add entity extractor; haven't fixed entity handling as it's likely to change and I got all kinds of funky issues...

annevk · annevk · commit 0883df6743ae · 2007-06-18T08:40:50.000Z
--HG--
extra : convert_revision : svn%3Aacbfec75-9323-0410-a652-858a13e371e0/trunk%40742
diff --git a/utils/extract-entities.py b/utils/extract-entities.py
@@ -0,0 +1,13 @@
+import urllib2
+entitiesPage = "http://www.whatwg.org/specs/web-apps/current-work/multipage/section-entities.html"
+output = ""
+for line in urllib2.urlopen(entitiesPage).readlines():
+    entityNameSig = "     <td> <code title=\"\">"
+    entityValueSig = "     </td><td> "
+    if line.startswith(entityNameSig):
+        x = len(entityNameSig)
+        output += "    \"" + line[x:-8] + "\": "
+    elif line.startswith(entityValueSig):
+        x = len(entityValueSig)
+        output += "u\"" + line[x:-1].replace("U+", "\\u") + "\",\n"
+print output
