For the last few weeks several of the Lindens that are good news sources have had little to say. They are working on various fixes in infrastructure, Experience Tools, Pathfinding, and other stuff. While some of the things are interesting most is rather tedious and boring. The result has been my consolidating news for various subjects into a single article.
This week and to some extent last week I have been breaking news back into separate articles. I’m doing the same this week. The left over interesting stuff is in this article.
SVC-4444 – Objects become fully permissive when rezzed under certain conditions. Baker Linden and Andrew have been looking at this item and working on fixes. It is a complex permissions issue that touches several inventory and transaction services. The work seems to have spun off into fixing other parts of the SL system so item 4444 can be fixed. Seems this is going to take some time.
Andrew Linden even said they gave up on 4444 until they get the associated problems fixed.
VWR-28921 – “Permissions” button in contents tab incorrectly uses slammed perms. This one of the ‘associated to 4444’ problem they have fixed.
VWR-27853 – Unable to change parcel restrictions for a scripts-disabled parcel in a damage-enabled region. Simon says this one is fixed.
VWR-28604 – Object by multiple creators shows creator as “(unknown)” in Inventory. Simon Linden says, “28604 isn’t fully fixed, but the ‘unknown’ is now ‘unknown / multiple’. It’s just a string change … actually extracting data or getting better info isn’t easy at that stage.”
SVC-4968 – Group won’t load – too many members. This is a big issue for a number of groups in Second Life. Andrew and Baker Linden are looking at how this can handled and fixed.
Some where around 10k members it becomes very difficult to manage groups.
SCR-79 – llMatchGroup() – Checks if an object or agent is active in a specified group. This came up in a recent UG meeting and Kelly Linden has been looking at it.
Many of the scripts in Second Life give permissions based on which group one belongs to. Consider the various shops that have a free gifts for those in the store’s group.
One of the problems is having to make the group active, so a script can check whether you belong to the group. There is a privacy issue involved. Some of us hide some group memberships. Whether one is embarrassed by or just not wanting to be hassled about their membership in Vestigial Virgins, Kink Kings or Luscious Lesbians, we are free to choose what information we share. Abuse of some scripting functions can take that freedom and privacy away.
But, it would be easier for scripters and residents if the script could check all your ‘public’ groups at once. This would save you having to change which group is active. However, it would give the script access to more information about you.
Kelly has found a way that can easily check groups you have listed as public. He is thinking that llGetAgentGroups(key agent_id) would be a solution. It would return a list of the group you are in that are NOT hidden. It won’t tell you which is active.
Kelly considered fixing llGetObjectDetails to work for agents IF their active group was showing in their profile. But, it is not currently possible to get the active group ID of another person.
Andrew was thinking about whether privacy would be maintained if an object that was owned by a group could check to see if a resident belonged to that group whether the user had the group hidden or not.
Kelly says it would be possible to do llAgentInGroup(key agent, key group) that returns true if the target person were in that group and the group was listed in their profile.
UPDATE: 2012-06-19 – There is an update on SCR-79 and group checks. It looks like it is not going to happen. As Kelly Linden researched the project he ran into data retrieval problems that make the project too complex for a near term solution.
You may not know that much of the viewer’s performance, as measured by FPS (Frames per Second), is dependent on the Render Pipeline. The pipeline is the code that takes all the pieces and parts of the digital world and draws the image we see on the screen.
To see how many components and steps are in the process turn on Fast Timers (Ctrl-Shift-9 – it’s a on/off toggle).
The rumor is that Runitia Linden is working on the pipeline to improve it. Some are concerned that work may complicate things for the Mesh Deformer. We have no evidence that is so.
If you ever visited Blue Mars, you know that OpenGL, the graphics system used by SL and BM, is capable of more than what we have seen in Second Life. With my old 8800 I was seeing 20-30 FPS in SL at high settings and 90+ in Blue Mars with way better imagery. So, I am really hoping this rumor is true.
The infrastructure package with the Phase I Multi-threaded Region Crossing stuff made it to the main channel. As of Friday it was working well.
The Blue Steel and Magnum release channels continue to run the packages from the previous week. Oskar is expecting Magnum to advance next week.
Blue Steel has the ‘Pre-Baked Avatar Appearance’ package. I’ll have to ask Nyx Linden how the research on the Blurry Avatar (Bake Fail) problem is going.
Magnum is running the package with a fix for: SVC-7902: Problem of not being able to rez on my land continues.
Le Tigre according to Oskar was ‘uncooperative.’ The package released there is about the infrastructure change of the release process, how regions are updated and packages rolled out. Infrastructure changes don’t change how the region works. They change how the server handle the region simulator, if that makes sense to you. An analogy might be using Photoshop to make images you use in a Word Document. Changes to the Photoshop program (infrastructure) would not change the Word (the simulator) program that holds the image.
Because of the problems they stopped Wednesday’s roll to Le Tigre. They did some changes and tried roll again on Thursday. They noticed that some regions were taking a really long time to connect and show neighboring regions. After an hour they connected and were fine. The result is the roll out was taking a LONG time and Oskar thought it might still be in progress during the meeting. Since roll outs start early in the Pacific Time and it was late afternoon, that is an unusually long time for a roll out.
Oskar says a number of server maintenance fixes are in the QA pipeline.
If you have changed your password and seen problems on ADITI (SVC-7727), then you know what a pain it is. There is a fix working through QA. It is ready for testing in ADITI and with any luck we will see it get rolled to a region in ADITI in the next couple of weeks.