Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!ucsd!tut.cis.ohio-state.edu!AUREL.CNS.CALTECH.EDU!bfox
From: bfox@AUREL.CNS.CALTECH.EDU (Brian Fox)
Newsgroups: gnu.bash.bug
Subject: Bind patch errors - free dumps core
Message-ID: <8911142253.AA00915@aurel.cns.caltech.edu.>
Date: 14 Nov 89 22:53:40 GMT
References: <1989Nov14.162036.17956@rpi.edu>
Sender: daemon@tut.cis.ohio-state.edu
Reply-To: bfox@aurel.cns.caltech.edu
Distribution: gnu
Organization: GNUs Not Usenet
Lines: 29


   Date: 14 Nov 89 16:20:36 GMT
   From: rpi!night@brutus.cs.uiuc.edu  (Trip Martin)
   Organization: Rensselaer Polytechnic Institute, Troy NY
   References: <ANDREWT.89Nov13123127@watsnew.waterloo.edu>
   Sender: bug-bash-request@prep.ai.mit.edu

   In <ANDREWT.89Nov13123127@watsnew.waterloo.edu> andrewt@watsnew.waterloo.edu (Andrew Thomas) writes:
   >
   >Regarding the patches posted recently for bind, I have some
   >complaints.  any attempt to rebind a macro, either to nil or to a
   >different key string results in a core dump due to free being called
   >on an invalid pointer and a message to bfox, which I am sure he does
   >not want to see.  Any information on how to fix this would be
   >appreciated.

   Those patches were not meant to be posted.  I was replying to Brian, 
   and didn't notice the cc to gnu-bash-bug.  My apologies for any problems
   it may cause.  I'll be more careful in the future.  

   I'll look into the bug though.  Also, you can't do a macro bind to nil.
   It should give you an error for that.  If you want to disable a key, do 
   a function bind to nil (i.e. no "macro" keyword). 

When fully implemented, it won't give an error binding a macro to nil;
it will make a macro which produces "nil".  You might want to run New
Implementation of Lisp, you know.

Brian