[Gcl-devel] Re: Still lost

[Camm Maguire]

Greetings, and thanks!  Here is my partially fixed list with the
latest gcl in case you see something different -- will try to fix the
others soon.

Take care,

=============================================================================
(defparameter *terrible-forms*
  '((write-line 0 0)
    (adjust-array 1 2)
    (write-string 0 0)
    (symbol-macrolet (a))
    (untrace 1 2 3)
    (with-accessors '(a) b c 1 2 3 "keywords")
    (typecase '(1 2 3) 4)
    (with-simple-restart 1 2 3 4 5)
;        (ccase 1 2 3)
    (pprint-logical-block 134790128 135401728)
;       (etypecase '(1 2 3) 4)
;       (ctypecase 1 2 3 4 5)
;       (first 0)
;       (second 0)
;       (third 0)
;       (fourth 0)
    (restart-bind 1 2 3 4 5)
    (restart-case (quote ("structure-object-")) "%cclosure-env-nthcdr")
;       (fifth 0)
;       (sixth 0)
;       (seventh 0)
;       (eighth 0)
;       (ninth 0)
;       (tenth 0)
    (ignore-errors (values-list (loop for i from 1 to 100 collect i))) ; 
doesn't ignore the error
;        (macrolet (quote (( nil   nil   nil )))) ; typed four times to the top 
level seg. violation
    (handler-case 0 1 2 3 4 5 6)
    (byte-position  1 )
    (byte-size  1 )
    (deposit-field  1  2  3 )
    (dpb  1  2  3 )
;        (ecase  1  2  3  4  5  6  7  8  9 )
    (proclaim (list (function car)))
    (handler-bind  1  2  3  4  5)
    (ldb  1  2 )
    (ldb-test  1  2 )
    (mask-field  1  2 )
;       (nth  1  2 )
    (coerce (quote (b . 17))(quote (c . 17)))
    (subtypep (quote (b . 17))(quote (c . 17))(quote (d . 17)))
    (read-from-string  1  2 )
    (svref  1  2 )
    (vector-pop  1 )
    (vector-push  1  2 )
    (vector-push-extend  1  2  3 )))
=============================================================================

Robert Boyer <address@hidden> writes:

> Below are a number of simple nonsense forms that cause bad
> things when typed to the read-eval-print-loop of GCL 2.7.0.
> Mostly what one gets are segmentation violations or control
> stack overflows.  The two items with semicolon comments are
> oddities.
> 
> There is much more to be found along these lines, I suspect.
> In particular, it is is necessary to get past the simplest
> checks, e.g., the number of arguments and the proper keyword
> patterns, and then to pass in utter nonsense in just the
> right places to get more serious errors.
> 
> Bob
> 
> (defparameter *terrible-forms*
>   '((write-line 0 0)
>     (adjust-array 1 2)
>     (write-string 0 0)
>     (symbol-macrolet (a))
>     (untrace 1 2 3)
>     (with-accessors '(a) b c 1 2 3 "keywords")
>     (typecase '(1 2 3) 4)
>     (with-simple-restart 1 2 3 4 5)
>     (ccase 1 2 3)
>     (pprint-logical-block 134790128 135401728)
>     (etypecase '(1 2 3) 4)
>     (ctypecase 1 2 3 4 5)
>     (first 0)
>     (second 0)
>     (third 0)
>     (fourth 0)
>     (restart-bind 1 2 3 4 5)
>     (restart-case (quote ("structure-object-")) "%cclosure-env-nthcdr")
>     (fifth 0)
>     (sixth 0)
>     (seventh 0)
>     (eighth 0)
>     (ninth 0)
>     (tenth 0)
>     (ignore-errors (values-list (loop for i from 1 to 100 collect i))) ; 
> doesn't ignore the error
>     (macrolet (quote (( nil   nil   nil )))) ; typed four times to the top 
> level seg. violation
>     (handler-case 0 1 2 3 4 5 6)
>     (byte-position  1 )
>     (byte-size  1 )
>     (deposit-field  1  2  3 )
>     (dpb  1  2  3 )
>     (ecase  1  2  3  4  5  6  7  8  9 )
>     (proclaim (list (function car)))
>     (handler-bind  1  2  3  4  5)
>     (ldb  1  2 )
>     (ldb-test  1  2 )
>     (mask-field  1  2 )
>     (nth  1  2 )
>     (coerce (quote (b . 17))(quote (c . 17)))
>     (subtypep (quote (b . 17))(quote (c . 17))(quote (d . 17)))
>     (read-from-string  1  2 )
>     (svref  1  2 )
>     (vector-pop  1 )
>     (vector-push  1  2 )
>     (vector-push-extend  1  2  3 )))
> 
> 
> 

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