Le 04/03/2013 11:18, Eero Tamminen a écrit :
Hi, On maanantai 04 maaliskuu 2013, Helmut Karlowski wrote:I do all of them, and they don't provide a valid fix that avoids violating supervisor mode.1. Who says it's a violation at all? Compendium?Yes, Compedium says: You should never call the AES in supervisor mode. In addition, supervisor mode should be entered and left in the same stack context (same 'C' function) or stack corruption can result. (This is from the description for Super().)
The compendium is old, sorry this is not because the TOS is not able to do that, that a new AES should not be able to do it!
TOS, NAES crash Magic, MyAES workXaAES near work as I can see (calling objc_draw() in a userdef), it doesn't crash.
This is only a problem of management of the stack, not so difficult to do.If we look at documentation for gemdos process message you should not call gemdos system, this is stupid, Magic do it perfectly, it's a cool way for example weak up a process and clean easy to implement AES messages management rather to use infinit loop, MyAES was at first design like this, this was very cool and easy and it work fine under Magic and crash under Mint :-(
System should be modern, perhaps we are not able to do it but old restriction should not be a feature when possible we should fix it.
Olivier
- Eero