Server roll outs have not been going that well this year. Tuesday a new roll was planed and started. It went well. Then Wednesday came and release channels were rolling out. The roll to Magnum went bad.
Oskar says, ‘This is the code that’s been in a few RC’s this year. It is looking stable.’ This is the code that rolled to Blue Steel this week.
New LSL function: integer llSetRegionPos(vector position) – The object with the script will move the root prim position to the given location. The position is any position within the region. If the position is below ground, it will be set to the ground level at that X,Y spot. The function has no delay or throttle.
- Returns 1 if the object is successfully placed within 0.1 m of position.
- Returns 0 and does not move the object if position is more than 10m off region or above 4096m.
- Returns 0 and does not move the object if the object is dynamic (has physics enabled).
- Returns 0 and does not move the object if the object can not move to position due to object entry rules, prim limits, bans, etc.
“frame_number” option added to llGetEnv() – Returns an integer that represents the current ‘frame’ of the simulator. The function is generally only useful for specific debugging cases.
- SVC-7466 – A note card holds more data than a script can read
- SVC-7520 – Keyframe motion doesn’t move towards the correct position when specified time is not an exact multiple of 1/45 seconds
- SVC-7485 – llSetKeyframedMotion cannot stop animation if none is running … sounds less important than it is …
- SVC-7493 – Weird mesh land impact issue
- Fixed several simulator crash bugs and potential memory leaks.
- Fixed a note card crashing bug.
Same as Blue Steel.
This server maintenance package failed. Read SVC-7604, which is next, to learn more.
SVC-7604 – llList2List() causes script crashes in RC Magnum 12.01.14.247742
The new code began crashing regions. The roll was rolled back. Now the code is running in ADITI. The Lindens need people to test real products that use the llList2List() function. Once testing is complete on ADITI it can move back to a release channel.
This ‘fail’ is due to the limited testing that can be done with test scripts. Real products are much more varied and users have thought up unique ways to use the function. The new version of it needs to be put through the paces. That can either be in ADITI or the AGNI release channels. People have no choice about helping when it moves to AGNI.
As Kelly Linden said, ‘The only way we won’t have a repeat next week is, if we get real products that were effected tested on aditi. Sandbox Goguen was what I was using, but I can double check.’
So, if you have problems with one of your products, get a copy over to ADITI in the regions; Ritch, Rosedale, Sandbox Goguen, or Sandbox Island Extension and test it.
If you know people that script, tell them about this article and problem. See if they can help out.
The functions being tested are: llList2List, llDeleteSubList, llGetListEntryType, llListInsertList, llListFindList, llListReplaceList, llDeleteSubString, llStringLength, llInsertString, and llSubStringIndex.
Remember. This affects scripts compiled to Mono.
SVC-7608 – Sims are not visible when diagonally opposed.
The Lindens are asking you to add information to the JIRA, if you have seen this happening. I suspect even knowing regions where it happens will help.
Maestro had been tracking a bandwidth issue in the new code . I wasn’t clear if they mean new to the main grid, release channel, or ADITI. I hate it when I miss a meeting.
The next server maintenance package has improvements for region crossings.
Kelly says he is working on bug fixes, crash problems, and the bandwidth thing.
Linden Realm’s Tools
Everyone wants to know when. Oskar says when they make it to testing he’ll let us know. Otherwise he knows as much about it as we do. Hey, it’s his story and he’s sticking to it.