There are a few interesting things that came up in Andrew’s late Friday afternoon meeting.
Andrew has been building a script to turn on Encroachment Return across the grid. The feature made it through QA and has been rolled out for some time. But, enabling it in thousands of servers is not small chore.
The script was run Thursday morning. So, you should now be able to return encroaching objects.
Andrew says there will be no SL Blog article announcing the change. They will be updating the Knowledge Base information, sometime soon.
There is considerable content on the mainland that is Mole built. If the properties on the content are correctly set, and they usually are, encroaching Mole objects cannot be returned.
Feature explanation: Objects are considered on your property when the center of the object or the center of the root prim is on your property. Those objects you have been able to return. But, if the center of the object was off your land and only a part of it encroached, you could not return it. The ‘Encroachment Return’ feature changes that. Making the change for prims was complicated enough. Including mesh objects made it more complex. Now they think the Lindens think they have it working well and have enabled it.
‘Mole’ is the name for the folks that build for the Linden Dept. of Public Works (LDPW). They build roads and stuff on the mainland and other continents. You can learn more about them in the Wiki at the above link.
Convex Hull Shape
Last week someone asked about what shape the physics engine used when a cube was changed to convex hull physics shape. Andrew had to find a Linden that knew. He came back with the answer this week, which is: yes. A prim cube set to convex hull uses Havok’s simple cube shape.
The change also pushes the cube into the new Land Impact cost accounting. See: Reducing Prim Count
Keyframe Objects Stopping
Those using llSetKeyframedMotion() have found that region restarts stop the behavior. More reports of the problem are trickling in. See: SVC-7556 – Objects with llSetKeyframedMotion() stop running after region restarts.
Andrew said, “Falcon wasn’t aware of that bug. He hopes to get to fixing it soon (pathfinding is almost over). It is a bug. We’ve had some reports.”
If you have experienced the problem and can reproduce it, please add your information to the JIRA.
There are rumors of accounts being quickly closed after becoming delinquent. Someone had asked Andrew about it and he did some checking. The Wiki has an article titled Linden Lab Official:Delinquency policy.
Accounts are blocked after 30 days. Accounts are recoverable for some time after that. Andrew said, “The account names may never go away, but after very long times the inventories might be cleared.”
Andrew brought up an inconsistency in Linden programming standards and asked about it.
“I had a question about script errors… Sometimes the script will chat to the debug channel, other times it sends the message just to the owner’s viewer (if the owner is present). It depends on which error the script hit. But, there is no clear policy on the server implementation. Sometimes we do one thing, sometimes the other. I was wondering if we should/could standardize on something?”
Seems most at the meeting did not have a strong preference. The only consensus is that error messages should start of with: “Doh!”
For now it looks like the error reporting messages from objects may change to only appear on the owner’s viewer.
Some like the idea of an Error Event, so they could write error handlers. Falcon mentioned it. The event and handling processes is common in other programming languages. But, he thinks adding an on_error event could be complicated.
LSL Syntax and Tooltips.
Andrew has been working to get an easily updatable process in place for the LSL. He said, “Not done yet, but I’m going to try to update some project viewer with updated tooltips, and keep working toward a server-side cap that will provide the latest syntax.”
Simon Linden reports that the World Map is working better. You can zoom out now without some of the region tiles disappearing.
Falcon repeated the news from an earlier meeting that Pathfinding tools will be out early in Week 12.
He asked, “Please come to Aditi next week (regions to be noted on the Pathfinding alpha web site) to test your physics objects. A LOT of physics changes have been made for this project. Some things will probably break. We can probably fix most of the ones that do. There will probably be some we can’t.
Volume detect and collide both work for Pathfinding characters; character vs character, character vs object, and character vs volume detect, which means you can detect a character moving through a volume detect object.
Falcon also said, “There will be an Agni beta for Pathfinding in the coming weeks. The beta will only be open to region owners (though once Pathfinding is released grid wide, it will not require you to own a region to use it). If you are interested, please e-mail firstname.lastname@example.org.”
Falcon also told us, “We’re doing an experiment with llCastRay, reducing the resource limits (that is, making it easier to cast more rays). This should solve the resource pool problem, but we’ll be monitoring for any griefer-related or performance issues. We’re gonna go with “raising” them for now. If all goes well, we’ll remove the code entirely. [For now] I’ve just disabled the code but you need to write your scripts as if it were there because it might come back. I hope to revisit the keyframing bugs after Pathfinding ships.”
Summing It Up
The last two weeks have not seen much making it to the main grid. However, some features in open beta are nearing completion and will soon roll to the main grid. Others are going into closed beta: Experience Tools.
I think we are in for an interesting year.