Raymond Toy <toy.raymond at gmail.com> writes:
> Done. Mostly works, except there's a typo in CMU_WRITE. You have
> "$$(CMU_MAXIMA)". That should have only one $. (It creates
> maxima.core in the file named, literally, "$(CMU_MAXIMA)".
That's odd. I thought $$(FOO) did the delayed expansion thing in Make
(and I think I copied it from the existing Makefile.am). Anyway, since
$(CMU_MAXIMA) apparently works, I'll switch it to that.
> And it seems sharefiles.mk is updated at the end. Shouldn't it be
> done first? Then the value that is used is in the core is not the
> final value.
Ahah! At the moment, "all" depends on both the core and on
sharefiles.mk, so I guess the pair get built in an arbitrary order. I'll
switch things round so the core depends on sharefiles.mk. Thanks.
> All of this sharefiles.mk stuff could go away if we could search
> share/ at start up. I had this working for most lisps, but gcl just
> couldn't be coerced to search the directory in a meaningful way.
Hmm. That would be a nice improvement, I agree. I might spend a bit of
time trying to get it working on GCL (if that's the painful one). I take
it that GCL is implementing CL's pathnames weirdly or something?
> Rupert> Yep, but I don't think this can possibly work. For SBCL,
> Rupert> where I know a little bit about what's going on, you don't
> Rupert> only need a matching SBCL but also the contents of
> Rupert> $SBCL_HOME have to match (a load of .fasl files, which
> Rupert> don't all get loaded immediately, I think). Since we don't
> Rupert> copy all of these, we're doomed to failure anyway so I
> Rupert> don't think we should be bothering trying to copy the
> Rupert> executable.
>
> Hmm. I don't think it will work for cmucl any more either since you
> need at least the unicode data base. But that can be all forced to be
> loaded into the core, so it won't need to be referenced anymore.
>
> I should look into this, but I almost never install, so motivation is
> low. :-(
Likewise. Hmm, I'll spend some time in the next couple of days and make
a couple of patches removing the "fig leaf" stuff we have at the moment,
since it definitely *doesn't* work. I'll append them to the bs2 branch.
Rupert
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 315 bytes
Desc: not available
URL: <http://www.math.utexas.edu/pipermail/maxima/attachments/20130319/ccd1de23/attachment.pgp>;