o Some of the error messages did not close '<p>' tags.

      - Collector #2002: fixed broken 'ls -R' functionality (didn't
        recurse properly subclasses of OFS.Folder)

Note that it's already automatically put on the PYTHONPATH by Zope, it just
wasn't there yet by default.

        (RNFR now returns 350 as status code instead 250)

- Report number of failures/errors

- Update with Mark's comments:

        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.

      - 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 because the service had no 'redirection
        pipe' anymore (because it's shutting down) to write to,
        causing the service to not restart automatically.

      - Collector #1976: FTP STOR command would load the file being
        uploaded in memory. Changed to use a TemporaryFile.

Collector #1904: On Mac OS X avoid a spurious OSError when zopectl exits.

ObjectManager now has an hasObject method to test presence. This
brings it in line with BTreeFolder.

Makes running CMF and Plone tests a lot less annoying.

Usage: ./bin/zopectl test --keepbytecode --nowarnings --dir Products/CMFPlone

It used to require that test.py be in $ZOPE_HOME/bin/;  now it will use
$ZOPE_HOME as a fallback.

- remove unused import.
- change SimpleItem.py to use python logging module
- log at error when standard_error_message can't be rendered
- give more useful info about errors rendering standard_error_message to the browser.
- all conflict errors are counted and logged at info
- logging makes it clear where the conflict has been resolved and where it hasn't
- errors seen by the user are rendered with standard_error_message and are sent to the error_log which will likely copy them to the event log, depending on the users setup.
- also checking in a functional test for generating write conflict errors.
  
(I'll be committing to CHANGES.TXT shortly, I just wanted to keep the merges simpler)

- all conflict errors are counted and logged at info, as they were before Florent's change
- logging makes it clear where the conflict has been resolved and where it hasn't
- errors seen by the user are rendered with standard_error_message and are sent to the error_log which will likely copy them to the event log, depending on the users setup.
- also checking in a functional test for generating write conflict errors.

(I'll be committing to CHANGES.TXT shortly, I just wanted to keep the merges simpler)

- log at error when standard_error_message can't be rendered
- give more useful info about errors rendering standard_error_message to the browser.

We now check ownership and proxy roles if an executable object is on the stack.

Note that we have removed the C implementation of 'checkPermission'.

Improved logging of ConflictErrors. Now a log is made at level BLATHER
with traceback for any conflict retried. In addition, a log is made at
level ERROR for a conflict that can't be retried anymore and is returned
to the browser as an error. Nothing is logged anymore at level INFO.

of a re-index.

indexes do; this way one can simply re-catalog a site upgraded from a Zope 2.7
instance.

  
o The new tests are amphibious:  they exercise both the Python and the C
  implementations, ensuring that they remain in sync.