Improving Samba 3 automated testing

[Volker Lendecke]

On Fri, Mar 03, 2006 at 07:17:51AM -0700, Jim McDonough wrote:
> I was thinking that we could have some way of flagging individual tests as 
> to whether they would be for a samba3 run or not.  The default would be to 
> say 'no', and one by one we can allow the tests for samba3 runs.

It's more complicated than that... The Samba4 tests are not
fine-grained enough. The often-cited RAW-OPLOCK test is a
good example: Samba3 is expected to pass all of the subtests
if it has 'kernel oplocks = no', with kernel oplocks under
Linux enabled the LevelII oplock tests fail. Do we want to
cover that kind of scenario?

> I just think trying to periodicaly pull over samba4 torture code into 
> samba3 will be a bigger headache than this.

My opinion also is that we should always run on latest
samba4 svn. If something breaks, it might as well be samba4
that is guilty. And fixing that is something worthwile as
well.

Volker
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.samba.org/archive/samba-technical/attachments/20060303/b34a76eb/attachment.bin