Re: [Axiom-developer] A curious algebra failure

[Martin Rubey]

Gabriel Dos Reis <address@hidden> writes:

> As you suggested, I suspect it should have been written like
> 
>      if S has SetCategory and S has "*"
> 
> The trouble is that we don't want to test just for the name "*".  We
> want to test for the signature
>           
>           "*":(%, %) -> %
> 
> I don't know how to write that in Spad -- Ralf?

I do not know whether it is helpful, but I wrote code like

diffDSF: DIFFSPECSF
diffDSF(s, n) == 
-- I have to help the compiler here a little to choose the right signature...
    if SUP F has _*: (NonNegativeInteger, SUP F) -> SUP F
    then D(s, n)

I don't think that 

    if S has _*: (%, %) -> %

would work, but

    if S has _*: (S, S) -> S

should.

Warning: I didn't check whether this is what you want to fix your bug.


Martin