Bring r21771 over from Five's 1.3 branch: Five 1.3 is released.

r21753 | efge | 2006-01-06 18:58:06 +0100 (Fri, 06 Jan 2006)
Fix cleanup of five:traversable.

r21752 | regebro | 2006-01-06 18:51:19 +0100 (Fri, 06 Jan 2006)
If one class was set to have a localsite hook twice, removing the hook
would be attempted twice during the cleanup of unit tests, and the tests
would fail.

the XML preamble is adjusted to the real encoding of the content
as specified through the 'charset' within the content-type
property.

I have so made a hash out of this merge from the chrism-clockserver-merge branch.  I have no idea what I'm doing, clearly.  But this checkin gets us to the place I intended originally.

(Content-Type header now contains a charset.)

      - Collector #1939: When running as a service, Zope could
        potentially collect too much log output filling the NT Event
        Log. When that happened, a 'print' during exception handling
        would cause an IOError in the restart code causing the service
        not to restart automatically.

        Problem is that a service/pythonw.exe process *always* has an
        invalid sys.stdout.  But due to the magic of buffering, small
        print statements would not fail - but once the file actually
        got written to, the error happened.  Never a problem when
        debugging, as the process has a console, and hence a valid
        stdout.

  Merge a couple of fixes that are related to http://www.zope.org/Collectors/Zope/1947
  and http://mail.zope.org/pipermail/zope-dev/2005-December/026236.html from the Five
  repository:
  
  Log message for r21321:
    Explain what this test does and fix misleading and risky usage of pprint
  
  Log message for r21322:
    Fix this test and re-enable it
  
  Log message for r21325:
    Add changes entry for some recently fixed bug
    update Five 1.2 changes history
    bump version number after 1.3c "release"

(Reverted r40251 as I quite obviously didn't know what I was doing.)