THE GFA3.0 BUGHUNT by Stefan Posthuma
This article consists of previously published material. All
credits go to the Stichting ST. The material is used with their
After long labours, Frank Ostrowski created his GfA BASIC 3.0.
This program was much honoured after its release, and Frank
Ostrowksi became Immortal in the world of the ST. But alas! The
program contains some faults that are able to grief and hinder
the one that is working with it.
Let's cut the Tolkien jazz for a sec.
As said above, GfA 3.0 contains bugs. The known bugs sofar are:
- register A7 will not be dumped when bombs occur.
- fourth parameter of DEFLINE is a dummy.
- arrays of booleans are +1 when they are TRUE.
- BYTE arrays have several problems (when INTS occur)
- string comparison sometimes fails when comparing strings of
- bombs of system crash during SSORT with strings.
- SSORT on BYTE arrays doesn't work out, QSORT on BYTE arrays
- string comparison sometimes fails during sort.
- INTEGER and WORD arrays are sorted the wrong way.
- SINQ() and ACOS() bomb out when they are used as an integer.
- a!=(2=3) gives TRUE
- INSERT and DELETE bomb when used on byte array and screw up
immensely when used on string arrays (DON'T USE!).
The BYTE array problem has not been solved. (too bad! I needed
- LSET and RSET can screw up things.
- several overflow tests not OK.
- RETURN a$ does not work properly.
Fortunately, a few guys from the Stichting ST have created a
patch program to solve all these problems. Great! The GfA BASIC
program can be found in the PROGRAMS folder on this disk. Just
select the GFABASIC.PRG you want to patch and things will work
There is one more bug however that cannot be patched. DUMP bombs
out when you try to DUMP a non-exsisting array. (Not DIMensioned
If you discover another bug in GfA, I suggest you either tell ST
NEWS about it, or contact the Stichting ST. Their adress:
NL-2300 EC Leiden