diff --git a/ChangeLog b/ChangeLog index 9c41c6f400..33bb1489af 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,3 +1,9 @@ +Mon Dec 8 13:02:11 2003 Minero Aoki + + * lib/uri/common.rb: new method URI.regexp. [ruby-dev:22121] + + * test/uri/test_common.rb: add test for URI.regexp. + Mon Dec 8 12:44:14 2003 Yukihiro Matsumoto * pack.c: define swap16 and swap32 only if they are not diff --git a/lib/uri/common.rb b/lib/uri/common.rb index 5d6a3b5519..fbb3558a32 100644 --- a/lib/uri/common.rb +++ b/lib/uri/common.rb @@ -396,28 +396,46 @@ module URI --- URI::extract(str[, schemes]) =end - def self.extract(str, schemes = []) - urls = [] - regexp = ABS_URI_REF - unless schemes.empty? - regexp = Regexp.new('(?=' + schemes.collect{|s| - Regexp.quote(s + ':') - }.join('|') + ')' + PATTERN::X_ABS_URI, - Regexp::EXTENDED, 'N') - end - - str.scan(regexp) { - if block_given? - yield($&) - else - urls << $& - end - } - + def self.extract(str, schemes = nil, &block) if block_given? - return nil + str.scan(regexp(schemes)) { yield $& } + nil else - return urls + result = [] + str.scan(regexp(schemes)) { result.push $& } + result + end + end + +=begin + +--- URI::regexp([match_schemes]) + + Returns a Regexp object which matches to URI-like strings. + If MATCH_SCHEMES given, resulting regexp matches to URIs + whose scheme is one of the MATCH_SCHEMES. + + The Regexp object returned by this method includes arbitrary + number of capture group (parentheses). Never rely on its + number. + + # extract first URI from html_string + html_string.slice(URI.regexp) + + # remove ftp URIs + html_string.sub(URI.regexp(['ftp']) + + # You should not rely on the number of parentheses + html_string.scan(URI.regexp) do |*matches| + p $& + end + +=end + def self.regexp(schemes = nil) + unless schemes + ABS_URI_REF + else + /(?=#{Regexp.union(*schemes)}:)#{PATTERN::X_ABS_URI}/xn end end