Re: todo for 0.14.4

On Thu, 11 Aug 2005, bryan o'connor wrote:

>>   low-hanging fruit.
>
> here are two that i haven't figured out yet plus another
> vector-push-extend regression.  there are a few more
> print-length/print-level failures that i'm working on now
> that i'd like to clean up before 0.14.4.

Great; thanks.

>
> 1.
>
>> (add-method #'print-object)
>
> should throw program-error because the wrong number of
> args was given.  num arg checking seems to be bypassed.
>

Welcome to OpenMCL Version (Beta: DarwinPPC32) 0.14.4-pre-050809a!
? (add-method #'print-object)
> Error in process listener(1): Too few args to: #<STANDARD-GENERIC-FUNCTION ADD-METHOD #x8074E56>
> While executing: NTH-ARG-DCODE-TOO-FEW-ARGS
> Type :POP to abort.
Type :? for other options.

It apparently catches this at least some of the time; I wonder if there's
some path through the discriminating code that skips the check.

>
> 2.
>
>> (documentation #'hash-table-p t)
> (NIL LAMBDA (CCL::HASH) (DECLARE (CCL::GLOBAL-FUNCTION-NAME HASH-TABLE-P))
> (BLOCK HASH-TABLE-P (CCL::ISTRUCT-TYPEP CCL::HASH 'HASH-TABLE)))
>
> seems to happen because #'hash-table-p is declared inline.

Ugh.  Fixed in CVS soon.

>
>
> 3.
>
> vector types 174 (unsigned-byte >32)? and 158 are missing
> in %uvector-replace on ppc64.

Fixed earlier today, thanks.

>
> _______________________________________________
> Bug-openmcl mailing list
> [email protected]
> http://clozure.com/mailman/listinfo/bug-openmcl
>
>