On Nov 26, 2008, at 7:57 PM, Derick Eddington wrote:
On Wed, 2008-11-26 at 12:12 -0800, marcomaggi wrote:
On Nov 26, 10:58 am, Abdulaziz Ghuloum <[EMAIL PROTECTED]> wrote:
Some people might have two
versions of Ikarus: a 32-bit version and a 64-bit version on the
same machine. Modifying the library source codes by embedding
offsets and sizes as constant literals is far from harmless.
I never used a 64 bits system, but it is my understanding that if
one wants to have 2 Ikaruses, one 32b and one 64b, he has to
install both versions of all the foreign C language libraries. Am
I wrong? 2 gmp, 2 libffi, 2 libwhatever...
You're not wrong. But how the OS or C works is not in our control.
So, Ikarus or not one already has a mess to deal with.
We don't need to exacerbate the mess, right? What we should strive
for is trying to minimize the mess.
But there doesn't need to be 2 copies of the Scheme libraries. The
Scheme libraries should use a properly abstracted high-level FFI
library
that knows whether it's on 32-bit or 64-bit and uses this knowledge to
select the corresponding configurations. How these configurations get
created for specific foreign libraries and how the high-level FFI
library finds them and correlates them with the high-level client
Scheme
library defining the foreign interface, remains to be figured out but
it's certainly possible.
Correct on all counts.
That said, let me build a Nausicaa version 0.1 that works
before the end of the year. Will you?
Sure. We just don't want you to waste your time on something other
Schemers aren't going to want to use.
Yes, and I still would like to see a description (document, draft,
whatever. nothing fancy, no ACM formatting) of what this is supposed
to be and how it is supposed to work. Is that possible Marco?
... so we should be patient, and projects are for personal
satisfaction and exploration too.
Yes, and especially in our setting. I mean, nobody is pushing any
deadlines and nobody is under any pressure to push something less
than satisfactory out the door. Neither management nor marketing
is after us, so, we might as well do something worthwhile. Quick
and dirty hacks are not worth much in the long run. (IMHO)
Aziz,,,
