RE: [Axiom-developer] Order of loading of lisp files?

[C Y]

--- Bill Page <address@hidden> wrote:

> CY,
> 
> On July 22, 2006 11:34 AM you wrote:
> > 
> > Is there some way to identify the order in which lisp, boot,
> > and spad files are loaded into the lisp environment once gcl
> > is built?  I'm trying to figure out how to attempt a sbcl build
> > and I'm not quite sure how to "plug in" sbcl instead of the
> > axiom copy of gcl.
> 
> That you ask the question in this way suggests that you have
> a lot to learn about the way Axiom is built! But then I have
> always admired the way you seem to be willing to ask these
> questions. :-)

Well, I guess idiocy has some virtue after all ;-).

> Seriously. If more people asked for things on
> this list, we would be getting a lot further than we have in
> the number of possible Axiom developers.

At a guess, I think we're all still waiting to see if the Aldor
situation sorts itself out.  However, since all the neat toys in the
lisp world seem to run on the more ANSI lisp implementations out there,
a lot of experimentation won't be able to happen until we can join them
there.

> The short answer to your question of course is that you must
> look at the Makefile.pamplet, or rather the set of makefiles
> in each of the root, boot, interp and algebra directories.
> The order is given there both explicitly and to lesser extent
> implicitly, in the manner allowed by make.

I have started trying to do that (it was more involved than I had
guessed - I had to follow several levels of make file down to the src
dir level before I even identified what I needed).  Unfortunately,
while that of course gives me the broad outlines of what I need, there
still seem to be a few gotchas I haven't ironed out.  To say nothing of
trying to tweak the lisp to load in sbcl.  For example, the boot make
file didn't indicate that I need interp/utils.lisp loaded before I
could do boottocl on a file (I'm sure that logic was somewhere but I
missed it the first time around).

> The longer answer is that the current Axiom build process is
> not really organized in a manner that would allow you to just
> "plug in" a different lisp. It might be easier for you to work
> with the Debian version of the build which does not build the
> lisp system from scratch. 

That's a good idea.

> Another thing that you might look at
> is the CMUCL build of Axiom that was done by Juergen Weiss
> about 3 years ago. See the most recent reference to this here:

I took a look at that, but I see I didn't look closely enough.  I
couldn't figure out how to build it, but I now see the Make files are
hidden in the tree.  That's far and away the most promising, since a
lot of the fixes for CMUCL are likely to apply to sbcl as well. 

> There is one statement that Juergen made here in this email
> that has really stuck in my mind. In reply to Tim Daly Jr.
> question regarding the common lisp port:
> 
> > ...
> >> And, if it's been done, how come people are still talking
> >> about getting it done?
> >
> 
> He wrote:
> 
> >  "because people do not listen ;-)..."

Well, I'm embarassed to admit I did download it but couldn't figure out
how to load it. I think I'm making more progress now, although it is
clear 3 years has resulted in enough changes in CMU to need some
updating.

In the interest of listening, does Tim Daly Jr. in fact have an asdf
file for Axiom, however incomplete?  Eventually we will want to be
looking at one if we do things the "lispish" way (and I think that
would be a very good idea - the make files are good for bootstrapping
and managing external requirements, but once we hit the lisp level asdf
is the way to go).  If one is already started I would very much prefer
to work with that than start from scratch ;-)
 
> Well, I am certainly listening, but unfortunately we have not
> heard anything more from either of them since that exchange. :(

I think it has gone no further because no annotated diff has been
prepared for Tim yet ;-).

I'll see if I can run the CMUCL version.

Cheers, and thanks.

CY

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com