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

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 8/8/2011 6:18 PM, Vincent Ladeuil wrote:
>>>>>> John A Meinel <email address hidden> writes:
>
> <snip/>
>
>> I don't see what issue it is masking.
>
> That, contrary to the dev belief, fdatasync is not available on
> BSDish systems.
>
>> We have a test suite that is not testing that 'fdatasync' is being
>> called when it is available on some platforms.
>
> Huh ? The test ensures that fdatasync is called when it is
> available.
>
> <snip/>
>
>> So I disagree that the "better" fix is to skip the test. I still
>> think the "better" fix is to fake an fdatasync call,
>
> Why would I fake something the python devs didn't implement ?
>
>> and test that it is being called when it is available.
>
> That's what the test does. Please read it.
>

Which *isn't* being run on some platforms. The point is to have the test
suite be as platform agnostic as possible, so that I can run it on
Windows/Linux and not worry that it will fail on a different platform
that I'm *not* running it on.

John
=:->

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk5A7BkACgkQJdeBCYSNAAOlAQCgxK+CA1P92MZ78Y7Js4yuk2a3
p6IAoKTHV39F2ZIpWkWtu/AJU3lZWSj4
=jWfU
-----END PGP SIGNATURE-----