| Commit message (Collapse) | Author | Files | Lines |
|
|
|
sculpties) correctly. [Really need to add a ExtraParams field to the sqlite database though, but for now I have combined their data so that we don't lose backward compatibility, know a couple of people have been using the datastore already].
Now have a rough day/night cycle (the movement of the sun needs to be made smoother but for now it is better than we had I think).
Added dalien's patch (issue 294) for saving and loading prims to a xml file (think he will be modifying these to be import/export functions and maybe writing a xml datastore for backups).
Some preliminary work on task inventory (ie object's/prim's inventory).
Added place holder data for AvatarProperties (ie a avatar's profile). Should we store this sort of data on the user server or have another server for it (a normal webserver should work).
Added a few more method to IClientAPI.
Sure there is something I'm forgeting.
|
|
with Rotation in entitybase (couldn't override as they are different types (LL vs Axiom) and didn't want to add new).
When you take prims into inventory (or delete them), they should now be removed from the prim datastore, so they no longer reappear in-world when you restart opensim.
|
|
* This meant introducing AbsolutePosition on all objects (since SimChat wants that)
|
|
standalone mode and using sqlite).
In standalone mode, if you have account authenticate turned on (setting in opensim.ini) then when you create a new account, a set of inventory is created for that account and stored in database (currently only a set of empty folders). Then during login the database is search for that set and sent to the client in the login response.
More functions will be added soon, like creating new folders (and a bit later items) from the client inventory window.
|
|
the root prim of the child group will actually get linked, working on linking the rest now).
Multiple prim groups are now stored in the sqlite database and are reloaded correctly.
|
|
that we can tell what prims belong to the same SceneObjectGroup. If sdague has a different method in mind when he gets back then he can change it then.
|
|
Temporary have had to rename the OpenSim.DataStore.MonoSqlite project to OpenSim.DataStore.MonoSqlite1, as I'm not sure what was done to stop the old project name being included in the VS2005 solution.
Also some config changes:
OpenSim now has a INI (OpenSim.ini) file that it will read some config settings from (if the ini file exists).
Added Mono.Data.SqliteClient.dll so that we can use the same code for sqlite on Windows and mono/linux. (from what I can tell Mono class libraries have a MIT license so there should be no problems with us including this dll).
So now to get the basic prim storage working , you need to first create the sqlite database file from the sqlite3-prims.sql in share directory. Then in the OpenSim.ini file, change the storage_plugin so it points to OpenSim.DataStore.MonoSqlite1.dll (storage_plugin = OpenSim.DataStore.MonoSqlite1.dll). Then in your region.xml files change the DataStore value so it is the name of your database file (at the moment you need a different sqlite3 database file for each region).
|
|
|
|
change in OpenSimMain), then basic ( with a few limits at moment) prim database backup will work.
|
|
sqlite *but* isn't being added back to the scene on load because
some information (like rootpart) isn't currently exposed enough
to save/restore, and I don't want to change the SceneObjectGroup
definition without MW around to discuss.
A couple of minor changes on the object interface for SceneObjectGroup
and tweaks to this class, and we have persistant prims again.
|
|
must run away on vacation. See you all in a week and change.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
now the fun begins of connecting all the dots
|
|
|
|
a Dictionary, and programatically building up sql commands for our dataset.
Prims done, prim shapes next
|
|
stupid and nuke all this work.
|
|
|
|
|
|
|
|
|
|
|
|
current objects to disk, and does updates on those objects usefully.
It is now functional enough to mean that moving to new objects is
just lots of plug and chug.
|
|
|
|
Not working yet.
|
|
this up to the sim in my env now to make sure we're getting data flow
in a reasonable way at this point.
|
|
main prebuild yet. Going to attempt to get some tracing on current
SceneObjects first
|
|
to do a bit more live hacking on data flow
|
|
|