Download
#!/usr/bin/perl -w
# vi: wm=79:tw=79 :
# 2004-09-22 21:57 EDT $Revision: 1.7 $
# Luis Mondesi <lemsx1@hotmail.com>
#
# DESCRIPTION: Converts a bookmarks.html(firefox/mozilla/netscape) file
# to bookmarks.rss (1.0)
#
# USAGE: bookmarks2rss.pl [OPTIONS] > file.rss
#
# There is no need to edit anything below. To run simply do:
#
# bookmark2rss.pl /path/to/bookmarks.html # will print to STDOUT use -o
# # file.rss to specify where to output
#
# or
#
# bookmarks2rss.pl # which assumes firefox's bookmarks are in:
# # ~/.mozilla/firefox/*default*/bookmarks.html. Will output
# # to STDOUT
#
# The resulting output will be printed to STDOUT unless --output= is given
#
# TODO:
# - should live bookmarks be included? hint: instead of "href" look for FEEDURL attribute
#
# CHANGELOG:
# - 2004-11-22 14:02 EST Added support for finding mozilla's bookmarks.html
# files automatically. More cleanups
# - 2004-11-22 13:24 EST Added better support for HTML::Parser, now XML::RSS
# can include better things, like nice Text names
# - 2004-11-22 13:23 EST Perl 5.8.x fixed utf8 problems. ton of cleanups
# LICENSE: GPL
package Bookmarks2rss;
use vars qw(@ISA);
@ISA = qw(HTML::Parser);
use strict;
use utf8;
$|++;
# standard Perl modules
use Getopt::Long;
Getopt::Long::Configure('bundling');
# non-standard modules
require HTML::Parser;
use XML::RSS;
use POSIX qw(setlocale ctime);
# create XML::RSS object
my $rss = new XML::RSS (version => '1.0');
# create HTML::Parser object:
my $p = new Bookmarks2rss;
# some defaults (to customize from command line)
my $SITE = ""; # --site
my $DESC = "My Firefox Bookmarks";
my $DATE = "";
my $SUBJECT = "bookmarks";
my $CREATOR = ""; # --creator
my $PUBLISHER = "$CREATOR"; # --publisher
my $COPYRIGHT = "";
my $LOCALE = ""; # --language
my $UPDATED = "daily"; # --update-period
my $FREQ = "1"; # --update-frequency
my $UPDATEBASE = "1901-01-01T00:00+00:00";
my $OUTPUT="";
# command line arguments:
GetOptions(
# flags
# strings
's|site=s' => \$SITE,
'd|desc=s' => \$DESC,
'subject=s' => \$SUBJECT,
'creator=s' => \$CREATOR, # email?
'publisher=s' => \$PUBLISHER,
'copyright=s' => \$COPYRIGHT,
'l|language=s' => \$LOCALE,
'updated=s' => \$UPDATED,
'update-frequency=s' => $FREQ,
'update-base' => \$UPDATEBASE,
'o|output=s' => \$OUTPUT,
# numbers
);
# supporting functions:
sub start
{
my($self,$tag,$attr,$attrseq,$orig) = @_;
if ( $tag eq 'a')
{
if ($self->{cur_url} = $attr->{href})
{
$self->{got_href}++;
}
}
}
sub end
{
my ($self,$tag) = @_;
$self->{got_href}-- if ($tag eq 'a' && $self->{got_href} )
}
sub text
{
my ($self,$text ) = @_;
if ($self->{got_href} )
{
# $self->{URLS}{$self->{cur_url}} .= $text;
# Add item:
$rss->add_item( title => "$text", link => $self->{cur_url} );
}
}
sub find_bookmarks
{
my $file = shift;
my @out = ();
if ( defined($file) && -f "$file" )
{
push(@out,$file);
return @out;
} else {
@out = glob ($ENV{HOME}."/.mozilla/firefox/*default*/bookmarks.html");
return @out;
}
return undef; # failed?
}
sub usage_die
{
print STDERR "Usage: $0 [/path/to/bookmarks.html]
";
exit(1);
}
# Parse away!
#
# First, setup our RDF "channel"
# some defaults:
$CREATOR = $PUBLISHER if ( $CREATOR eq "" ); # not too smart but works :-)
$PUBLISHER = $CREATOR if ( $PUBLISHER eq "" );
$SITE = "public" if ( $SITE eq "" ); # hostname --long might be leaking too much info
$LOCALE = $ENV{LANGUAGE} if ( $LOCALE eq "" and exists $ENV{LANGUAGE} );
$LOCALE = $ENV{LANG} if ( $LOCALE eq "" and exists $ENV{LANG} ); # fallback in case $LANGUAGE was not set
$LOCALE = "C" if ( $LOCALE eq "" ); # uff!
POSIX::setlocale( &POSIX::LC_ALL, $LOCALE );
$DATE = ctime(time) if ($DATE eq ""); # ctime format: Sat Nov 19 21:05:57 1994
$rss->channel(
title => "$SITE",
link => "http://$SITE",
description => "$DESC",
dc => {
date => "$DATE",
subject => "$SUBJECT",
creator => "$CREATOR",
publisher => "$PUBLISHER",
rights => "$COPYRIGHT",
language => "$LOCALE",
},
syn => {
updatePeriod => "$UPDATED",
updateFrequency => "$FREQ",
updateBase => "$UPDATEBASE",
},
);
# find all bookmark files and parse all:
my @ifile = find_bookmarks(shift);# || usage_die($!);
foreach my $f (@ifile)
{
if ( -f "$f" ) # redundant? you bet!
{
$p->parse_file($f); # TODO if two or more bookmark files are found,
# the last one would override the content of the
# first parse file. perhaps this should be reported to STDERR?
} else {
print STDERR "$f is not a file!
";
}
}
# print RSS to file or STDOUT
if ( $OUTPUT ne "" )
{
open (OUTPUT,">:utf8","$OUTPUT") || die $!;
print OUTPUT $rss->as_string;
close(OUTPUT);
} else {
## FIXME Perl gives "wide character" warning for UTF-8 locales
print STDOUT $rss->as_string;
}
