fixes#57231, fixes#57421
Within the context of parsing navigation geometry, this commit:
- added missing transform of `MultiMeshInstance`
- changed all transforms to global ones so that they don't need to be
calculated by hand
By default, when installing from Android App Bundles the native
libraries are not extracted. They are loaded directly from the APK.
See: https://stackoverflow.com/a/56551499
Passing only the file name to dlopen, without the location, makes it
search the native library in all locations, including inside the apk.
A common source of errors is to call functions (such as round()) expecting them to work in place, but them actually being designed only to return the processed value. Not using the return value in this case in indicative of a bug, and can be flagged as a warning by using the [[nodiscard]] attribute.
Keep applying the windows entropy patch (UWP support).
Remove no longer needed padlock patch.
Update thirdparty README to reflect changes, and new source inclusion
criteria.
(cherry picked from commit e375cbd094)
I don't see any reference to gzip/svgz supported in the nanosvg library,
and the handful of test gzip compressed svgz files I tried failed loading.
Also cleaning a couple missing includes in platform export code.
The command line option is the same you would pass to Mono, but it
begins with `--mono-debugger-agent=` instead of `--debugger-agent=`.
This is useful for platforms where it's difficult for the user to
configure it via environment variables, like Android.
Change the entire navigation system.
Remove editor prefix from nav mesh generator class. It is now used for baking
at runtime as well.
Navigation supports obstacle avoidance now with the RVO2 library.
Nav system will also automatically link all nav meshes together to form one
overall complete nav map.
When reloading C# classes and keep their properties values they are
retrieved and stored in a state list.
Retrieving the properties was only getting the fields of the C# class
and not inherited fields so those properties values were lost on reload.
Now we also try to find the field in the parent classes.
When folders are moved/removed from the file system, the `.csproj`
may need to be edited to update the path of C# scripts or remove them.
If a C# solution has not been created, the `.csproj` file does not exist
and therefore there is no need to edit it.
Due to the async nature of WebRTC implementations, the multiplayer peer
could end up having queued packets from a given connection before it is
able to emit the "peer_added" signal.
This commit ensures that packets from peers which are not notified yet
are skipped by `get_packet` and `get_available_packet_count`.
Two main changes:
- Better handling of concave shapes to make sure the queries don't
return a triangle index instead of shape index.
Note: A concave shape within a compound shape will always return a shape
index of 0 because of Bullet limitations.
- Extra check for compound shapes in some queries to avoid undefined
behavior, because the shape index can have an uninitialized value with
convex shapes in some cases.
(cherry picked from commit 02d40de30d)
Removes miniz as a bundled dependency, relies on our own zlib instead.
Includes a couple commits ahead of `v1.0.1` tag to fix MinGW builds.
(cherry picked from commit 46d3effa99)
Moves copy and paste in their own functions so copy_nodes_request and paste_nodes_request are able to work.
Applies paste offset to the last mouse clicked position.
Settings that aren't within a subsection are difficult to reach when
other settings do have a subsection.
This also adds documentation for the project setting.
In all physics servers, body_get_direct_state() now silently returns
nullptr when the body has been already freed or is removed from space,
so the client code can detect this state and invalidate the body rid.
In 2D, there is no change in behavior (just no more errors).
In 3D, the Bullet server returned a valid direct body state when the
body was removed from the physics space, but in this case it didn't
make sense to use the information from the body state.
The hash symbol creates spurious issue references on GitHub if
the message is posted outside a code block, which means some issues
have a lot more references than originally intended.
Godot uses Variant parameters for calls to script methods.
Up until now we were boxing such parameters when marshalling
them for invokation, even if they were value types.
Now Godot allocates the marshalled parameters on the stack,
reducing the GC allocations resulted from boxing.
Keep track of MeshInstance and GeometryInstance override materials in the GLTFMesh object.
Ensure all arrays are non-empty to conform with "minItems":1 in glTF spec.
Fleshed out the "Optimize Mesh" options found in the mesh import UI
Gave a checkbox to every vertex attribute that can be compressed
Surfaced option to enable/disable Octahedral compression for
normal/tangent vectors
Also surfaces the vertex position compression option which previously
inaccessible because the defaults did not compress vertex positions
Supports all current importers (obj, fbx, collada, gltf)
Ensures that the `get_property_list` and `get_script_property_list`
methods push the script properties to the end of the given list, this
prevents the script property from appearing after the script variables.
Sets `AlignOperands` to `DontAlign`.
`clang-format` developers seem to mostly care about space-based indentation and
every other version of clang-format breaks the bad mismatch of tabs and spaces
that it seems to use for operand alignment. So it's better without, so that it
respects our two-tabs `ContinuationIndentWidth`.
GLB chunk padding length calculation was backwards and missing for the BIN chunk.
Fixed error caused by "skins":[] when no skins were present.
Finally, encode animations before textures to avoid accessor misalignment due to texture byteLength.
It triggers a crash when playing V9 videos.
Could likely be fixed if anyone wants to work on it, but so far nobody seems to
want to and WebM support is dropped in 4.0, so this workaround should help for
now.
Fixes#50862.
This provides more realistic lighting with a very small performance cost.
The option is available in both GLES3 and GLES2, and can be enabled in
the Project Settings. This goes well with the ACES Fitted tonemapping mode
that was recently added.
When enabled, this also makes upgrading Godot 3.x projects to Godot 4.0 easier,
since lighting in 3.x will better match how it'll look in Godot 4.0.
And do the dedent and stripping for both translated and
non-translated strings for consistency, and so that we
don't need to do it at the call site.
(cherry picked from commit a16031beb6)
Create GLTFSkeleton at the same time we create GLTFNode objects.
Create GLTFSkin at the same time we export MeshInstance3D
Fixes export of blend shape arrays for meshes with multiple surfaces.
Fixes array indexing issues in export of glTF morph target animations.
Converts BoneAttachment3D nodes during normal node creation: this avoids
special cases during mesh export, and especially exporting skeletons or meshes
which are children of BoneAttachment3D.
Co-authored-by: K. S. Ernest (iFire) Lee <ernest.lee@chibifire.com>
This can be used to free some CPU cores when baking lightmaps.
When using fewer CPU cores, lightmap baking is slower but background
tasks aren't slowed down as much.
Return a const String reference to make sure that when its address
is used as a handle in NativeScriptLanguage::init_library(),
it refers to the heap instead of a temporary object on the stack.
Stop include Bullet headers using `-isystem` for GCC/Clang as it misleads
SCons into not properly rebuilding all files when headers change.
This means we also need to make sure Bullet builds without warning, and
current version fares fairly well, there were just a couple to fix (patch
included).
Increase minimum version for distro packages to 2.90 (this was never released
as the "next" version after 2.89 was 3.05... but that covers it too).
Prevents a crash when calling test_body_motion. Call reload_kinematic_shapes
from init_kinematic_utilities as they are always called together.
(cherry picked from commits 112985c5bc
and 3540e716f9)
It used to call `enet_host_service` until all events were consumed, but
that also meant constantly polling the connection leading to potentially
unbounded processing time.
It now only service the connection once, and instead consumes all the
retrieved events via `enet_host_check_events`.
It used to always send them reliably when transfer mode was unreliable
or ordered if the packet size was more then the enet host MTU (1400
bytes by default).
This commit also adds a warning when debug is enabled to explain the
effects of sending fragmented packets unreliably.
Pawel Lampe
Marcel Admiraal
Rémi Verschelde
Ignacio Roldán Etcheverry
David Snopek
Yuri Sizov
rafallus
Max Hilbrunner
Hugo Locurcio
lawnjelly
Dardan Bujupaj
Raul Santos
Rémi Verschelde
Fabio Alessandrelli
Haoyu Qiu
GlyphTheWolf
Arnav Vijaywargiya
K. S. Ernest (iFire) Lee
Bartłomiej T. Listwon
LATRio
SaracenOne
bruvzg
kobewi
RedMser
Jake Young
Zae
Aaron Franke
Francois Belair
Cory Petkovsek
Vincent
PouleyKetchoupp
jitspoe
David Cambré
Daniel Kolesa
Mark Riedesel
Pedro J. Estébanez
BimDav
Lyuma
Omar El Sheikh
Ev1lbl0w
M. Huri
Infinixius
Rafał Mikrut
Kirill Diduk
Stéphane Fortin
Kongfa Waroros
Gergely Kis
Manuele Finocchiaro
Jordan Schidlowsky
Germanrc