arybicki@pnet02.gryphon.com (Adam Rybicki) (07/18/89)
Hi All,
I am working on a program that uses iff.library v16.1 which was
posted in .binaries.amiga. I have encountered a problem that is either with
iff.library or my hardware. The problem happens also when iff.library calls
are made from the sample program included with the .zoo file. Here is how I
can reproduce the problem:
- I ask my program to load IFF pictures from the disk
a few times (must do it more than 10 times).
- I exit my program--nothing happens yet.
- I type 'avail' at CLI prompt--I crash with Guru 00000003
which is an address error.
I have memwatch running all the time and it doesn't complain.
Metascope debugger stops 'avail' before the guru. Then I see that A0 (and D0
for that matter) contain some bogus value 0x23232323.
Has anyone else encountered this problem? If it is not a problem with
iff.library than what else is likely to cause such an error?
Thanks
----- Adam -----
UUCP: {ames!elroy, <backbone>}!gryphon!pnet02!arybicki
INET: arybicki@pnet02.gryphon.com or arybicki.lax1b@xerox.comarybicki@pnet02.gryphon.com (Adam Rybicki) (07/20/89)
Here is a reply I got in mail to my original plea for help. It basicaly
concludes that iff.library v16.1 has a bug that trashes the free memory list.
------------------------------------------------------------------------
You're pretty right on iff.library, it screws up the free memory list.
This happens from time to time, depending on the particular picture. I had
some Brushes and some ILBM with XModem garbage, and one caused a bug similar
to yours, crashing in Avail, the other caused my 2 1/2 mega fast ram to
fragment incredibly. Putting the particular pics through DPaint solved the
problem.
I traced the bug to OpenIFF(), that is, if you OpenIFF a picture, you're
sold if it causes trouble, as you have no possible way to check IFFError()
before OpenIFF(), and afterwards it's too late. Certainly not the kind of
bug you'd want in a program where other people will input their pictures!
Bottom Line: the iff.library causes this, be careful.
You're welcome to post this reply, I can't get to post currently - emacs
messes this terminal. Actually, I would like you to post this.
Yours, Henrik Clausen, hrc@daimi.dk Fido: 2:234/24.1
-------------------------------------------------------------------------
If we could only find out how to contact the iff.library's author...
UUCP: {ames!elroy, <backbone>}!gryphon!pnet02!arybicki
INET: arybicki@pnet02.gryphon.com or arybicki.lax1b@xerox.com