occasionally, but are fixed on restart (issue 1919).
This patch introduces the following changes:
1. when a user teleports out of Region A, remove that user's profile
from the Region A user profile cache
2. when a user crosses between regions out of Region A, remove that
user's profile from the Region A user profile cache
3. the user profile cache's session ID member can now be set (written),
and is updated each time a connection with a new avatar is established (ie: a
new avatar enters the region)
4. when a region server looks up a user profile and a cache miss
occurs, fetch the user profile from the user server first instead of
immediately returning null
ScenePresence. Pace prim delivery to about 200 updates/s max. Break
a long-held lock that caused the notorious MapBlockQuery XMLRPC timeout
on agent login. Eliminate 60 second timeout at "Waiting for region handshake".
Fix region crossing/teleport response. Fix appearance in primmy regions.
llRezObject changed to llRezAtRoot. llRezObject implemented
to call llRezAtRoot until I can figure out how to calculate
a bounding box. Partial implementation of llPushObject
(no angular impulse applied)
patch attached replaces the tree walk algorithm used to build the
folder hierarchy with a single database query. That is, we replace
1 database query per folder with 1 query for the root folder's
properties and 1 query to retrieve the entire collection of folders for a user.
In case you run the server with mono --debug OpenSim.exe
line numbers are reported as <filename>:<linenumber>, so no
"at line" is found. That led to an exception, which is caught
since r5766. The attached patch fixes this for mono;
Bringing in an old private patch, increasing region registration timeout
to 90 seconds. That allows regions witj > 10000 prims to register
successfully. Running them.... well, YMMV
* If there's a Socket.AlreadyInProgress, just silently ignore this one
* Tried to refactor the Reset and BeginRecieve logic into something a little more readable, little less duplicated
Increase GetUserProfile timeout from 6 to 30 seconds. Because
the worker thread fo the XMLRPC data pump sits on a lock
while the system starts, regions with upwards of 6000 prims
throw and exception when that thread gets blocked for more
than 6 seconds, due to this timeout, and won't start.
When a region is re-registered (e.g. because of region restart),
it is removed from m_regions. If I understood that correctly,
the old entry should be replaced, instead.
This fixes a sitting problem which popped up on loading regions from
archive and turned out to be caused by the fact that the archiving
mechanisms all preserve the sit-target information that is set in the
object at the time the image is captured. This caused the new region
to pick a sit-target prim that did not correspond to the prim on which
the script was running, so the script's changed event is driven with
an invalid avatar UUID.
I have modified the ArchiveReadRequest class so that any sit information
that survives the archiving process is deleted before the object is
instantiated. This change has just been checked in.
Mike Mazur