoded in %XX format and must be decoded, but not
    # until split up into key=value pairs: file=gcc&node=Invoking%20GCC
    if ($request =~ /=/) {	# Form created key=value pairs
	%request = &UrlDecode(split(/[&=]/, $request));
	if (!defined $request{'debug'}) {
	    # Do nothing
	} elsif ($request{'debug'} =~ /^Y(es)?$|^On$|^True$/i) {
	    $DEBUG = 1;
	    &Debug("debug=$request{'debug'}\nDEBUG enabled!");
	    &Debug($id);
	    &Debug("QUERY_STRING: $QUERY_STRING") if (defined $QUERY_STRING);
	    &Debug('ARGV: "', join('", "', @ARGV), '"') if @ARGV;
	} elsif ($request{'debug'} =~ /N(o)?$|^Off$|^False$|^$/i) {
	    $DEBUG && &Debug("debug=$request{'debug'}\nIgnored!");
	} else {
	    $DEBUG = 1;
	    &Debug("debug=$request{'debug'}\nSay what???\nDEBUG enabled!");
	    &Debug($id);
	    &Debug("QUERY_STRING: $QUERY_STRING") if (defined $QUERY_STRING);
	    &Debug('ARGV: "', join('", "', @ARGV), '"') if @ARGV;
	}
	if ($nodename = ($request{'query'} || $request{'isindex'})) {
	    if ($nodename !~ /^\(/ && $request{'file'}) {
		$nodename = "(".$request{'file'}.")".$nodename;
	    }
	} elsif ($request{'file'}) {
	    $nodename = "(".$request{'file'}.")".$request{'node'};
	} else {
	    $nodename = "(dir)";
	}
    } else {			# Simple request for a node
	$nodename = &UrlDecode($request);
    }
} elsif (@ARGV) {
    # The argument string is already decoded, bet special characters are
    # backslash escaped: \(gcc\)Invoking\ GCC
    ($nodename = join('+', @ARGV)) =~ s/\\(\W)/$1/g;;
} else {
    $nodename = "(dir)";
}
$nodename = "(dir)" unless $nodename;
$nodename = "(".$nodename unless ($nodename =~ /^\(/);
$nodename = $nodename.")" unless ($nodename =~ /\)/);
$DEBUG && &Debug("Nodename: $nodename\n");
&info2html($nodename);

if ($DOCREF) {
    print
	"<HR>\n",
	"<EM>automatically generated by</EM> ",
	"<A HREF=\"$DOCREF\"><STRONG>$pg</STRONG></A>",
	" <EM>version $version</EM>\n</BODY>\n</HTML>\n";
} else {
    print
	"<HR>\n",
	"<EM>automatically generated by</EM> ",
	"<STRONG>$pg</STRONG>",
	" <EM>version $version</EM>\n</BODY>\n</HTML>\n";
}

exit(0);

#----------------- SUBROUTINES -------------------------------------------------
#

# Handle request for one info-node
sub info2html {
    local($node) = @_;
    local($file, $node_file, $node_name, $fullnode, $link, $linkh, $h_file);
    local($directory, $basefile, $handle, $pos, $entrypos);
    local($cachedfile, $cachedpos);
    local($info_img, $cache, $orglen, $regexp, $menu, $end, $listing, $active);
    local($matches, $blank, $lastblank, $seenMenu, $indirect, $inentry);
    $info_img = "<IMG SRC=\"$INFO_ICON\" ALIGN=$ALIGN ALT=\"\"> " if $INFO_ICON;

    # Nodename looks like one of these:
    # (file)label	- Both file and label of the Info node given
    # (file)		- Label defaults to "Top"
    # 			- File defaults to "dir", Label defaults to "Top"

    if ($node =~ /^\(([^\)]*)\)(.+)$/) {
	($node_file, $node_name) = ($1, $2);
    } elsif ($node =~ /^\(([^\)]*)\)$/) {
	($node_file, $node_name) = ($1, "Top");
    } elsif (!$node) {
	($node_file, $node_name) = ("dir", "Top");
    } else {
	&Error("Malformed node: $node");
	return(&info2html("(dir)Top"));
    }
    $fullnode = "($node_file)$node_name";
    
    ($target = $node_name) =~ tr/A-Z/a-z/;
    ($regexp = $target) =~ s/(\W)/\\$1/g; # Escape special characters
    $DEBUG && &Debug("Nodename: $node\nfile: $node_file\ntarget: $target");

    ($file, $pos) = &TryCache("($node_file)$target", $regexp) if $CACHE;
    if ($file) {
	$cachedfile = $file;
	$cachedpos = $pos;
	($directory, $basefile) = ($cachedfile =~ m|(.*)/([^/]*)$|);
    } else {
	($directory, $basefile) = &FindFile($node_file);
	unless ($basefile) {
	    &Error("Couldn't find Info file \"$node_file\".");
	    &UpdateCache();
	    return(($fullnode =~ /^\(dir\)(Top)?$/i) || &info2html("(dir)Top"));
	}
	$file = "$directory/$basefile";
	$pos = 0;
	unless ($file = &OpenFile($file)) {
	    &Error("Couldn't open Info file \"$node_file\".");
	    &UpdateCache();
	    return(($fullnode =~ /^\(dir\)(Top)?$/i) || &info2html("(dir)Top"));
	}
    }
    # Figure out what file to specify in links to other targets within same file
    $link = $node_file;		# This seems to be the safest choice
    $linkh = &HTMLify($link);	# HTML'ified $link
    
  FileLoop:
    while ($NFILES > 0) {
	$handle = $file;
	$DEBUG && &Debug("Now reading from $handle");
	while ($_ = (shift @INPUT || scalar(<$handle>))) {
	    $orglen = length($_);
	    chop;
	    #study;			# study actually seems to hurt!
	    /^[\037\f]/ && do {
		if ($active) {
		    print "</DL>\n" if $menu; $menu = 0; # End menu
		    print "</PRE>\n" if $listing; $listing = 0; # End text
		    close($handle);
		    $DEBUG && &Debug("Closed file $handle");
		    last FileLoop;
		}
		$active = 0;
		$seenMenu = 0;
		$indirect = 0;
		$inentry = 1;
		$entrypos = $pos;
		next;
	    };
	    next unless $inentry;

	    ($inentry == 1) && do  {
		local($h_node, $h_next, $h_prev, $h_up);
		local($n) = 0;
		/^tag table:/i && do {
		    # we don't use the tag table
		    $inentry = 0;
		    next;
		};
		/^indirect:/i && do {
		    # this entry is a list of filenames to include:
		    #
		    #	gcc.info-1: 1131
		    #	gcc.info-2: 49880
		    #	gcc.info-3: 99426
		    $inentry++;
		    $indirect++;
		    next;
		};

		# top line:
		# File: info,  Node: Add,  Up: Top,  Prev: Expert,  Next: Menus 
		#
		# Parse the header line. If one of the fields
		#	Node: Up: Next: Previous: File:
		# is found, then a variable 'h_node' is set for
		# the field 'node:', 'h_next' for 'next:', etc.
		#
		/\bNode: *([^,\t]*)/i && ($h_node = $1) =~ s/\s+$//;
		/\bUp: *([^,\t]*)/i && ($h_up = $1) =~ s/\s+$//;
		/\bPrev: *([^,\t]*)/i && ($h_prev = $1) =~ s/\s+$//;
		/\bPrevious: *([^,\t]*)/i && ($h_prev = $1) =~ s/\s+$//;
		/\bNext: *([^,\t]*)/i && ($h_next = $1) =~ s/\s+$//;
		
		if ($h_node =~ m/^$regexp$/i) {
		    $active = 1;
		    $matches++;
		    /\bFile: *([^ ,\t]*)/i && ($h_file = $1);
		    $h_file = $node_file unless $h_file;
		    # Update cache if necessary
		    if ($CACHE &&
			(($cachedfile ne $file) ||
			 ($cachedpos ne $entrypos))) {
			$CACHE{"($node_file)$target"} = "$entrypos\0$file";
		    }
		    print
			"<HTML>\n<TITLE>Info Node: ",
			&HTMLify("($h_file)$h_node"),
			"</TITLE>\n<BODY BGCOLOR=#FFFFFF>\n",
			"<H1>",
			"$info_img",
			&HTMLify("($h_file)$h_node"),
			"</H1>\n";
		    print "<FORM METHOD=\"GET\" ACTION=\"$SCRIPT_NAME\">\n"
			if $INPUTFORM;
		    print "<HR>\n";
		    if (defined $h_next) {
			$h_next = &HTMLify($h_next);
			print
			    "Next: <B>",
			    &Anchor($linkh, $h_next, $h_next, $NEXT_ICON, 1),
			    "</B><TT>  </TT>";
			$n++;
		    }
		    if (defined $h_prev) {
			$h_prev = &HTMLify($h_prev);
			print
			    "Prev: <B>",
			    &Anchor($linkh, $h_prev, $h_prev, $PREV_ICON, 1),
			    "</B><TT>  </TT>";
			$n++;
		    }
		    if (defined $h_up) {
			$h_up = &HTMLify($h_up);
			print
			    "Up: <B>",
			    &Anchor($linkh, $h_up, $h_up, $UP_ICON, 1),
			    "</B><TT>  </TT>";
			$n++;
		    }
		    print "<HR>\n";
		    print("<INPUT TYPE=\"hidden\" NAME=\"file\"",
			  " VALUE=\"$linkh\">",
			  "<INPUT TYPE=\"submit\" VALUE=\"Goto:\">",
			  "<INPUT TYPE=\"text\" NAME=\"query\"",
#			  " VALUE=\"$linkh\"",
			  " SIZE=30> ",
			  "Enter <EM>node</EM> , <EM>(file)</EM> ",
			  "or <EM>(file)node</EM>\n")
			if $INPUTFORM;
		    print "</FORM>\n" if $INPUTFORM;
		    print "<PRE>\n" unless $listing; $listing = 1; # Start text
		} elsif ($CACHE) {
		    $CACHE{"($node_file)\L$h_node"} = "$entrypos\0$file";
		}
		$inentry++;
		next;
	    };

	    ($inentry == 2 && $indirect) && do  {
		# each line of this entry consists of two fields,
		# a filename and an offset, separated by a colon.
		# For example:
		#	texinfo-1: 1077
		local($includefile, $offset) = split(/:/);
		unless ($includefile =~ /^\//) {
			$includefile = "$directory/$includefile";
		}
		$DEBUG && &Debug("#include $includefile");
		# should save: $inentry $indirect $pos
		push(@inentry, $inentry);
		push(@indirect, $indirect);
		push(@pos, $pos);
		push(@file, $file);
		$inentry = 0;
		$indirect = 0;
		$pos = 0;
		($file = &OpenFile($includefile)) || return(0);
		next FileLoop;
	    };

	    next unless $active;
	    $_ = &HTMLify($_) if /[<>&]/; # Test added for performance reasons
	    #study;			# study actually seems to hurt!
	    
	    $lastblank = $blank; $blank = 0;
	    /^$/ && do {
		print "\n";
		$blank = 1;
		next;
	    };

	    if (($end) = /^\*\s+Menu:(.*)$/) {
		# start of a menu:
		$seenMenu = 1;
		print "</PRE>\n" if $listing; $listing = 0; # End text
		print "$end";
		print "\n<DL>" unless $menu; $menu = 1; # Start menu
		next;
	    };

	    /^\*/ && do {
		#---- SAMPLE LINES: -----------------------------------------
		# * Sample::.		Sample info.
		#
		# * Info: (info).	Documentation browsing system.
		# 
		# * Bison: (bison/bison)
		# 		A Parser generator in the same style as yacc.
		# * Random: (Random) Random    Random Number Generator
		#------------------------------------------------------------

		if ($menu == 0 && $seenMenu) {
		    print "</PRE></STRONG>\n" if $listing; $listing = 0; # End text
		    print "\n<DL>" unless $menu; $menu = 1; # Start menu
		}

		# * foo::
		/^\*\s+([^:]+)::/ && do {
		    $rest_of_line = $';
		    print
			"<DT>", &Anchor($linkh, $1, $1, $MENU_ICON),
			"<DD>";
		    $rest_of_line =~ s/^[\s\.]+//;
		    print $rest_of_line, "\n";
		    next;
		};

		# * foo: (bar)beer OR (bar)
		/^\*\s+([^:]+):\s*\(([^\) \t\n]+)\)([^\t\n\.,]*)/ && do {
		    $rest_of_line = $';
		    print
			"<DT>", &Anchor($linkh, "($2)$3",$1, $MENU_ICON),
			"<DD>";
		    $rest_of_line =~ s/^[\s\.]+//;
		    print $rest_of_line, "\n";
		    next;
		};

		# * foo: beer.
		/^\*\s+([^:]+):\s*([^\t,\n\.]+)/ && do {
		    $rest_of_line = $';
		    print
			"<DT>", &Anchor($linkh, $2, $1, $MENU_ICON),
			"<DD>", $2, ". ";
		    $rest_of_line =~ s/^[\s\.]+//;
		    print $rest_of_line, "\n";
		    next;
		};

		# no match: ignore silently
	    };

	    $menu && $lastblank && do {
		print "</DL>\n" if $menu; $menu = 0; # End menu
		print "<STRONG><PRE>\n" unless $listing; $listing = 1; # Start text
	    };

	    $menu && do {
		s/^\s+//;
	    };

	    /\*note/i && do {
		# cross reference entry:
		# "*note nodename::."
		# "*note Cross-reference-name: nodename."
		local($n) = 0;
		# There can be multiple notes in a line, so find them all...
		while (1) {
		    # *note \nfoo... (reference split over newline)
		    # *note foo\nbar... (reference split over newline)
		    # *note foo: bar\nbleh... (reference split over newline)
		    if (/\*note\s*$/i ||
			/\*note\s+[^:\.]+$/i ||
			/\*note\s+[^:\.]+:\s+[^:\.\t]+$/i) {
			# Merge with next line
			local($line) = scalar(<$handle>);
			$pos += length($line);
			chop($line);
			$_ .= "\n" . &HTMLify($line);
		    }
		    # *note foo:
		    if (/\*note(\s+)([^:\.]+)::/i) {
			s//:=:NOTE:=:/;	# insert unique (I hope) marker
			local($spc, $ref, $lbl) = ($1, $2, $2);
			local($note) = "<B>Note:</B>$spc";
			$note .= &Anchor($linkh, $ref, $lbl);
			s/:=:NOTE:=:/$note/;
			$n++;
			next;
		    }

		    # * foo: (bar)beer OR (bar)
		    if (/\*note(\s+)([^:]+):\s+\(([^\)\s]+)\)([^\t.,]*)(.?)/i) {
			s//:=:NOTE:=:/;	# insert unique (I hope) marker
			local($spc, $ref, $lbl) = ($1, "($3)$4", "$2$5");
			local($nl) = ($ref =~ /\n/) ? "\n" : "";
			local($note) = "<B>Note:</B>$spc";
			$note .= &Anchor($linkh, $ref, $lbl);
			s/:=:NOTE:=:/$note$nl/;
			$n++;
			next;
		    }

		    # * foo: beer.
		    if (/\*note(\s+)([^:]+):\s+([^\t,\.]+)(.?)/i) {
			s//:=:NOTE:=:/;	# insert unique (I hope) marker
			local($spc, $ref, $lbl) = ($1, $3, "$2$4");
			local($nl) = ($ref =~ /\n/) ? "\n" : "";
			local($note) = "<B>Note:</B>$spc";
			$note .= &Anchor($linkh, $ref, $lbl);
			s/:=:NOTE:=:/$note$nl/;
			$n++;
			next;
		    }

		    last;
		}
	    };

	    print "$_\n";
	} continue {
	    $pos += $orglen unless $active;
	}
	print "</DL>\n" if $menu; $menu = 0; # End menu

	# clear status variables;
	$active = 0;
	$seenMenu = 0;
	$indirect = 0;
	$inentry = 0;
	$lastblank = 0;

	$DEBUG && &Debug("End of file $handle");
	close($handle); $NFILES--;
	$DEBUG && &Debug("Closed file $handle");
	$inentry = pop(@inentry);
	$indirect = pop(@indirect);
	$pos = pop(@pos);
	$file = pop(@file);
	last if $matches;
    }
    while ($file  = pop(@file)) {
	$handle = $file;
	close($handle); $NFILES--;
	$DEBUG && &Debug("Closed file $handle");
	$inentry = pop(@inentry); # Not really necessary
	$indirect = pop(@indirect); # Not really ncessary
	$pos = pop(@pos);	# Not really necessary
    }
    unless ($matches) {
	&Error("Couldn't find target: \"$node_name\" in file \"$node_file\".");
	if ($CACHE && $cachedfile) {
	    $CACHE{"($node_file)$target"} = undef;
	    if ($cachedpos eq "0") {
		$CACHE{"($node_file)"} = undef;
	    }
	}
	&UpdateCache();
	return(($fullnode =~ /\)Top$/i) || &info2html("($node_file)Top"));
    }
    &UpdateCache();
    return($matches);
}

#---------------------------------------------------------------------------

sub UrlDecode {

#	Decode a URL encoded string or array of strings 
#	1.	Change "+" to space, since FORMS change space to "+"
#	2.	Change "%XX" to character with hex value "XX"

    foreach (@_) {
	tr/+/ /;
	s/%(..)/pack("c",hex($1))/ge;
    }
    wantarray ? @_ : $_[$[];
}

sub Anchor {
    local($link, $ref, $label, $icon, $iconlink) = @_;
    local($file, $name, $img, $href);

    $DEBUG && &Debug("Anchor($link, $ref, $label)");
    # (foo)bar
    if ($ref =~ m/^\(([^\)]+)\)\s*([^\t,\.]*)/) {
	$file = $1;
	$name = $2;
    } elsif ($link =~ /^dir$|\/dir$/i) {
	$DEBUG && &Debug("(dir) node - Menu \"$ref\" means \"($ref)\"");
	$file = $ref;
	$name = "";
    } else {
	$file = $link;
	$name = $ref;
    }
    $name =~ s/\s+$//;		# Strip trailing blanks
    $href = "($file)$name";
    # Escape special characters in URL to %XX form.
    # Since encoding is done to %XX form we must first encode "%" itself.
    # The HTML special characters "<", ">" and "&" are already HTML'ified to
    # "&lt;", "&gt;" and "&amp;" so we must *not* further encode "&" here,
    # but for good measure we can encode any "<" and ">" that slip through...
    $href =~ s/%/%25/g;		# %
    $href =~ s/([<>\#\+\?\=\"\\])/sprintf("%%%X",ord($1))/ge;
    $href =~ s/\s+/+/g;		# Encode multiple blanks as a "+" encoded space
    $href = "$PREFIX$href";
    if ($icon) {
	$img = "<IMG SRC=\"$icon\" ALIGN=$ALIGN ALT=\"\*\"> ";
    }
    if ($iconlink) {
	return "<A HREF=\"$href\">$img$label</A>";
    } else {
	return "$img<A HREF=\"$href\">$label</A>";
    }
}

sub HTMLify {
    local($_) = @_;
    s/&/&amp\;/g;
    s/</&lt\;/g;
    s/>/&gt\;/g;
    $_;
}

sub FindFile {
    local($orgname) = @_;
    local($name) = $orgname;
    local($dir, $fil);
    $DEBUG && &Debug("FindFile: \"$name\"");
    
    ($dir, $fil) = &FindFileNoAlt($name);
    if ($dir) {
	$CACHE{"($orgname)"} = "0\0$dir/$fil" unless ($orgname =~ /\//);
	return($dir, $fil);
    }
    # Try a possible alternative...
    $fil = $name;
    $fil =~ s/[-\.]info$//;
    $fil =~ tr/A-Z/a-z/;
    $name = $ALTERNATIVE{$fil};
    $DEBUG && &Debug("\$ALTERNATIVE{$fil} = $name");
    return(undef) unless $name;
    $DEBUG && &Debug("Trying with the alternative \"$name\"...");
    ($dir, $fil) = &FindFileNoAlt($name);
    if ($dir) {
	$CACHE{"($orgname)"} = "0\0$dir/$fil" unless ($orgname =~ /\//);
	$CACHE{"($name)"} = "0\0$dir/$fil" unless ($name =~ /\//);
	return($dir, $fil);
    } else {
	return(undef);
    }
}
   
sub FindFileNoAlt {
    local($name) = @_;
    local($aname) = $name;
    local(@list);
    local($dir, $fil);
    local($regexp, $aregexp);

    $aname =~ s/\.gz$|\.Z$//;
    if ($aname =~ /\.info$/) {
	$aname =~ s/\.info$//;
    } elsif ($aname =~ /-info$/) {
	$aname =~ s/-info$/.info/;
    } else {
	$aname =~ s/$/.info/;
    }
    $DEBUG && &Debug("FindFileNoAlt: \"$name\", Alt=\"$aname\"");

    ($regexp = $name) =~ s/(\W)/\\$1/g; # Escape special characters
    if ($name =~ /\.gz$|\.Z$/) {
	# Don't add gzip'ped and compress file to the regular expression
    } elsif (@GZCAT && @ZCAT) {
	$regexp .= "(\\.gz|\\.Z)?";
    } elsif (@GZCAT) {
	$regexp .= "(\\.gz)?";
    } elsif (@ZCAT) {
	$regexp .= "(\\.Z)?";
    }
    ($aregexp = $aname) =~ s/(\W)/\\$1/g; # Escape special characters
    if (@GZCAT && @ZCAT) {
	$aregexp .= "(\\.gz|\\.Z)?";
    } elsif (@GZCAT) {
	$aregexp .= "(\\.gz)?";
    } elsif (@ZCAT) {
	$aregexp .= "(\\.Z)?";
    }
    $DEBUG && &Debug("\$regexp=/$regexp/ \$aregexp=/$aregexp/");
    # Try absolute match for $name...
    if ($name =~ /\//) {
	($dir, $fil) = ($name =~ m|(.*)/([^/]*)$|);
	if ($ALLOWPATH || grep($_ eq $dir, @INFOPATH)) {
	    @list = ($name);
	    push(@list, "$name.gz") if (@GZCAT && !($name =~ /\.gz$|\.Z$/));
	    push(@list, "$name.Z") if (@ZCAT && !($name =~ /\.gz$|\.Z$/));
	    push(@list, $aname);
	    push(@list, "$aname.gz") if (@GZCAT && !($aname =~ /\.gz$|\.Z$/));
	    push(@list, "$aname.Z") if (@ZCAT && !($aname =~ /\.gz$|\.Z$/));
	    foreach (@list) {
		$DEBUG && &Debug("Trying absolute match for \"$_\"...");
		if (-f $_) {
		    ($dir, $fil) = ($_ =~ m|(.*)/([^/]*)$|);
		    return($dir, $fil);
		}
	    }
	    # Remove path component
	    $name =~ s,^.*/([^/]*)$,$1,;
	    $aname =~ s,^.*/([^/]*)$,$1,;
	    $DEBUG && &Debug("Stripped path from filename: $name");
	} elsif ($ALLOWRELPATH && !($name =~ /^\//) && !($name =~ /\.\./)) {
	    @list = ();
            foreach $dir (@INFOPATH) {
	       push(@list, "$dir/$name");
	       push(@list, "$dir/$name.gz") if (@GZCAT && !($name =~ /\.gz$|\.Z$/));
	       push(@list, "$dir/$name.Z") if (@ZCAT && !($name =~ /\.gz$|\.Z$/));
	       push(@list, "$dir/$aname");
	       push(@list, "$dir/$aname.gz") if (@GZCAT && !($aname =~ /\.gz$|\.Z$/));
	       push(@list, "$dir/$aname.Z") if (@ZCAT && !($aname =~ /\.gz$|\.Z$/));
	    }
	    foreach (@list) {
		$DEBUG && &Debug("Trying absolute match for \"$_\"...");
		if (-f $_) {
		    ($dir, $fil) = ($_ =~ m|(.*)/([^/]*)$|);
		    return($dir, $fil);
		}
	    }
	    # Remove path component
	    $name =~ s,^.*/([^/]*)$,$1,;
	    $aname =~ s,^.*/([^/]*)$,$1,;
	    $DEBUG && &Debug("Stripped path from filename: $name");
	} else {
	    $DEBUG && &Debug("Warning: Absolute path-names not allowed!");
	    $name =~ s,^.*/([^/]*)$,$1,;
	    $aname =~ s,^.*/([^/]*)$,$1,;
	    $DEBUG && &Debug("Stripped path from filename: $name");
	}
    }

    # Try exact match for $name in all directories...
    $DEBUG && &Debug("Trying exact match for \"$name\"...");
    foreach $dir (@INFOPATH) {
	@list = ("$dir/$name");
	push(@list, "$dir/$name.gz") if (@GZCAT && !($name =~ /\.gz$|\.Z$/));
	push(@list, "$dir/$name.Z") if (@ZCAT && !($name =~ /\.gz$|\.Z$/));
	foreach (@list) {
	    $DEBUG && &Debug("Trying exact match for \"$_\"...");
	    if (-f $_) {
		($dir, $fil) = ($_ =~ m|(.*)/([^/]*)$|);
		return($dir, $fil);
	    }
	}
    }
    # Try exact match for $aname in all directories...
    $DEBUG && &Debug("Trying exact match for \"$aname\"...");
    foreach $dir (@INFOPATH) {
	@list = ("$dir/$aname");
	push(@list, "$dir/$aname.gz") if (@GZCAT && !($aname =~ /\.gz$|\.Z$/));
	push(@list, "$dir/$aname.Z") if (@ZCAT && !($aname =~ /\.gz$|\.Z$/));
	foreach (@list) {
	    $DEBUG && &Debug("Trying exact match for \"$_\"...");
	    if (-f $_) {
		($dir, $fil) = ($_ =~ m|(.*)/([^/]*)$|);
		return($dir, $fil);
	    }
	}
    }
    # Try caseless match for $name in all directories...
    $DEBUG && &Debug("Trying caseless match for \"$name\"...");
    @list = ();
    foreach $dir (@INFOPATH) {
	opendir(DIR, $dir);
	push (@list, grep(s/^/$dir\//, sort grep(/^$regexp$/i, readdir(DIR))));
	closedir(DIR);
    }
    if ($#list > 0) {		# One or more matches, return first match
	($dir, $fil) = ($list[0] =~ m|(.*)/([^/]*)$|);
	return($dir, $fil);
    } elsif ($#list == 0) {	# No matches
	($dir, $fil) = ($list[0] =~ m|(.*)/([^/]*)$|);
	return($dir, $fil);
    }
    # Try caseless match for $aname in all directories...
    $DEBUG && &Debug("Trying caseless match for \"$aname\"...");
    @list = ();
    foreach $dir (@INFOPATH) {
	opendir(DIR, $dir);
	push (@list, grep(s/^/$dir\//, sort grep(/^$aregexp$/i, readdir(DIR))));
	closedir(DIR);
    }
    if ($#list > 0) {		# One or more matches, return first match
	($dir, $fil) = ($list[0] =~ m|(.*)/([^/]*)$|);
	return($dir, $fil);
    } elsif ($#list == 0) {	# No matches
	($dir, $fil) = ($list[0] =~ m|(.*)/([^/]*)$|);
	return($dir, $fil);
    }
    # Bummer - no matches at all
    return(undef);
}

sub OpenFile {
    local($filename) = @_;
    local($handle, $pid, $file, $directory);
    
    if ($filename =~ /\//) {
	($directory, $filename) = ($filename =~ m|(.*)/([^/]*)$|);
    }
    $file = "$directory/$filename";
    unless (-f $file) {
	if (@GZCAT && -f "$file.gz") {
	    $filename .= ".gz";
	    $file .= ".gz";
	} elsif (@ZCAT && -f "$file.Z") {
	    $filename .= ".Z";
	    $file .= ".Z";
	} else {
	    $DEBUG && &Debug("No such file: $file");
	    return(undef);
	}
    }
    $DEBUG && &Debug("Trying to open file \"$file\"...");
    $handle = $file;
    if ($filename =~ /\.gz$/) {
	if (@GZCAT) {
	    select((select(STDOUT), $| = 1)[0]); # Non-buffered STDOUT
	    select((select(STDERR), $| = 1)[0]); # Non-buffered STDERR
	    $pid = open($handle, "-|");
	    if ($pid) {		# This is the parent!
		$NFILES++;
		$DEBUG && &Debug("Opened pipe: @GZCAT $file |");
		return($file);
	    } elsif (defined $pid) { # This is the child!
		exec(@GZCAT, $file) || die "Could not exec: $!\n";
	    } else { 		# Pipe failed!
		$DEBUG && &Debug("Could not open pipe: $!");
		return(undef);
	    }
	} else {
	    $DEBUG && &Debug("Cannot use gzip'ped file: $file");
	    return(undef);
	}
    } elsif ($filename =~ /\.Z$/) {
	if (@ZCAT) {
	    select((select(STDOUT), $| = 1)[0]); # Non-buffered STDOUT
	    select((select(STDERR), $| = 1)[0]); # Non-buffered STDERR
	    $pid = open($handle, "-|");
	    if ($pid) {		# This is the parent!
		$NFILES++;
		$DEBUG && &Debug("Opened pipe: @ZCAT $file |");
		return($file);
	    } elsif (defined $pid) { # This is the child!
		exec(@ZCAT, $file) || die "Could not exec: $!\n";
	    } else {		# Pipe failed!
		$DEBUG && &Debug("Could not open pipe: $!");
		return(undef);
	    }
	} else {
	    $DEBUG && &Debug("Cannot use compressed file: $file");
	    return(undef);
	}
    } else {			# Not a compressed or gzip'ped file
	if (open($handle, $file)) {
	    $NFILES++;
	    $DEBUG && &Debug("Opened file \"$file\"");
	    return($file);
	} else {
	    $DEBUG && &Debug("Could not open file: $!");
	    return(undef);
	}
    }
}

# Try to lookup the file and position of the node in the cache
sub TryCache {
    local($cachekey, $regexp) = @_;
    local($handle, $line, $h_node, $pos, $dummy);
    local($cachevalue, $cachedpos, $cachedfile, $cachedir, $newkey, $file);
    undef @INPUT;
    $DEBUG && &Debug("Trying cached entry for \"$cachekey\"...");
    if ($CACHE) {
	unless ($CACHE_OPENED) {
	    if (eval 'dbmopen(%cache, $CACHE, 0644) || die "$!\n"') {
		$CACHE_OPENED = 1;
	    } else {
		$DEBUG && &Debug("Couldn't open cache: $@");
		undef $CACHE;
	    }
	}
	if ($CACHE_OPENED) {
	    $cachevalue = $cache{$cachekey};
	} else {
	    undef $CACHE;
	    return(undef);
	}
    } else {
	undef $CACHE;
	return(undef);
    }
    if (!$cachevalue) {
	if (!$ALLOWRELPATH && ($cachekey =~ m,\(.*/.*\).*,)) {
	    # Remove path and try again
	    ($newkey = $cachekey) =~ s,^\([^\)]*/([^/\)]*)\),($1),;
	    $DEBUG && &Debug("New key: $newkey");
	    return(&TryCache($newkey, $regexp));
	} elsif ($regexp && ($cachekey =~ /^\([^\)]*\).+/)) {
	    # Remove target and try again
	    ($newkey = $cachekey) =~ s,^\(([^\)]*)\).*,($1),;
	    $DEBUG && &Debug("New key: $newkey");
	    return(&TryCache($newkey, undef));
	} else {
	    $DEBUG && &Debug("Cached entry not found!");
	    return(undef);
	}
    }
    ($cachedpos, $cachedfile) = split("\0", $cachevalue);
    $DEBUG && &Debug("Cached entry found: pos=$cachedpos in \"$cachedfile\"");
    if ($cachedfile =~ /\//) {
	($cachedir = $cachedfile) =~ s,(.*)/[^/]*$,$1,;
	if (!$ALLOWPATH && !grep($_ eq $cachedir, @INFOPATH)) {
	    $DEBUG && &Debug("Warning: Absolute path-names not allowed!");
	    $CACHE{$cachekey} = undef;
	    return(undef);
	}
    }
    if ($cachedpos < 0) {
	$DEBUG && &Debug("Warning: Negative cached position ignored!");
	$cachedpos = 0;
	$CACHE{$cachekey} = undef;
    }
    unless ($file = &OpenFile($cachedfile)) {
	$CACHE{$cachekey} = undef;
	return(undef);
    }
    if ($file ne $cachedfile) {
	$CACHE{$cachekey} = "$cachedpos\0$file";
    }
    $handle = $file;
    $DEBUG && &Debug("Now reading from $handle");
    # Seek forward to the cached position by using seek() or read()
    # Note that seek() will not work with a pipe!
    unless (seek($handle, $cachedpos, 0) ||
	    (read($handle, $dummy, $cachedpos) == $cachedpos)) {
	close($handle); $NFILES--;
	$CACHE{$cachekey} = undef;
	return(undef);
    }
    undef $dummy;
    $DEBUG && &Debug("Position: $cachedpos");
    unless ($regexp) {
	return($file, $pos);
    }
    if ($line = <$handle>) {
	push(@INPUT, $line);	# Save line for later
	chop($line);
	$DEBUG && &Debug("line: [$line]");
	if ($line =~ /^[\037\f]/) {
	    $DEBUG && &Debug("Found node start");
	    if ($line = <$handle>) {
		push(@INPUT, $line); # Save line for later
		chop($line);
		$DEBUG && &Debug("line: [$line]");
		if ($line =~ /\bnode: *([^,\t]*)/i) {
		    $h_node = $1;
		    $h_node =~ s/\s+$//; # delete trailing spaces
		    if ($h_node =~ m/^$regexp$/i) {
			$DEBUG && &Debug("Found the node!");
			$pos = $cachedpos;
			return($file, $pos);
		    }
		}
	    }
	}
    }
    undef @INPUT;
    undef $pos;
    $CACHE{$cachekey} = undef;
    close($handle);
    return(undef);
}

# Update the cache lookup DBM database with any saved entries in %CACHE
sub UpdateCache {
    local($key, $value, $pos, $file);
    if ($CACHE && %CACHE && &LockCache()) {
	unless ($CACHE_OPENED) {
	    if (eval 'dbmopen(%cache, $CACHE, 0644) || die "$!\n"') {
		$CACHE_OPENED = 1;
	    } else {
		$DEBUG && &Debug("Couldn't open cache: $@");
		undef $CACHE;
	    }
	}
	if ($CACHE_OPENED) {
	    while (($key, $value) = each %CACHE) {
		if (defined $value) {
		    $cache{$key} = $value;
		    if ($DEBUG) {
			($pos, $file) = split("\0", $value);
			&Debug("cache{$key} set to: pos=$pos in \"$file\"");
		    }
		} else {
		    delete $cache{$key};
		    $DEBUG && &Debug("cache{$key} deleted");
		}
	    }
	    undef %CACHE;
	    eval 'dbmclose(%cache) || die "$!\n'; $CACHE_OPENED = 0;
	    &UnLockCache();
	    return(1);
	} else {
	    $DEBUG && &Debug("Couldn't open DBM file: $!");
	    undef $CACHE;
	    &UnLockCache();
	    return(0);
	}
    } else {
	undef $CACHE;
	return(0);
    }
}

# Lock the lookup cache DBM database
#
# See the dbm(3) manual page. Here is an excerpt from dbm(3) on SunOS 4.1.3:
#
# BUGS
#      ...
#      There are no interlocks and no reliable cache flushing; thus
#      concurrent updating and reading is risky.
#
sub LockCache {
    return(1) unless $CACHE_LOCKING; # Just fake it unless cache locking is used
    local($file) = $CACHE . ".lock";
    unless (open(LOCKFILE, ">$file")) {
	$DEBUG && &Debug("Couldn't open CACHE lockfile \"$file\": $!");
	return(0);
    }
    unless (eval 'flock(LOCKFILE, $LOCK_EX) || die "$!\n"') {
	$DEBUG && &Debug("Couldn't lock CACHE lockfile \"$file\": $@");
	close(LOCKFILE);
	return(0);
    }
    $DEBUG && &Debug("Locked CACHE lockfile \"$file\"");
    return(1);
}

# Unlock the cache lookup DBM database
sub UnLockCache {
    return(1) unless $CACHE_LOCKING; # Just fake it unless cache locking is used
    local($file) = $CACHE . ".lock";
    unless (eval 'flock(LOCKFILE, $LOCK_UN) || die "$!\n"') {
	$DEBUG && &Debug("Couldn't unlock CACHE lockfile \"$file\": $@");
	close(LOCKFILE);
	return(0);
    }
    close(LOCKFILE);
    $DEBUG && &Debug("Unlocked CACHE lockfile \"$file\"");
    return(1);
}

# Print an HTML error message
sub Error {
    local($reason) = @_;
    print "<STRONG>Sorry! - $reason</STRONG>\n<P>\n";
    return(0);
}

# Print debug information if debugging is enabled
sub Debug {
    # Print out text if debugging enabled
    if ($DEBUG) {
	print $DEBUG_PREFIX;
	if ($DEBUG_HTMLIFY) {
	    foreach (@_) {
		print &HTMLify($_);
	    }
	} else {
	    print @_;
	}
	print $DEBUG_POSTFIX;
    }
}

__END__