.gitignore @@ -1,3 +1,6 @@ +* 0.5.13 - Wed Apr 22 11:12:40 EDT 2009 + - Added :narrow option so find() skips feed_validate and A links. + * 0.5.12 - Fri Mar 20 12:34:48 EDT 2009 - Added support for "feed://" URLs ChangeLog @@ -50,7 +50,7 @@ module Feedbag end end - def self.find(url) + def self.find(url, args = nil) $feeds = [] url_uri = URI.parse(url) @@ -65,20 +65,22 @@ module Feedbag #url = "#{url_uri.scheme or 'http'}://#{url_uri.host}#{url_uri.path}" # check if feed_valid is avail - begin - require "feed_validator" - v = W3C::FeedValidator.new - v.validate_url(url) - return self.add_feed(url, nil) if v.valid? - rescue LoadError - # scoo - rescue REXML::ParseException - # usually indicates timeout - # TODO: actually find out timeout. use Terminator? - $stderr.puts "Feed looked like feed but might not have passed validation or timed out" - rescue => ex - $stderr.puts "#{ex.class} error ocurred with: `#{url}': #{ex.message}" - end + unless args[:narrow] + begin + require "feed_validator" + v = W3C::FeedValidator.new + v.validate_url(url) + return self.add_feed(url, nil) if v.valid? + rescue LoadError + # scoo + rescue REXML::ParseException + # usually indicates timeout + # TODO: actually find out timeout. use Terminator? + $stderr.puts "Feed looked like feed but might not have passed validation or timed out" + rescue => ex + $stderr.puts "#{ex.class} error ocurred with: `#{url}': #{ex.message}" + end + end begin html = open(url) do |f| @@ -102,19 +104,21 @@ module Feedbag end end - (doc/"a").each do |a| - next unless a["href"] - if self.looks_like_feed?(a["href"]) and (a["href"] =~ /\// or a["href"] =~ /#{url_uri.host}/) - self.add_feed(a["href"], url, $base_uri) - end - end - - (doc/"a").each do |a| - next unless a["href"] - if self.looks_like_feed?(a["href"]) - self.add_feed(a["href"], url, $base_uri) - end - end + unless args[:narrow] + (doc/"a").each do |a| + next unless a["href"] + if self.looks_like_feed?(a["href"]) and (a["href"] =~ /\// or a["href"] =~ /#{url_uri.host}/) + self.add_feed(a["href"], url, $base_uri) + end + end + + (doc/"a").each do |a| + next unless a["href"] + if self.looks_like_feed?(a["href"]) + self.add_feed(a["href"], url, $base_uri) + end + end + end end rescue Timeout::Error => err lib/feedbag.rb cd5124a584af96a6fb3c08a56ae8966403c02974 David Moreno david@axiombox.com http://github.com/damog/feedbag/commit/fae3019476757691429d6e502bc516b2483dff40 fae3019476757691429d6e502bc516b2483dff40 2009-04-22T08:14:11-07:00 2009-04-22T08:14:11-07:00 0.5.13 dfe28b7783b96fecc3d6797d4610764058c3ff04 David Moreno david@axiombox.com