[Oberon] Case statements containing base type labels

[Andreas Pirklbauer]

Isn’t the problem that the evaluation order of the Oberon “type" case statement (CASE p OF basetype .. | extendedtype.. | .. END) is not well defined. This inevitably means that different compilers will (necessarily) end up implementing different semantics.. which is bad.

In my view:

a) The type case statement should never have been merged with the “numerical” case statement - not because it makes the parser a little bit easier, but mainly because they are in fact separate statements.

b) The evaluation order of labels should be defined in the language - the order doesn’t necessarily have to be declaration order, it could for example also be defined by the type hierarchy itself (e.g. highest extension level first).