Doug,

...

> I think the proposed solution is fine for the moment. I
> expect that Ann,
> and probably others, will consider it unacceptable in the
> long run due to
> the exclusion of gstat access. Her mantra is "don't break
> past behavior"
> and operationally it creates problems to lock out gstat.
> Whether it is a
> good trade off is up to each of us to decide. I'll side with
> you - fix the bug.

Actually, I spoke to Ann prior to her departure for a much due 3 week
vacation, and she herself said that she could live with the broken
functionality and was prepared to defend the decision to use exclusive
access.

The bug is so significant that it needs the sledge hammer which we will
be using.

To not use the proposed approach only exposes users to a needless and
significant risk -- which in the long run for a database engine is a
much greater threat than the annoyance of some tools not working.


Sean