Write times fix for S3

[Jeremy Allison]

On Mon, Sep 08, 2008 at 07:48:11AM +0200, Stefan (metze) Metzmacher wrote:
> 
> We don't need special cases for SMBwrite truncate.
> 
> I think SMBwrite truncate, SET_END_OF_FILE and SET_ALLOCATION_SIZE
> just need to be handled like an explicit set of the write time.

I don't think that's the case. It's not sticky like an explicit
set of the write time.

> I'm adding some more code to the torture tests to verify this...

That would be good.

> It's correct that it's never reset! And subsequent SMBwriteX's doesn't
> update the write time!

I'm going to add tests for this.

> Please don't apply your patch as it's not correct.
> We need more torture tests first, before we can think of the
> correct fix.

Sorry, I already did. It passes all the torture tests
we have. I actually have a data model in mind that
would explain this behavior, I'm planning to add
more torture tests to determine if this is the model
that Windows uses and correct the code as needed.

Jeremy.