Deploys for the week of 2011-05-16

[Tika Oberueng]

Indeed, it appears that LeTigre has re-rolled with the same version number as before, 11.05.11.229644 instead of the expected 11.05.17.230139... Version number update issue, or did the old code go back out? With Magnum's issues today, plus who knows what else, it's been a rough day. Hopefully they get better for everyone!

[Cincia Singh]

I'm confused (nothing new but it causes me to ask questions). If the main channel AND the RC channels have all been rolled and are now supposedly running 11.05.19.xxxxxxx, why is Quordlepleen region on the mainland still running 11.05.05.228940 (apparently a Blue Steel release from May 5th)?

[Maestro Linden]

Hi Innula, the fix for SCR-66 is currently in Second Life RC LeTigre 11.05.11.229644.  This fix will probably be on all regions in 2 weeks, barring any unexpected changes or bugs.

[Innula Zenovka]

Thanks for the prompt reply, Maestro.   That's very good news and very helpful to know, too.

[MB Robonaught]

Is it possible to do a test run on a single Magnum RC channel before deploying to all? If so contact me I'll be glad to help gather information for you to decide to go ahead or not

[Void Singer]

it's not only possible, it's how things are run.... Magnum, BlueSteel, and LeTigre all get series of intial fixes/changes, and after a week if they evaluated to see if they are behaving as expected, if so, the simplest or most well tested of the lot get promoted to the main channel, the others get tweaks and/or extended testing. there are also the "Snack" servers now wich are being used to test very simple additions/changes for quick inclusion.

if you mean a single region... probably not, although I don't suppose it hurts to ask... sometimes single regions are duplicated to beta grid for extended pre testing.

[MB Robonaught]

If there is anything we can do let me know. The sim I own is a car racing sim and if anyone can tax a server we do. Our races are often decided by thousandths of a sec and not to mention we hit the end of the sim at high speeds. So it'd be a great place where gathering data on how script times and memory from various sources can point out areas where efficiency can be greatly approved for all aspects in SL. We've already done testing, though limited with the tools we have,on things like how certain scripted cars gain an advantage by being poorly written an thus gaining scripting priority. So if anything we can offer is of interest please contact me direct

[Void Singer]

I can suggest testing region crossing behavior for those vehicles on the Snack Sandbox regions (search from the map) as there are supposed to be some improvements for MONO scripts and region crossing there, among other things...

IMHO LSO scripts are still better suited for cross region performance, but the gap is narrowing.

[MB Robonaught]

The cars we use rarely cross sim boundaries as they are designed for straight line racing the length of a Sim. The track itself is as long as a Sim is. The testing I suggest is to gather information on how fast the server can keep up to this type of racing( drag racing). For example Reaction time which is a critical part is the amount of time where the script turns a light on to when the sever receives a command to move forward at a very high rate of speed and acceleration. In the fastest class a car can transverse an entire sim in around 0.94 seconds at a SL speed approaching 480mph. While this is going on the scripts need to keep track of which starting prim is hit first along with which prim is hit first at the end within .0000001 of a second response. So as you can see any server lag spells doom for a race. My suggestion is if we watch what makes and breaks a race and make it more efficient think how smoothly normal everyday operations would become from working on the extreme side

[Void Singer]

there are other performance changes on those servers, and region crossing was only suggested as a major thing to test.... I'd test performance by comparison where possible to make sure and be alerted to any early issues