[Gcl-devel] Re: some ANSI compliance bugs maybe?

[Camm Maguire]

Greetings!

Both Paul and I are pretty much continually editing cvs -- he adding
tests and me trying to fix failures.  Occasionally tagged versions get
out of skew like this.  YOu just need to comment three lines in two
files in the ansi-tests directory to get the suite to run, as a
temporary workaround to functions gcl does not yet implement.  Here
are the needed edits:

=============================================================================
Index: ansi-tests/documentation.lsp
===================================================================
RCS file: /cvsroot/gcl/gcl/ansi-tests/documentation.lsp,v
retrieving revision 1.4
diff -u -r1.4 documentation.lsp
--- ansi-tests/documentation.lsp        18 May 2005 12:26:58 -0000      1.4
+++ ansi-tests/documentation.lsp        29 Jun 2005 21:53:59 -0000
@@ -626,8 +626,8 @@
 (defmethod (setf documentation-test-class-1-doc-accessor) ((newdoc string) 
(obj documentation-test-class-1))
   (setf (slot-value obj 'my-doc) newdoc))
 
-(defmethod documentation ((obj documentation-test-class-1) (doctype (eql t)))
-  (documentation-test-class-1-doc-accessor obj))
+;(defmethod documentation ((obj documentation-test-class-1) (doctype (eql t)))
+;  (documentation-test-class-1-doc-accessor obj))
 
 (defmethod (setf documentation) ((newdoc string) (obj 
documentation-test-class-1) (doctype (eql t)))
   (setf (documentation-test-class-1-doc-accessor obj) newdoc))
Index: ansi-tests/load-data-and-control-flow.lsp
===================================================================
RCS file: /cvsroot/gcl/gcl/ansi-tests/load-data-and-control-flow.lsp,v
retrieving revision 1.7
diff -u -r1.7 load-data-and-control-flow.lsp
--- ansi-tests/load-data-and-control-flow.lsp   25 Jun 2005 02:03:12 -0000      
1.7
+++ ansi-tests/load-data-and-control-flow.lsp   29 Jun 2005 21:53:59 -0000
@@ -9,7 +9,7 @@
 (load "return.lsp")
 (load "return-from.lsp")
 (load "defsetf.lsp")
-(load "define-setf-expander.lsp")
+;(load "define-setf-expander.lsp")
 
 (load "and.lsp")
 (load "apply.lsp")
=============================================================================


In cvs head, we're now at:

1822 out of 21164 total tests failed: ....

Take care,

Robert Boyer <address@hidden> writes:

> Now what have I done wrong?
> 
> I tried, or so I thought, the precise recipe you just gave me for looking
> into GCL ANSI compliance, but got this error:
> 
>    % cd ansi-tests && echo '(load "gclload.lsp")' | ../unixport/saved_ansi_gcl
>    GCL (GNU Common Lisp)  2.7.0 ANSI    Jun 27 2005 11:26:37
>    Source License: LGPL(gcl,gmp), GPL(unexec,bfd)
>    Binary License:  GPL due to GPL'ed components: (BFD UNEXEC)
>    Modifications of this banner must retain notice of a compatible license
>    Dedicated to the memory of W. Schelter
> 
>    Use (help) to get some basic information on how to use GCL.
> 
>    >
>    Loading gclload.lsp
>    Loading gclload1.lsp
>    Finished loading gclload1.lsp
>    Error in MULTIPLE-VALUE-LIST [or a callee]: The function 
> DEFINE-SETF-EXPANDER is undefined.
> 
>    Fast links are on: do (use-fast-links nil) for debugging
>    Broken at MULTIPLE-VALUE-LIST.  Type :H for Help.
>     1 (Continue) Retry loading file "define-setf-expander.lsp".
>     2 Retry loading file "load-data-and-control-flow.lsp".
>     3 Retry loading file "gclload2.lsp".
>     4 Retry loading file "gclload.lsp".
>     5 Return to top level.
>    dbl:CL-TEST>>% 
> 
> This in "t2".
> 
...
> 
> As far as I am concerned, there is zero priority to any ANSI related
> complaint I have reported except as it fits into your desired schedule for
> increasing ANSI compliance.  None of the reported bugs are in my way at all.
> 
> By the way, on an unrelated matter, there is something absolutely thrilling
> about seeing this from (room t) on harper:
> 
>    immfix size:          4,611,686,018,427,387,904 fixnums
> 
> Way to go!
> 

Great!  Though this huge amount of space could probably be better used
as heap, and is not due to some unfortunate shared library load
address most likely....  If you ever come across a solution for
moving the shared library start in Linux, I'd alsways be most
interested. 

Take care,

> Bob
> 
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah