Hello Dmitry,

> Look at the tracker for CORE-115, CORE-142, CORE-85 and CORE-925.
> Some of them mention "<> ANY" and "ALL", but it's the same as "NOT
> IN".

Thanks, I had a look at them and it turns out that the bug was due to
a wrong index choice. That explains why I couldn't reproduce Félix's
results in 1.5.3 - I had set up a simple testcase w/o keys or indices.


Greetings,
Paul Vinkenoog

[Non-text portions of this message have been removed]