Sandcastle fix?

<< < (4/4)

dizzy:
I've looked it over, and it seems more like a problem with robustness than a multi-tile issue. Say, for example, you have a lot reset while there are sand castles on your lot. That would cause them to re-init IIRC and would thus leave your sand castles in a hung state.

Code:

# Group = 0x7FA5A9EA, Instance = 0x1001
# Title = Function - Main
#
# Format = 0x8009, Params = 0, Locals = 0
# Tree type = 0x0, Header flag = 0x0, Tree version = 0xFFFF8009 (-32759), Cache flags = 0

     0: Idle(450 (0x1C2)); true: 1, false: error
     1: My Attribute 0x3 > 0; true: 3, false: 0

You see, if the Build interaction isn't forced by the Destination, the Kill Me Countdown attribute will never be > 0, and thus you will have invisible sand castles till you bulldoze the lot or use some object deleter.

Now that I stop and think about it, I think you might have stuck sand castles merely from canceling the build interaction as well.

J. M. Pescado:
Quote from: dizzy on 2007 September 19, 10:52:39

I've looked it over, and it seems more like a problem with robustness than a multi-tile issue. Say, for example, you have a lot reset while there are sand castles on your lot. That would cause them to re-init IIRC and would thus leave your sand castles in a hung state.
The Sand Castle init does not reset Attribute 3, so the Sand Castle Self-Destruct Timer would be retained.


Quote from: dizzy on 2007 September 19, 10:52:39

You see, if the Build interaction isn't forced by the Destination, the Kill Me Countdown attribute will never be > 0, and thus you will have invisible sand castles till you bulldoze the lot or use some object deleter.

Now that I stop and think about it, I think you might have stuck sand castles merely from canceling the build interaction as well.
I think this is actually a real possibility, one which the Jeffistani fix does nothing to address. Experimenting with this now.

UPDATE: Found the real bug. It's not where Jeffy found it, and Jeffy's fix will not stop it. Fix out Real Soon Now.

What ACTUALLY happens is not that the sand castles are not being flagged for destruction properly after being kicked, stomped, or abandoned, but what happens if the action is cancelled while the sim is still routing to the sand castle site: While a started construction process is properly flagged for termination, if the sim fails to reach the sand castle due to routing failure or the action is cancelled, the abandoned sand castle (which has never been started and therefore is invisible and cannot be kicked or stomped) is abandoned forever. This now properly flags all sandcastles for termination and destroys them immediately if they are still invisible.

Someone should inform Jeffy that his fix has missed the real issue, since New Jeffistan seems to lack a feedback area and I would probably be banned on sight there anyway, nor do I feel like jumping through the ridiculous registration hoops I've heard they have there.

Navigation

[0] Message Index

[*] Previous page