Re: [Gcl-devel] ecase/ccase

[Paul F. Dietz]

Camm Maguire wrote:
> Greetings!  Just committed a few fixes.  Paul, are these tests right?
> Isn't the gcl behavior now correct?
>
> Test CL-TEST::ECASE.4 failed
> Form: (CL-TEST::CLASSIFY-ERROR (ECASE 1 (T NIL)))
> Expected value: TYPE-ERROR
> Actual value: NIL.
> Test CL-TEST::ECASE.5 failed
> Form: (CL-TEST::CLASSIFY-ERROR  (ECASE 1 (OTHERWISE NIL)))
> Expected value: TYPE-ERROR
> Actual value: NIL.
>  CL-TEST::ECASE.6 CL-TEST::ECASE.7 CL-TEST::ECASE.8 CL-TEST::ECASE.9
>  CL-TEST::ECASE.10 CL-TEST::ECASE.11 CL-TEST::ECASE.12
>  CL-TEST::ECASE.13 CL-TEST::ECASE.14 CL-TEST::ECASE.15
>  CL-TEST::ECASE.16
> Test CL-TEST::ECASE.17 failed
> Form: (CL-TEST::CLASSIFY-ERROR
>           (ECASE 'CL-TEST::A (CL-TEST::B 0) (CL-TEST::C 1)
>                  (OTHERWISE 2)))
> Expected value: TYPE-ERROR
> Actual value: 2.


ECASE doesn't take an OTHERWISE clause.  OTHERWISE is just another
constant there.  Since A is not EQ to OTHERWISE, it should not match
that clause.  (Similarly for T in ECASE.4.)

        Paul