Re: STk 2.2 is out!

From: Greg A. Woods <>
Date: Fri, 22 Sep 95 11:16:58 -0400 (EDT)

[ On Fri, September 22, 1995 at 10:15:38 (MDT), Frank Ridderbusch wrote: ]
> Subject: Re: STk 2.2 is out!
> As it turns out, it seems that there is a compiler dependency. I took
> the time and recompiled everything with GCC 2.6.2 and voila, it works
> (as was hinted in another message). I seems, that this dependency was
> introduced sometime during version STk 2.1.2 or 2.1.3. Since that
> time, I couldn't build with my system compiler. Before that, I never
> had this kind of problems.

GRRR!!! I *HATE* ANSI-C prototypes!!! They give programmers far far
too much rope, and we often wrap this rope very firmly around our necks!

Here's the fix to STk 2.2 for the weird failure when loading any STklos

ttyp7:<woods_at_most> $ cvs diff Src/list.c
Index: Src/list.c
RCS file: /local/src-CVS/misc/STk/Src/list.c,v
retrieving revision
diff -r1.1.1.1 list.c
< if (len >= 0) return STk_makeinteger((double) len);
>   if (len >= 0) return STk_makeinteger((long) len);
The one damn bug that would instantly break even an mc68k machine!
Long live lint(1)!  Hip Hip Hurray!
BTW, I'll now start checking in my changes to allow STk to compile on
non-ANSI platforms, and should be able to provide a patch soon.
Note though that I'm not going to lint the code just yet -- a quick look
at just the calls to STk_makeinteger() show there are many, many, many
function parameter casts necessary.  Luckily it's unlikely anyone will
need to run this code on a 16-bit machine, and all the 64-bit machines I
know have ANSI-C compilers natively.
Also, for whatever GCC version failed to compile this correctly (I'm
certain the prototype was in scope!) has a *VERY* serious bug -- don't
use it!!!!
							Greg A. Woods
+1 416 443-1734			VE3TCP			robohack!woods
Planix, Inc. <>; Secrets of the Weird <>
Received on Fri Sep 22 1995 - 17:19:14 CEST

This archive was generated by hypermail 2.3.0 : Mon Jul 21 2014 - 19:38:59 CEST