It made minor adjustments difficult as the camera moved every time
Align Transform With View was used.
This closes#36738.
(cherry picked from commit a636631e33)
The indexes for the ScriptEditorItemData entries were not getting
updated after sorting. This would cause a page to be open but with a
different tab selected. Whenever _update_script_names was called next,
it would correct this indexing. Now we correct it immediately following
the tab sort.
(cherry picked from commit fe53c1b0bd)
This fixes numerous false positives coming out of the culling system.
AABB checks are now a full separating-axis check against the frustum, with the points of the frustum being compared to the planes of the box just as the points of the box were being compared to the planes of the frustum. This fixes large objects behind the camera not being culled correctly.
Some systems that used frustums that were (sometimes mistakenly?) unbounded on one or more side have been modified to be fully enclosed.
This makes the script name appear before the scene file name,
which ensures it's always visible even if the list of scripts is too
narrow to display the full name.
This only impacts built-in scripts with custom resource names.
Unnamed resources will still use `<scene_file>::<id>` naming
in the list of scripts.
(cherry picked from commit e90beade15)
The Request Docs button is partly responsible for layout overflow
issues on narrow displays, such as #31133.
It also tended to attract spam and low-effort issues that were
difficult to act upon. A "Send Docs Feedback" menu option has been added
to replace it.
(cherry picked from commit 2f6f029a75)
Fixes#18809.
The in-editor ortho camera used a far clipping plane of 8192 units, and was
placed 4096 units away from the camera cursor.
This was far enough to cause culling issues from floating point precision loss
on objects smaller than one unit.
This change modifies the near/far clipping planes of the ortho camera to use
those specified in the editor (and currently used by the perspective camera).
The frustum is still centered around the camera cursor location, as it was
before.
Changes made:
* Added dirty bit for SpatialEditorSelectedItem's last_xform
* SpatialEditorViewport checks that dirt bit too before skipping the selection
(cherry picked from commit 19825436d4)
Alignment of scene pixels on screen pixel ensure a crisp rendering of small features (such as text). Unfortunately, alignment of top left pixel on screen adds a lot of jittering when zooming at high zoom factor.
This change allow to snap the top left scene pixel on the closest screen pixel (not only the top-left most), and we do so only when the scale factor is an integer.
(cherry picked from commit 1c02906a6f)
The AABB for an empty Spatial has 0 size, since the stored and
compared Transform was scaled by the AABB size, it would completely
destroy the rotation information. If there is no rotation
information, the gizmo doesn't update when the rotation changes.
(cherry picked from commit a0af3094b4)
The new default keys were chosen to match common graphics editing
software. A modifier is no longer required to use the Select tool,
making it faster to use.
This closes#34170.
(cherry picked from commit 88213b54ba)
It was removed after the implementation of VHACD. Generating a single
shape can lead to better performance, so it may still be desired.
This also adds tooltips for several options in the Mesh menu.
This closes#35692.
(cherry picked from commit 90af009f2e)
Suppose that the user wants to use some guidelines in 2D mode. The
user has enabled "Use Pixel Snap", and configured the "Grid Step" to
1px.
On some zoom levels, when dragging the guidelines step by step, some
offsets shows the wrong value. The offsets that are wrong vary - it is
affected by the zoom level, so some zoom levels do not display this
problem.
For example, a user may see this while dragging the guideline:
0px 1px 1px 3px 4px 5px 5px 7px 8px
whereby 2px and 6px are missing.
This is due to a floating-point error. The values are printed as
(truncated) integers, but they are actually decimals, so they were
actually 1.9999 and 5.9999 for the missing cases.
Let's fix that by rounding up the values before printing them to get rid
of the errors.
This fixes#35010.
Fixes#26637.
Fixes#19900.
The viewport_size returned by get_viewport_size was previously incorrect, being half the correct value. The function is renamed to get_viewport_half_extents, and now returns a Vector2.
Code which called this function has also been modified accordingly.
This PR also fixes shadow culling when using ortho cameras, because the correct input for CameraMatrix::set_orthogonal should be the full HEIGHT from get_viewport_half_extents, and not half the width.
It also fixes state.ubo_data.viewport_size in rasterizer_scene_gles3.cpp to be the width and the height of the viewport in pixels as stated in the documentation, rather than the current value which is half the viewport extents in worldspace, presumed to be a bug.
