|
|
|
[
Permlink
| « Hide
]
WarKirby Magojiro added a comment - 15/Dec/07 01:12 PM
annoying.
I've always tended to assign start parameters somewhere myself anyway, but yes this is definitely wrong.
What if the script is loaded with llRemoteLoadScriptPin()? does the start parameter come back there? I noticed that the on_rez isn't triggered for that but llRemoteLoadScriptPin does have a start_param so I wonder where that gets loaded?
Just tested with the Beta Grid Refresh done on Jan 11, 2008 (Simulator v1.18.6.77074). The bug is still present and causes llGetStartParameter() to always return zero.
Many of my scripts - including the Rez-Foo - use llGetStartParameter in various points in the script. I'd say this issue is rather critical because this function is pretty wide spread in many scripts out there.
Now that Havoc 4 is being tested on the main grid the problem shows far more often then before. Apparently some changes to the order of operations for rezing objects caused us to essentially lose the setting of the start parameter. Oops.
This will be fixed with the next havok4 deploy. One thing to note is that this value is set when the object is rezed, there is no way to "fix" objects that are already in world. If you rez an object this way in the current havok4 it will have a start parameter of 0, even after the next update. However if the same object does the same rez after the next update it will work as expected. Given the general uses of this I don't think this is a huge deal, but wanted to let people know. This is a very big deal for us on ACVA, the functionality of BabelSwarm relies on the start parameter working. What's the ETA on the next update?
Sadly, not fixed in todays mono beta server refresh (according to my testing). Will it need the next client ?
This has totally broken a family of my objects. I am hoping for a fix soon, or should I start the task of reprogramming to communicate another way, without using the startparameter ? It is very important to note that this was a havok4 bug, not a mono bug. Havok4 and mono are sharing a grid, but not the same code. Mono beta updates will not effect the havok4 beta updates and visa versa. This fix for havok4 will be in the next havok4 update. If you found this bug in mono, please file a new jira with the appropriate Affects Version(s).
Is there an estimate on when the next Havok4 update will be?
I have done some further testing on this bug and found that the llGetStartParameter is still loosing it's value when the script is reset. I am sure the value was not lost when a script reset under havok1. ...and if it did, it shouldnt! A lot of legacy items issue a reset in the on_rez event to get around memory leaks and other old bugs like llGetOwner not updating the owner. Darling. This bug was fixed months ago and is working fine. llGetStartParameter should be cleared on script reset, it's most likely stored in script memory like any other variable. Everything in script memory is cleared on reset. If anything is left then it's a bug, since the whole point of a script reset is to clear everything out and start fresh.
From the LSL Wiki on llGetStartParameter: "If the script is reset (using llResetScript or other means), this value is set to 0". |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||