improve whitespace matching

fix #36
inukshuk · Aug 26, 2015 · 147ca48 · 147ca48
1 parent 5991d6d
commit 147ca48
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/lib/anystyle/parser/parser.rb b/lib/anystyle/parser/parser.rb
@@ -10,8 +10,8 @@ class Parser
         :pattern => File.expand_path('../support/anystyle.pat', __FILE__),
         :compact => true,
         :threads => 4,
-        :separator => /\s+|\b(\d\S*:)/,
-        :tagged_separator => /\s+|(<\/?[^>]+>)/,
+        :separator => /[[:space:]]+|\b(\d[^[:space:]]*:)/,
+        :tagged_separator => /[[:space:]]+|(<\/?[^>]+>)/,
         :strip => /[^[:alnum:]]/,
         :format => :normalized,
         :xml_entities => Hash[*%w{ &amp; & &lt; < &gt; > &apos; ' &quot; " }],