Re: wolf 3d-- bug? -- the culprit?

[Rubywand <rubywand@swbell.net>]

Nathan Mates writes ...
>
> .... 
>    Given that I found crash-causing code in the ROM code of my ROM 3
> box, I *highly* doubt it's Hierarchic. I have a fix on my ROM 01
> system (similar code is in RAM and easier to patch). And, I do have
> the Hierarchic init v1.6.1 on my ROM 01 system. Wolfy works great with
> Hierarchic and my patch.
>

     Neat! Is the patch available on your web site? If so, what is the
file to download?

 
>    The problem is due to an OS bug (which can be worked around at the
> app level), in which critical tools can be incompletely dumped from
> RAM under the wrong circumstances. The only reason why it has probably
> remained undetected so far is that the conditions to set it off are
> quite rare-- an app which has called _SetTSPtr to manually patch out
> a toolset.
> 
>    Adding or removing inits can mask or expose memory trashing
> bugs. You may shift things around to a configuration where critical
> sections of ram get zapped, or relatively unused parts are trashed. I
> can say definitively that crashes on my system were caused by an OS
> bug, not any init, not any addon. And that's proof by code, not proof
> by absence of crashes.
>

     This is pretty interesting. Are we talking about situations where
an application does more than one _SetTSPtr for the same toolset?

     Could you explain the bug in more detail and explain the fix?


Rubywand