[cmucl-help] bugs in C::SOURCE-LOCATION and string reversal
Matt Kaufmann
kaufmann at cs.utexas.edu
Thu May 16 15:37:37 UTC 2013
Thanks for dealing with these issues.
Regarding:
>> We also found another issue. If *default-pathname-defaults* is set to
>> some non-default value, the dumped core retains the value. Not sure
>> what the correct behaviour should be for this.
I did a little test with CCL, LispWorks, and GCL and found that they
retain the value too. So apparently it's the application's
responsibility (i.e., my responsibility) to deal with this.
Interestingly, SBCL and Allegro CL apparently do not retain the value,
instead setting it at startup to match the current working directory;
and CLISP seems to set it to #P"" at startup.
I'll probably just arrange that when my application starts up, if
(pathname-directory *default-pathname-defaults*) is non-nil, then
*default-pathname-defaults* is assigned the value of (make-pathname).
Regards,
Matt
From: Raymond Toy <toy.raymond at gmail.com>
Date: Wed, 15 May 2013 20:34:20 -0700
>>>>> "Matt" == Matt Kaufmann <kaufmann at cs.utexas.edu> writes:
Matt> The first bug pertains to C::SOURCE-LOCATION. Below is part of the
Matt> log, showing three errors during compilation followed by a break
Matt> during load of the resulting compiled file.
Matt> ;
Matt> ;
Matt> ; File: /v/filer4b/v11q002/acl2space/acl2/devel/books/centaur/gl/gl at expansion.lsp
Matt> ; In: DEFMACRO GL-BDD-MODE => DEFATTACH GL::BFR-MODE
Matt> ; (DEFATTACH GL::BFR-MODE GL::BFR-BDD)
Matt> ; --> DEFPARAMETER PROGN LISP::SET-DEFVAR-SOURCE-LOCATION
Matt> ; --> LISP::SET-DEFVAR-SOURCE-LOCATION
Matt> ; ==>
Matt> ; (C::SOURCE-LOCATION)
Matt> ; Error: (during macroexpansion)
Matt> ; Type-error in KERNEL::OBJECT-NOT-TYPE-ERROR-HANDLER:
Matt> ; 27574 is not of type (UNSIGNED-BYTE 14)
With help from Matt, I've been able to reproduce this. I think the
issue is that the file gl at expansion.lsp is a 500 KB source file with
way too many forms in it and exceeding the limit 14 bits that the
compiler allocates for the top-level form number and the form number.
We also found another issue. If *default-pathname-defaults* is set to
some non-default value, the dumped core retains the value. Not sure
what the correct behaviour should be for this.
Ray
_______________________________________________
cmucl-help mailing list
cmucl-help at cmucl.cons.org
http://lists.zs64.net/mailman/listinfo/cmucl-help
More information about the cmucl-help
mailing list