From a3e4780b38a2302d2f70f3105c7fe7c5a7327118 Mon Sep 17 00:00:00 2001
From: calvin <calvin@e7d03fd6-7b0d-0410-9947-9c21f3af8025>
Date: Thu, 2 Feb 2006 21:40:57 +0000
Subject: [PATCH] quote attributes with unicode entity escapes

git-svn-id: https://linkchecker.svn.sourceforge.net/svnroot/linkchecker/trunk/linkchecker@3039 e7d03fd6-7b0d-0410-9947-9c21f3af8025
---
 linkcheck/HtmlParser/htmllib.py | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)

diff --git a/linkcheck/HtmlParser/htmllib.py b/linkcheck/HtmlParser/htmllib.py
index f116435f..819d4fa9 100644
--- a/linkcheck/HtmlParser/htmllib.py
+++ b/linkcheck/HtmlParser/htmllib.py
@@ -199,9 +199,19 @@ def quote_attrval (s):
     @return: the quoted HTML attribute
     @rtype: string
     """
-    s = s.replace('&', "&amp;")
-    s = s.replace('"', "&quot;")
-    return s
+    res = []
+    for c in s:
+        if ord(c) <= 127:
+            # ASCII
+            if c == u'&':
+                res.append(u"&amp;")
+            elif c == u'"':
+                res.append(u"&quot;")
+            else:
+                res.append(c)
+        else:
+            res.append(u"&#%d;" % ord(c))
+    return u"".join(res)
 
 
 def quote_val (s):