"arno" wrote ...
Names are not mentioned in the
article but they could have a problem, too.

They most definitely are a problem.

I tested one of my files with cleaning
and without and I got the same
results (9 queries run 6 times, some
to linked foxpro, some to linked mdb,
some to the open xls)

You have included an open workbook and got a memory leak, so your test
proves nothing we don't already know. From your earlier post, it seems
you may be experiencing a double leak i.e. one from the open xls and
one from the Foxpro source.

Try isolating each data source i.e. one test to query only the
unlinked mdb tables, another to test to query only the foxpro data
(don't bother testing the open xls at all because this is a known
issue).

Have you ever played the board game Clue/Cluedo? If you call Oberst
von Gatow to the Musikzimmer with the Heizungsrohr, you could get
various results but what if you already hold two of those cards...?

Banging your head against the wall is
easy to reproduce but cannot replace a
crashtest for vehicles ;)

I suggest you only need to conduct the test twice: once without the
wall and once with (in that order). That should give you a good idea
whether it is the wall causing the pain. Warning: like the ADO bug,
this could result in memory loss and much leakage <g.

Jamie.

--