-Added a new method in Resource: reset_state , used for reloading the same resource from disk
-Added a new cache mode "replace" in ResourceLoader, which reuses existing loaded sub-resources but resets their data from disk (or replaces them if they chaged type)
-Because the correct sub-resource paths are always loaded now, this fixes bugs with subresource folding or subresource ordering when saving.
-For inspector refresh, the inspector now detects if a property change by polling a few times per second and then does update the control if so. This process is very cheap.
-For property list refresh, a new signal (property_list_changed) was added to Object. _change_notify() is replaced by notify_property_list_changed()
-Changed all objects using the old method to the signal, or just deleted the calls to _change_notify(<property>) since they are unnecesary now.
-Rendering server now uses a split RID allocate/initialize internally, this allows generating RIDs immediately but initialization to happen later on the proper thread (as rendering APIs generally requiere to call on the right thread).
-RenderingServerWrapMT is no more, multithreading is done in RenderingServerDefault.
-Some functions like texture or mesh creation, when renderer supports it, can register and return immediately (so no waiting for server API to flush, and saving staging and command buffer memory).
-3D physics server changed to be made multithread friendly.
-Added PhysicsServer3DWrapMT to use 3D physics server from multiple threads.
-Disablet Bullet (too much effort to make multithread friendly, this needs to be fixed eventually).
Setting each point's position was missing for 3D. Now enabling collision
render debug will display contact points for 3D physics, the same way it
does for 2D physics.
Note: Multimesh rendering seems not to work in this scenario on master,
but it's working fine on 3.2.
This changes the way 2D & 3D physics picking behaves in relation to pause:
- When pause is set, every collision object that is hovered or captured (3D only) is released from that condition, getting the relevant mouse-exit callback., unless its pause mode makes it immune from pause.
- During the pause. picking only considers collision objects immune from pause, sending input events and enter/exit callbacks to them as expected.
- When pause is left, nothing happens. This is a big difference with the classic behavior, which at this point would process all the input events that have been queued against the current state of the 2D/3D world (in other words, checking them against the current position of the objects instead of those at the time of the events).
-Always use temporal reproject, it just loos way better than any other filter.
-By always using termporal reproject, the shadowmap reduction can be done away with, massively improving performance.
-Disadvantage of temporal reproject is update latency so..
-Made sure a gaussian filter runs in XY after fog, this allows to keep stability and lower latency.
-Added more finegrained control in RenderingDevice API
-Optimized barriers (use less ones for thee same)
-General optimizations
-Shadows render all together unbarriered
-GI can render together with shadows.
-SDFGI can render together with depth-preoass.
-General fixes
-Added GPU detection
They are bound as both regular and virtual methods which makes ClassDB
report the methods twice when querying the API. The non-virtual binding
is removed since both methods only seem to be used as virtual.
This commit adds a view-dependant fade to the 3D viewport grid. It fades out
at steep view angles to hide the solid regions that appear far from the camera.
I also included a fade to hide the grid borders.
I added some improvements to the dynamic grid when the camera is in orthogonal mode.
It properly handles zoom now, and the grid center is now set to the intersection point
between the grid plane and the camera forward ray, keeping the grid
always visible.
- Based on C++11's `thread` and `thread_local`
- No more need to allocate-deallocate or check for null
- No pointer anymore, just a member variable
- Platform-specific implementations no longer needed (except for the few cases of non-portable functions)
- Simpler for `NO_THREADS`
- Thread ids are now the same across platforms (main is 1; others follow)
reduz
Rémi Verschelde
MarvinFF
Kongfa Waroros
PouleyKetchoupp
Pedro J. Estébanez
Rafał Mikrut
Kanabenki
kobewi
Oliver Dick
Rémi Verschelde
George Marques
muiroc
clayjohn
jfons
nemerle
bruvzg
Aaron Franke