Re: Compile failure on SBCL 1.0.22

On Dec 12, 2008, at 12:28 PM, Keith James wrote:

> On Friday 12 December 2008 18:03:52 Thomas Russ wrote:
>> You should use the most recent snapshot.  A lot of bug fixes have  
>> gone
>> in since our last official release.  In particular, there were a
>> number of changes to make PowerLoom work with the stricter type
>> inference and enforcement of SBCL and CMUCL.
>>
>> The current snapshot compiles for me with SBCL 1.0.12 (I haven't
>> upgraded in a while).
>>
>>> CL-USER> (list (lisp-implementation-version) (machine-type))
>>> ("1.0.22" "X86")
>
> Thanks for the quick reply. I experience the same errors with the  
> 3.2.38
> snapshot.
>
> I'm unable to backup to an older SBCL, so I've set
> *stella-compiler-optimization* to a higher debug level and rebuilt  
> so that I
> can have a look around.

OK.  I downloaded the 1.0.22 version of SBCL and ran into the same  
problem.
There seems to be a bug in our translator, in that it isn't producing  
the correct value.

For a quick workaround, locate the following file and make these  
changes:

    native/lisp/stella/primal.lisp

Change

   (CL:DEFVAR *HASH-BYTE-RANDOM-TABLE* :NULL_VALUE
      ...)

  to

   (CL:DEFVAR *HASH-BYTE-RANDOM-TABLE* #()
      ...)


We'll see about getting a corrected version out soon.


> (Incidentally, I get a clean build with CMUCL 19e.)
>
> -- 
> Keith
> _______________________________________________
> powerloom-forum mailing list
> [email protected]
> http://mailman.isi.edu/mailman/listinfo/powerloom-forum