[Openmcl-devel] clozurecl-1.5: maxima-5.20.1 testsuite failures

Gary Byers gb at clozure.com
Sun Apr 25 02:36:32 UTC 2010



On Sun, 25 Apr 2010, Andrey G. Grozin wrote:

> Testsuite failures are unfurprising:
>
> grozin at gdh-zimmer106 ~ $ maxima
> Maxima 5.21.1 http://maxima.sourceforge.net
> using Lisp Clozure Common Lisp Version 1.5  (LinuxX8632)
> Distributed under the GNU Public License. See the file COPYING.
> Dedicated to the memory of William Schelter.
> The function bug_report() provides bug reporting information.
> (%i1) :lisp *read-default-float-format*
> SINGLE-FLOAT
>
> So, this patch has not changed it to 'double-float :-(
>
> Andrey
>

I don't know.  Mailers can wrap long lines and information can generally
get lost in transit, but when I replace the :TOPLEVEL-FUNCTION argument
with something that binds *READ-DEFAULT-FLOAT-FORMAT* to DOUBLE-FLOAT
around a call to CL-USER::RUN, CL-USER::RUN certainly seems to be
called with that binding in effect:


[src/maxima-5.20.1] gb at rinpoche> ccl ./src/binary-openmcl/maxima
Maxima 5.20.1 http://maxima.sourceforge.net
using Lisp Clozure Common Lisp Version 1.5-r13651  (LinuxX8632)
Distributed under the GNU Public License. See the file COPYING.
Dedicated to the memory of William Schelter.
The function bug_report() provides bug reporting information.
(%i1) :lisp *read-default-float-format*
DOUBLE-FLOAT
(%i1)

If that doesn't work for you, then it would seem that either:

1) a simple binding of the special variable *READ-DEFAULT-FLOAT-FORMAT* doesn't
    work as expected

or

2) there's a typo in the patched :TOPLEVEL-FUNCTION argument.

I didn't see such a typo in your earlier mail, but I'd stil be more
inclined to suspect (2) than (1).

It's certainly possible that there's a third explanation here, but I
can't imagine what that'd be.

If we can't figure that out, it might be simpler to change CL-USER::RUN,
which already initializes some other lisp global variables (including
*READ-DEFAULT-FLOAT-FORMAT* on some platforms.)



More information about the Openmcl-devel mailing list