Date:         Wed, 16 Apr 1997 15:13:07 -0400
Reply-To:     "Rickards, Clinton S" <RickardsCS@AETNA.COM>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         "Rickards, Clinton S" <RickardsCS@AETNA.COM>
Subject:      FW: SCL bug with lists

The question is "is it a bug or is it mis-use?" Using curlist, or
anything else that has at its core "use the current value of ..."
implies that the user (in this case, the programmer) is responsible for
ensuring that all pointers, etc, are correct...
Clint
Clint Rickards, ARS IT, TNB2
Phone: 860-273-3420
 
>----------
>From:  JMS@online.realtech.de[SMTP:JMS@online.realtech.de]
>Sent:  Wed, Apr 16, 1997 6:42 am
>To:    SAS-L@VTVM1.CC.VT.EDU
>Subject:       SCL bug with lists
>
>I found a problem in SCL-lists which is not really a bug,
>but nearly. Look at the following SCL-code:
>
> c=makelist();     /* make a list */
> rc=curlist(c);    /* becomes current list */
>     <other stuff>
> rc=dellist(c);    /* finished - delete it */
>     <other stuff - or other frame entry>
> otherlist=makelist();
>
>Now, otherlist gets the same list identifier as
>the previous list "c". If now a function is invoked
>that writes to the curlist, the contents of
>"otherlist" are destroyed because the curlist
>internal variable still points to the list identifier
>of "c" which now is owned by "otherlist" !
>
>Jens Schlatter
>
>--
>mailto:jms@online.realtech.de
>http://online.realtech.de/schlatter/
>