RE: [MiNT] Gemjing and memory protection: possible?

[Odd Skancke <oskancke@online.no>]

On Wed, 1 Nov 2000, Skarstein, Jo Even wrote:

> > 
> > When I boot without jclktool I get 
> > pid 126 (taskbar) MEMORY VIOLATION type=privat ........
> > 
> I was going to suggest that you shouldn't use 3.00a, as Ozk can't even start
> it without it reporting a memory violation...
> 
> > Without jclktool and taskbar I get this:
> > 
> ....but then I read this. (Hey Ozk, perhaps the crashes aren't my fault after
> all? ;-)

 Hmm.. explain why the previous version of TaskBar works, then? Besides, I
use a fixed up version of GemJing, compiled with gcc 2.95.2 using mintlib
0.55.3. And I did check the AV handling in GemJing, which is correct.

> 
> > pid 118 (thing) MEMORY VIOLATION type=privat ........ 
> > pid 105 (aessys) MEMORY VIOLATION type=free RW=r AA=1C983C PC=121A4RE
> > BP=1212000
> > (aessys) Operating system killed
> > FATAL ERROR
> > 
> I really don't know what's happening, but it could be that GEMjing
> broadcasts an AV-message or something (just a guess) and that this kills the
> responding apps. It's not likely though (e.g. why should aessys suffer from
> this?), but try running Jay's GEM message trap program
> (http://jaysoft.atari.org/) and see if something like this happens.

 GemJing have never killed anyhting on my system, so I would rule that app
out of this problem. However, if GSXB is used by GemJing to install
end-of-frame interrupts (which the old version did not do, but 'my'
version does) one have to use the xdd versions of the ES1371 driver,
otherwize serious shit happends when the eof interrupt triggers :)

 I really doubt the above is related to AV-messages at all, since Thing
have always catched those on my system. Something more serious is wrong
here.

 Martin, does this only happend with GemJing? 
 
Regards,

 Odd Skancke - ozk@atari.org - http://assemsoft.atari.org