From: Giuseppe Bilotta <giuseppe.bilotta@gmail.com>
Date: Thu, 20 Nov 2008 14:17:27 +0000 (+0100)
Subject: rss plugin: prevent double UTF-8 deconding
X-Git-Url: https://git.netwichtig.de/gitweb/?a=commitdiff_plain;h=ac9eb1b02d4200006566ccd630dd678345008963;p=user%2Fhenk%2Fcode%2Fruby%2Frbot.git

rss plugin: prevent double UTF-8 deconding

The rss parser looks at the encoding specified into the XML file and
converts everything to UTF-8. Since we do the UTF-8 conversion
ourselves, monkey-patch the XML 'encoding' declaration to claim it's
UTF-8 already (as it actually is).
---

diff --git a/data/rbot/plugins/rss.rb b/data/rbot/plugins/rss.rb
index 9e85b416..45ee4a23 100644
--- a/data/rbot/plugins/rss.rb
+++ b/data/rbot/plugins/rss.rb
@@ -1092,6 +1092,13 @@ class RSSFeedsPlugin < Plugin
     # reassign the 0.9 RDFs to 1.0, and hope it goes right.
     xml.gsub!("xmlns=\"http://my.netscape.com/rdf/simple/0.9/\"",
               "xmlns=\"http://purl.org/rss/1.0/\"")
+    # make sure the parser doesn't double-convert in case the feed is not UTF-8
+    xml.sub!(/<\?xml (.*?)\?>/) do |match|
+      if /\bencoding=(['"])(.*?)\1/.match(match)
+        match.sub!(/\bencoding=(['"])(?:.*?)\1/,'encoding="UTF-8"')
+      end
+      match
+    end
     feed.mutex.synchronize do
       feed.xml = xml
     end