Consider the following function from src/interp/clam.boot:
numberOfEmptySlots cache==
count:= (CAAR cache ='$failed = 1; 0)
for x in tails rest cache while NE(x,cache) repeat
if CAAR x='$failed then count:= count+1
count
The SBCL type checker does not
04a20e2..d6c0bdd 100644
--- a/changelog
+++ b/changelog
@@ -1,3 +1,4 @@
+20070915 tpd src/algebra/carten minor edit for regression cleanup
20070914 wxh src/hyper/hyper fix bad bracing of )hd change
20070914 tpd src/algebra/fraction.spad remove double )spool command
20070914 tpd src/algebra/kl.spad
Gabriel Dos Reis wrote:
On Sat, 15 Sep 2007, Waldek Hebisch wrote:
| The problem is due to incomplete conversion form old Boot to Shoe.
| Originally clam.boot was translated by old Boot and NE was passed
| unmolested to Lisp output. In vmlisp.lisp we have:
|
| (defmacro ne (a b) `(not
root wrote:
Ed,
I've forwarded your email to Camm Maguire, the GCL maintainer.
It appears to be failing during GCL build. If you are using the
Axiom distribution there is a variable in the top level
Makefile that can be set to different distributions of GCL
(the old distros are
M. Edward (Ed) Borasky wrote:
root wrote:
Ed,
I've forwarded your email to Camm Maguire, the GCL maintainer.
It appears to be failing during GCL build. If you are using the
Axiom distribution there is a variable in the top level
Makefile that can be set to different distributions of GCL
Ed,
M. Edward (Ed) Borasky wrote:
root wrote:
Ed,
I've forwarded your email to Camm Maguire, the GCL maintainer.
It appears to be failing during GCL build. If you are using the
Axiom distribution there is a variable in the top level
Makefile that can be set to different distributions of
root wrote:
Axiom cannot use the installed lisp as it is currently configured.
If you wish you could copy the obj/linux/bin/lisp to your system
location and use it as your GCL image. The only change is that it
links some routines used for communications to sman, which should
be transparent
M. Edward (Ed) Borasky [EMAIL PROTECTED] writes:
| root wrote:
|
|
| Axiom cannot use the installed lisp as it is currently configured.
| If you wish you could copy the obj/linux/bin/lisp to your system
| location and use it as your GCL image. The only change is that it
| links some
Ah -- I'll try that. Meanwhile, the Axiom build with gcl-2.6.7 completed
fine on the AMD64, so I guess 2.6.8pre2 is broken. One minor hitch:
apparently when you do a make install, it's not doing a mkdir -p:
[EMAIL PROTECTED] /home/axiom/silver $ sudo make install
78 installing Axiom in
On Sat, 15 Sep 2007, M. Edward (Ed) Borasky wrote:
It turns out the gcl-2.6.7 tarball is there in zips, so I just edited
the Makefile for 2.6.7 and it's now building. Meanwhile, I would
actually like to use one of the installed Lisps on my systems to build
Axiom -- I have a choice of SBCL,
++
=
diff --git a/changelog b/changelog
index d6c0bdd..9baab4b 100644
--- a/changelog
+++ b/changelog
@@ -1,3 +1,7 @@
+20070915 tpd merge bug100 branch
+20070915 tpd src/input/Makefile add bug100.input regression test
+20070915 tpd src/input
root wrote:
Ah -- I'll try that. Meanwhile, the Axiom build with gcl-2.6.7 completed
fine on the AMD64, so I guess 2.6.8pre2 is broken. One minor hitch:
apparently when you do a make install, it's not doing a mkdir -p:
[EMAIL PROTECTED] /home/axiom/silver $ sudo make install
78 installing
root wrote:
Ah -- I'll try that. Meanwhile, the Axiom build with gcl-2.6.7 completed
fine on the AMD64, so I guess 2.6.8pre2 is broken. One minor hitch:
apparently when you do a make install, it's not doing a mkdir -p:
[EMAIL PROTECTED] /home/axiom/silver $ sudo make install
78 installing
root wrote:
root wrote:
Ah -- I'll try that. Meanwhile, the Axiom build with gcl-2.6.7 completed
fine on the AMD64, so I guess 2.6.8pre2 is broken. One minor hitch:
apparently when you do a make install, it's not doing a mkdir -p:
[EMAIL PROTECTED] /home/axiom/silver $ sudo make install
78
M. Edward (Ed) Borasky wrote:
root wrote:
root wrote:
Ah -- I'll try that. Meanwhile, the Axiom build with gcl-2.6.7 completed
fine on the AMD64, so I guess 2.6.8pre2 is broken. One minor hitch:
apparently when you do a make install, it's not doing a mkdir -p:
[EMAIL PROTECTED]
That should be easy to test, though -- do the whole build as root and
all the environment variables will be set all the way through. It won't
take me long to check it -- it's less than an hour on an Athlon64 X2
5200+ :).
envy :-) send the system to me and i'll test it here :-)
who makes the box?
root wrote:
That should be easy to test, though -- do the whole build as root and
all the environment variables will be set all the way through. It won't
take me long to check it -- it's less than an hour on an Athlon64 X2
5200+ :).
envy :-) send the system to me and i'll test it here :-)
17 matches
Mail list logo