Follow-up Comment #8, bug #15560 (project wesnoth):
I agree with the fact that advance=true shouldn't have occur in a W context.
(And revision 39404 is just horrible: purposely causing the clients to behave
differently was bound to cause an OOS.) But I don't agree with either option 1
or 2.
More precisely, option 1 is a strong regression with respect to the
expressiveness of WML. Option 2 seems to me that it will introduce more bugs
than it will solve: having two streams of data, replay on the one hand,
synchronization on the other hand, will just cause duplications and
complications. Especially since the replay needs to contain synchronization
data anyway (otherwise it wouldn't be usable as a replay afterward).
My opinion is that we should just get rid of W. Whenever the engine triggers
a situation where information is missing (that is, not the current side yet
there is no replay data), it should assert false so that we can fix it.
As for prestart/start events, they should just be assumed to be at the start
of turn 1, or something akin to it, in order to decide who is responsible for
the decisions. In fact, I think this is already the case and it works fine.
_______________________________________________________
Reply to this item at:
<http://gna.org/bugs/?15560>
_______________________________________________
Message posté via/par Gna!
http://gna.org/
_______________________________________________
Wesnoth-bugs mailing list
[email protected]
https://mail.gna.org/listinfo/wesnoth-bugs
