[Axiom-developer] Re: aldor patches

[Peter Broadbery]

On Mon, 2005-08-22 at 23:06 -0500, address@hidden wrote:
> Pete,
> 
> I'm looking at your patches and reading them in detail.
> 
> You mention a shell variable 'ALDORROOT' but don't tell me what
> value it is expected to have (a) at compile time and (b) at runtime.
> 

ALDORROOT is the location of your aldor distribution; it should always
be set to the place where aldor is installed on your machine - in the
NAG days we could install aldor within the axiom tree - I'm guessing
that this isn't viable with different licenses for the two systems. 

Fortunately the two things are fairly loosely coupled; I don't see any
reason for the aldor lisp generator to change dramatically, and axiom's
requirements are pretty stable. Moving versions should be less painful
than changing lisps for example.

> I'm assuming that the aldor compiler will be in 
>   mnt/linux/bin
> and the libraries would be in 
>   mnt/linux/lib
> 
> 
> Also in the Makefile.pamphlet you set:
> 
> INTERPSYS := ${OBJ}/${SYS}/bin/depsys
> 
> Is this what you intended? Why?

Not sure; I've probably just taken liberties with the conventional
names;  It should be depsys of course.  I probably initially used
interpsys, and then ran into a problem and changed the definition and
not the name.

> 
> I did learn the difference between '=' and ':=' in Makefiles from
> your comments so I'll look at back-fitting this into the system
> where appropriate.
> 
> more questions to follow...
> 
> t

OK - will try to reply quickly, but I'm out and about fairly frequently.

NB: As expected, I missed a bunch of files and certain ones need to have
'document' run on them explicitly - I'll try to prepare a makefile which
explicitly does the document thing first; attrib.as.head.pamphlet
contains a minor error in the \author line; & should be a comma or
something.

Peter.

-- 
Peter Broadbery <address@hidden>