Code review comment for lp:~vila/bzr/822649-bsd-no-fdatasync

>>>>> John A Meinel <email address hidden> writes:

    > Review: Approve
    > On 8/8/2011 2:52 PM, Vincent Ladeuil wrote:
    >> Vincent Ladeuil has proposed merging
    >> lp:~vila/bzr/822649-bsd-no-fdatasync into lp:bzr.
    >>
    >> Requested reviews: bzr-core (bzr-core) Related bugs: Bug #822649 in
    >> Bazaar:
    >> "bzrlib.tests.test_transport.TestLocalTransportWriteStream.test_local_fdatasync_calls_fdatasync
    >> fails on BSDs" https://bugs.launchpad.net/bzr/+bug/822649
    >>
    >> For more details, see:
    >> https://code.launchpad.net/~vila/bzr/822649-bsd-no-fdatasync/+merge/70733
    >>
    >> Don't try to run the test if os.fdatasync is not defined.

    > I'm wondering if it would be better to just teach 'overrideAttr' to not
    > fail when the target attribute doesn't exist.

For other cases, calling *override*Attr for an attribute that doesn't
exist and getting yelled at... sounds appropriate :)

Not failing in this specific case would have masked the issue.

Thanks for the prompt review.