Re: RESET vector in MiNT

[tlang@limes.de (Torsten Lang)]

Hi Petr,
> 
> MS>Why do we need the reset vector anyway?
> MS>
> MS>TOS *should* IMHO set everything that is needed after reset. OK, some older
> MS>versions don't reset cookie jar, _shell_p etc., but if a fix is needed for
> MS>these, it should only be installed if that specific TOS version is detected.
> 
> Exactly my thoughts. So could we remove the reset vector stuff since the
> next (beta) release and watch people if it makes any problems?
> 
Well, in general I think it's not up to MiNT to patch around in the OS. There
are enough bugs in different TOS versions which - to my mind - should be
corrected by a TSR. So you have the possibility to start that TSR where needed
(normally very early in the AUTO folder). On the Amiga Commodore also used a
centralized way to install essential patches to the OS: There's a "SetPatch"
program which analyses the OS and installs patches needed for the specific
OS version. This avoids the situation that "patch a has to be installed before
patch b and c, but patch c wants to be installed before a or otherwise disables
a" and so on. Another very concrete problem for example on my system is that
I can't use CBHD as driver since - only when autobooted - it patches around in
the system and disables some of my partitions :-(
So it could be a good idea to take these patches out of MiNT and put them into
a TSR - perhaps configurable via the HSMODEM setter.