Re: [Axiom-developer] MonoidWithZero

[Bill Page]

On 7/22/07, Franz Lehner wrote:
> > wh-sandbox uses a different bootstrap method than Axiom Silver and
> > build-improvements, so you might need to do things slightly
> > differently. But most important in any case is the contents of the
> > file:
> >
> > wh-sandbox/src/algebra/Makefile.pamphlet
> >
> > grep for 'AMR' (the abbreviation for AbelianMonoidRing) in this file.
> > E.g. in SPADLIST, CATLIST, CATDOMS
> >
> > I think that at least you will need to add your new 'MRCAT' to each of
> > these lists.
> indeed adding MRCAT to SPADLIST and CATLIST makes MRCAT.spad and  MRCAT.o
> appear in the build directory.
> Adding it to CATDOMS triggers an early error when compiling VECTOR.

You are right. I do not think it should be included in CATDOMS.

> However when it comes to compiling MRING, MRCAT is invisible as before:
>
>     initializing NRLIB MRING for MonoidRing
>     compiling into NRLIB MRING
>     cannot produce category object:
> (|MonoidRingCategory| R M)
>     finalizing NRLIB MRING
>     Internal Error
>     Unexpected error in call to system function getSlotFromFunctor
>
>
> > If you are still having trouble I will have time to try this myself
> > only much later today and I will try to give you more specific
> > suggestions then.
> thanks so far, later today will be tomorrow for me ...

After applying your patch and adding MRCAT to SPADLIST and CATLIST, so
far all I can do is confirm the error message that you show above. I
do not understand the cause but I suspect a bootstrap-related error. I
think we need the opinion of the expert here: Waldek, do you
understand the meaning and cause of the message "cannot produce
category object"? Is this a bootstrap problem or perhaps a problem in
the SPAD compiler?

Regards,
Bill Page.