[jcifs] timezone problems

[Julian Reschke]

Michael B Allen wrote:

> I'm just using copyTo. Specifially I was using examples/CopyTo to copy a
> file from an NT4 machine to a different directory on the same machine. I
> have made some changes regarding timestamps and attributes but I think
> they're limited to directories so you should be able to reproduce the
> behavior using 0.8.2. I can see the byte sequence of the time field coming
> in is the same as what's going out and I still get drift so ....

I'll try that.

Anyway, breaking setLastModified()/getLastModified() in order to fix 
copyTo() seems to be the wrong approach. If both calling sequence need 
different behaviours of the low-level date conversions, this 
definitively needs to be fixed differently.

>>At the end of the day, there should be a set of unit tests tha JCifs
>>should pass; that would discussions like this much easier :-)
> 
> 
> During stability periods I try to run examples/runtests.sh. It works as a
> poor mans test suite. When we determine the correct behavior perhaps we
> will have for this too.

I think we *do* know the correct behaviour for 
getLastModified()/setLastModified() -- it needs to be compatible with 
java.io.File (we had different tests posted to this list).

Regards, Julian

-- 
<green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760