[Gcl-devel] Re: load

[Camm Maguire]

Greetings!

Robert Boyer <address@hidden> writes:

> > If in unit one one has
> > (defun foo (x) (bar x))
> > and in unit two
> > (defun bar (x) (declare (fixnum x)) (1+ x))
> > (each of which could be compiled in separate sessions), loading these
> > compiled into the same image will break foo, without an automatic
> > recompile at load time.
> 
> If both foo and bar are compiled within the same
> 'with-compilation-unit', then neither of the two .o files need be
> written at all until after both foo and bar have been thoroughly
> scrutinized.

Agreed.  On the other hand, they still could be in different
compilation units, no?  Then the isse of load conflict remains, but
now has been mitigated.  If you are OK with this, I think proceeding
on to a with-compilation-unit along the lines of our earlier email is
a good idea.  If on the other hand we need to remove all recompiles
from load categorically, then we need a new idea.  Your opinion?

> 
> An approach that sometimes helped was to load the two .lisp files
> containing foo and bar before compiling either.
> 
> This is all very hard and you should ignore my comments unless you like
> them.
> 

I do.  with-compilation-unit should solve the problem for acl2, nqthm,
maxima, axiom ...., leaving possible issues only when one might try to
load both in the same image, and even then, only if they referred to
each other, which itself is unlikely.

Take care,

> Bob
> 
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah