Peter Mueller writes in his e-mail of 1993/10/24 [sic]:

Is the function `Normalizer(A,B)' , with `A' and `B' in a common
parent group, supposed to work even if `B' is not a subgroup of `A'?

Yes, it is supposed to work.

He continues

If yes, then the following seems to provide a bug:

[example deleted]

This is indeed a bug. It is still in GAP 3.3. I have looked at the code
and have tried to figure out what goes wrong, but cannot find the
problem. I can also produce smaller examples, but not small enough to
allow a step by step investigation of the backtracking algorithm.
I would have added the temporary hack to use

'Intersection( <U>, Normalizer( Parent(<U>,<V>), <V> )'

except that I'm not even certain that this is really the problem.
I will investigate this problem further.

Martin.

-- .- .-. - .. -.  .-.. --- ...- . ...  .- -. -. .. -.- .-
Martin Sch"onert,   Martin.Schoenert@Math.RWTH-Aachen.DE,   +49 241 804551
Lehrstuhl D f"ur Mathematik, Templergraben 64, RWTH, 52056 Aachen, Germany