In rare situations if a light is placed near colinear to a frustum edge, the extra culling plane derived can have an inaccurate normal due to floating point error.
This PR detects colinear triangles, and prevents adding a culling plane in this situation.
Add scene side discrete level of detail.
New node `LOD` for UI, and `LODManager` within `World` for automatically updating child visibilities based on distance from cameras.
* Re-introduces a property for portals to decide whether they are included in room bounds during room conversion.
* Adds a special case for portals that extend into the start room, which may be caused by level design inaccuracies.
Existing shadow caster culling using the BVH takes no account of the camera. This PR adds the highly encapsulated class VisualServerLightCuller which can cut down the casters in the shadow volume to only those which can cast shadows on the camera frustum.
This is used to:
* More accurately defer dirty updates to shadows when the shadow volume does not intersect the camera frustum.
* Tighter cull shadow casters to the view frustum.
Lights dirty state is now automatically managed:
* Continuous (tighter caster culling)
* Static (all casters are rendered)
1) Physics interpolated particles in global mode are specified in global space. In VisualServer they should therefore ignore local transform.
2) Additionally, the expected final_transform should be passed on to children, rather than the identity transform used on the local item.
3) Local bounds in hierarchical culling are fixed for items using identity transform, by calculating their local bound in local space from the global space particles.
The FontDrawer used in TextEdit was previously not being flushed before drawing auto-completion boxes. This was causing rendering artifacts.
This PR also increases the backward compatibility of the MultiRect OFF mode, by forcing a flush after each character.
Adds optional hierarchical culling to the 2D rendering (within VisualServer).
Each canvas item maintains a bound in local space of the item itself and all child / grandchild items. This allows branches to be culled at once when they don't intersect a viewport.
The bound Rect2 was previously incorrect because bone transforms need to be applied to verts in bone space, rather than local space. This was previously resulting in skinned Polygon2Ds being incorrectly culled.
Large groups of similar rects can be processed more efficiently using the MultiRect command. Processing common to the group can be done as a one off, instead of per rect.
Adds the new API to VisualServerCanvas, and uses the new functionality from Font, BitmapFont, DynamicFont and TileMap, via the VisualServerCanvasHelper class.
Later logic using the `pairable_mask` would catch cases preventing pairing callbacks between geometry. However the collision checks were still taking place, wasting performance.
Here we utilize the `tree_mask` feature of BVH to totally eliminate unnecessary collision checks between geometry.
As many open source projects have started doing it, we're removing the
current year from the copyright notice, so that we don't need to bump
it every year.
It seems like only the first year of publication is technically
relevant for copyright notices, and even that seems to be something
that many companies stopped listing altogether (in a version controlled
codebase, the commits are a much better source of date of publication
than a hardcoded copyright statement).
We also now list Godot Engine contributors first as we're collectively
the current maintainers of the project, and we clarify that the
"exclusive" copyright of the co-founders covers the timespan before
opensourcing (their further contributions are included as part of Godot
Engine contributors).
Also fixed "cf." Frenchism - it's meant as "refer to / see".
Backported from #70885.
To allow maximum flexibility in the initial rollout, VisualServer functions were not bound in order to prevent compatibility issues if the API changed.
These functions are now bound to allow direct use from the VisualServer.
The functions that used it already use a threadsafe FIFO queue
to communicate between threads and a sync to have the main thread
wait for the render thread.
Fixes#35718
Backport of this PR: https://github.com/godotengine/godot/pull/63597
This adds these as new Built-Ins to Spatial Shaders
* Object's Position in World Space
* Camera Position in World Space
* Camera Direction in World Space
* Object's Position in View Space
Some logging messages were still being issued when portals/debug/logging was false. This could be annoying in games that stream in parts of levels and repeatedly call `rooms_convert()`.
This PR allows all but essential logging to be disabled.
In situations where rooms are converted multiple times, the previous room hint ID can reference a room number that is out of range of the new number of rooms. This fixes the bug by checking the room hint ID is within range.
Adds special logic for handling skeleton bounding rect updates. Previously these were never being updated because the canvas item is never set to "rect_dirty".
Adds instances to the transform update list as well as the interpolate update list when unhiding them. This ensures that the system auto-detects non-moving objects, and removes them from the interpolate update list on the next tick, preventing unnecessary updates.
A previous optimization prevented instances being added to the interpolation lists when hidden to save processing. This caused a regression when unhiding nodes outside of the physics tick - the interpolated transforms would be stale until the next physics tick, causing a glitch.
This PR readds instances immediately to the interpolation lists when they are unhidden, preventing this glitch.
