Used name2codepoint directly; recode it.

[bookmarks_db.git] / Robots / parse_html.py

diff --git a/Robots/parse_html.py b/Robots/parse_html.py
index c7acb6426e7dfbf386702ffb18fe5f12b3bed1ce..680ad5573a741bb99402449a77d969202e49abde 100755 (executable)
--- a/Robots/parse_html.py
+++ b/Robots/parse_html.py
@@ -24,7 +24,7 @@ parsers.append(parse_html)
 
 
 import re
-from htmlentitydefs import entitydefs
+from htmlentitydefs import name2codepoint
 
 entity_re = re.compile("(&\w+;)")
 num_entity_re = re.compile("(&#[0-9]+;)")
@@ -32,20 +32,22 @@ num_entity_re = re.compile("(&#[0-9]+;)")
 def recode_entities(title, charset):
    output = []
    for part in entity_re.split(title):
-      if entity_re.match(part):
-         part = entitydefs.get(part[1:-1], part)
+      if part not in ("&", "<", ">", "&quote;", " ") and \
+            entity_re.match(part):
+         part = unichr(name2codepoint.get(part[1:-1], part)).encode(charset)
       output.append(part)
+   title = ''.join(output)
 
-   output2 = []
-   for part in num_entity_re.split(''.join(output)):
+   output = []
+   for part in num_entity_re.split(title):
       if num_entity_re.match(part):
          try:
             part = unichr(int(part[2:-1])).encode(charset)
          except UnicodeEncodeError:
             pass # Leave the entity as is
-      output2.append(part)
+      output.append(part)
 
-   return ''.join(output2)
+   return ''.join(output)
 
 
 def parse_html(filename, charset=None, log=None):
@@ -56,7 +58,16 @@ def parse_html(filename, charset=None, log=None):
          charset = None         # ...try charset from HTML
 
    for p in parsers:
-      parser = p(filename, charset)
+      charsets = [universal_charset, DEFAULT_CHARSET]
+      if charset not in charsets:
+         charsets.insert(0, charset)
+      parser = None
+      for c in charsets:
+         try:
+            parser = p(filename, c)
+            break
+         except UnicodeEncodeError:
+            pass
       if parser:
          break
       else:
@@ -77,7 +88,7 @@ def parse_html(filename, charset=None, log=None):
          if parser.meta_charset:
             if log: log("   META charset   : %s" % parser.charset)
          else:
-            if log: log("   HTTP charset   : %s" % parser.charset)
+            if log: log("   guessed charset: %s" % parser.charset)
          if log: log("   current charset: %s" % universal_charset)
          if log: log("   title          : %s" % title)
          try:
@@ -105,9 +116,7 @@ if __name__ == '__main__':
    from m_lib.defenc import default_encoding
    current_charset = default_encoding.replace("windows-", "cp")
 
-   parser = parse_html(sys.argv[1], universal_charset)
-   print parser.charset
-   print parser.title
-   print parser.title.decode(universal_charset).encode(current_charset, "replace")
-   print parser.refresh
-   print parser.icon
+   parser = parse_html(sys.argv[1], universal_charset,
+      log=lambda s: sys.stdout.write(s + '\n'))
+   print "   refresh:", parser.refresh
+   print "   icon   :", parser.icon