Make "make test" startup more reliable

[Volker Lendecke]

On Mon, Apr 11, 2011 at 03:50:53PM +0200, Michael Adam wrote:
> in principle it is very desirable to wrap operations in
> transactions like this. But as we already briefly discussed,
> the layering is wrong, since the calls to
> dcerpc_winreg_int_hklm_openkey() and so on
> between the transaction start/stop calls, do not
> use the regdb directly, not even reg_api, but goes
> the dcerpc path.

What would a better layering look like? The problem is that
every individual registry change requires a tdb transaction.
gencache has the same problem and solved it by two
databases: One transaction-based and one non-transactional.
We could apply a similar scheme and use the winreg_FlushKey
operation to trigger the sync from outside.

Volker

-- 
SerNet GmbH, Bahnhofsallee 1b, 37081 Göttingen
phone: +49-551-370000-0, fax: +49-551-370000-9
AG Göttingen, HRB 2816, GF: Dr. Johannes Loxen