Re: [hatari-devel] Fw: Zombi/Hatari

[ Thread Index | Date Index | More lists.tuxfamily.org/hatari-devel Archives ]


On 10/05/2012 23:18, Thomas Huth wrote:
Am Thu, 10 May 2012 22:49:40 +0200
schrieb Nicolas Pomarède<npomarede@xxxxxxxxxxxx>:


no I don't remember and didn't find any note about it in my files.
Do you mean it used to work at one point in previous Hatari version ?

I tried with tos 1.04 in stf mode and didn't get any freeze ; is
there a reproducable way to get the freeze ? (by clicking on specifc
places ?)

Actually it does not really "freeze", you just always die after moving
around for 10 to 20 rooms (for example, I do not manage to go down the
stairs and to go up completely again without dying).

Anyway, I think I've got a clue what might be wrong with the game: It
initially sets a new date/time stamp with the IKBD clock-set command -
which we do not support in Hatari yet. So when it later reads back the
time from the IKBD, it gets a time stamp that is far away from the time
expected by the program - so it thinks too much time has passed and you
die.

I am just trying to implement a fix for this ... but that might require
a new variable in the IKBD code that has to be saved in the memory
snapshots, too. So this will break the compatibility of the snapshots
with version 1.6.0 and 1.6.1 ... Nicolas, would that be ok for 1.6.2 or
do you prefer to keep 1.6.2 compatible with the two previous versions?

  Thomas

If we can't avoid extending the ikbd memory state, then it's fine with me if the snapshot changes. It's true it's never nice for the user to not be able to restore a previous snapshot but if it corrects emulation there's nothing we can do (having some kind of migration between snapshot version would be to complicated to handle and not worth the effort).

Also, I have some things I intend to work on after 1.6.2 and they will certainly change the snapshot too, so breaking it a little sooner ... :)

Nicolas






Mail converted by MHonArc 2.6.19+ http://listengine.tuxfamily.org/