[ZODB-Dev] Repozo tests -- not
tseaver at palladion.com
Wed Dec 2 11:15:27 EST 2009
-----BEGIN PGP SIGNED MESSAGE-----
Chris Withers wrote:
> Martin Aspeli wrote:
>> - there were no tests for the existing code
> There are tests, they're just manual and spew a lot of output even when
> passing, but they do (as best I could tell) exercise the backup and
> restore cycle quite heavily.
Jim is vetoing any changes to repozo until soembody updates that test to
run as an automated unit test. I find this a strange exercise in
leadership: this tool, which is important to much of the ZODB
community, happens to be one Jim *doesn't use at all* (because of ZRS).
A sufficiently cynical observer might even squint at that and see a
potential conflict of interest.
> I ran these before and after to verify they didn't fail but did exhibit
> the behaviour I was trying to fix before I started and didn't fail *and*
> didn't exhibit the behaviour I was trying to fix after.
>> - Tres wrote a trivial test for the trivial fix
> Indeed. His test exhibited the problem prior to the fix and showed it to
> be fixed after. (since the problem was a warning generated on import)
>> - You now found some different problem in the same code,
> ...but didn't say what it is. Jim seemed to hint it was a Python 3
> problem. I'd warrant there are much more serious problems to be worrying
> about for Python 3 first. If Python 3 compatibility is now a requirement
> for any patches submitted, then we're well and truly fucked.
Sniffing at sys.versioninfo was a silly way to do the fix anyhow (I
think Jim even suggested it): much better would have been to
conditionally import hashlib, and fall back to md6 only where it raises
an import error.
Tres Seaver +1 540-429-0999 tseaver at palladion.com
Palladion Software "Excellence by Design" http://palladion.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
-----END PGP SIGNATURE-----
More information about the ZODB-Dev