An NVIDIA profile is applied to the current executable to disable
threaded OpenGL optimizations on Windows (see #71472). But because the
application is only added to the profile upon the profile creation,
newer executables won't be added to the profile (e.g. if the profile is
created on first launch of Godot_v4.1-stable_win64.exe, when users
update the editor and launch Godot_v4.2-stable_win64.exe, the profile
will never be applied to this new executable).
This patch fixes that scenario by splitting creating the profile (if it
doesn't exist) and adding the application (if it doesn't have a profile
applied) into two separate steps.
Applications that have been manually added to a different profile aren't
overriden to avoid confusing users who know what they're doing.
This means no CPU occlusion culling (and not compiling Embree), unless
you compile custom export templates with `module_raycast_enabled=yes`.
This reduces the memory footprint significantly, and binary size.
Fixes#70621.
Co-authored-by: Hugo Locurcio <hugo.locurcio@hugo.pro>
The UWP platform port was never ported to the Godot 4.0+ API,
and it's now accumulating bitrot as it doesn't compile, and thus
we no longer propagate platform changes in it.
So we finally remove to acknowledge this state. There's still some
interest in reviving the UWP port eventually, especially as support
for Direct3D 12 will soon be merged, but when that happens it will
be easiest to redo it from scratch.
This allows to include script_instance.h directly in the
generated gdvirtual.gen.inc, and remove excessive includes
from the codebase.
This should also allow Resource to use GDVIRTUAL macros,
which wasn't possible previously due to a circular dependency.
Code now always double quotes the filename to use as command line
argument when calling explorer.exe. In particular, commas in a filename
would be interpreted by explorer.exe as separators for commands.
Similarly a trim_suffix for "file://" is assumed to be a mistake, this
could potentially be a PREfix that we want to strip, but never a suffix.
Since it didn't seem needed in the end, we removed it.
It wrongly returned 20 on array buffers, which used to be the enumerator
value of Godot 3.x's type PoolByteArray, and now is the value of type Color,
while it should return 29 which is the enumerator value for PackedByteArray.
When a hardware keyboard is connected, all key events come through so we can route them directly to the engine.
This is not the case for soft keyboards, for which the current logic was designed as it requires extra processing.
Upon investigating the extremely slow MSVC build times in #80513, I noticed
that while Godot policy is to never use exceptions, we weren't enforcing it
with compiler flags, and thus still included exception handling code and
stack unwinding.
This is wasteful on multiple aspects:
- Binary size: Around 20% binary size reduction with exceptions disabled
for both MSVC and GCC binaries.
- Compile time:
* More than 50% build time reduction with MSVC.
* 10% to 25% build time reduction with GCC + LTO.
- Performance: Possibly, needs to be benchmarked.
Since users may want to re-enable exceptions in their own thirdparty code
or the libraries they compile with Godot, this behavior can be toggled with
the `disable_exceptions` SCons option, which defaults to true.
This makes it easy to retrieve the project version at runtime
for display purposes, while simplifying the export preset configuration.
You can now leave the version empty unless you need to override it on a per-preset
basis.
Since export presets save the values of default values to the `export_presets.cfg`
file, this change only affects export presets created after this commit was merged.
This adds the ability for games to obtain platform-specific information about joypads such as their vendor/product ID, their XInput gamepad index or the real name of the device before it gets swapped out by the gamecontrollerdb's name.
This PR also includes a rebased version of #76045, this is because this PR is intended to be mainly to help people implementing Steam Input, as having the gamepad index is essential.
Body length cannot be reliably retrieved from the web.
Reading the "content-length" value will return a meaningless value when
the response is compressed, as reading will return uncompressed chunks
in any case, resulting in a mismatch between the detected body size and
the actual size returned by repeatedly calling read_response_body_chunk.
Additionally, while "content-length" is considered a safe CORS header,
"content-encoding" is not, so using the "content-encoding" to decide if
"content-length" is meaningful is not an option either.
We simply must accept the fact that browsers are awful when it comes to
networking APIs.
The previous packaging format for Godot Android plugins consisted of the plugin's `gdap` config file accompanied by binaries defined in the `gdap` file.
This format is now deprecated (starting with Godot 4.2), and instead Godot Android plugins are now packaged as `EditorExportPlugin` plugins.
The `EditorExportPlugin` class has been updated with the following methods to provide the necessary set of functionality:
- `_supports_platform`: returns true if the plugin supports the given platform
- `_get_android_dependencies`: retrieve the set of android dependencies (e.g: `org.godot.example:my-plugin:0.0.0`) provided by the plugin
- `_get_android_dependencies_maven_repos`: retrieve the urls of the maven repos for the provided android dependencies
- `_get_android_libraries`: retrieve the local paths of the android libraries (AAR files) provided by the plugin
- `_get_android_manifest_activity_element_contents`: update the contents of the `<activity>` element in the generated Android manifest
- `_get_android_manifest_application_element_contents`: update the contents of the `<application>` element in the generated Android manifest
- `_get_android_manifest_element_contents`: update the contents of the `<manifest>` element in the generated Android manifest
This hugely reduces the number of exports, making it acceptable for
browsers.
Note that dlink + threads is still not working due to upstream issues
with the pthread emulation library.
Should hopefully be solved once emscripten move to native WASM threads.
Decouples the Godot java entry point from the Android Fragment component. This enables the Godot component to be more easily reused across different types of Android components including Activities and Services.
We can't rely on the error code from `gio` or `kioclient5`, in my
rudimentary testing they return `1` for both missing files and other
situations like not having a Trash can on the mounted volume.
Fixes#79108.
During GDC and general testing on Steam Deck units, we found that single
gamepads would often register inputs twice under certain circumstances.
This was caused by SteamInput creating a new virtual device, which Godot
registers as a second gamepad. This resulted in two gamepad devices
reporting the same button presses, often leading to buggy input response
on games with no multi-device logic and other-wise could cause intended
Steam rebindings to not work as intended (for example, swapping o and x
on a playstation pad if that feature isn't supported by the game.)
SDL gets around this by taking in a list of devices that are to be
ignored. When valve sees a controller that wants to be rebound via
SteamInput, they push a new VID/PID entry onto the environment
variable `SDL_GAMECONTROLLER_IGNORE_DEVICES` for the original gamepad
so that all game inputs can be read from the virtual gamepad instead.
This leverages the same logic as we are already using SDL gamepad
related HID mappings.
We don't use that info for anything, and it generates unnecessary diffs
every time we bump the minor version (and CI failures if we forget to
sync some files from opt-in modules (mono, text_server_fb).
When trying to export a C# project, this displays an error message after
creating a export preset for an unsupported platform.
Support for these platforms is planned for a future release.
CPPcheck found most of them.
no need to assign the variable twice:
- AnimationTrackEditTypeAudio
- SSEffects
variable is assigned in all if-else clauses:
- EditorHelp
- AndroidInputHandler
- MenuBar
- ShaderCompiler
same if clause:
- ItemList
clearing an empty bitfield has no effect:
- Viewport
Follow-up to #75932.
Since these icons are only used by the export plugin, it makes sense to
move them and generate the headers there.
The whole `detect.is_active()` logic seems to be a leftover from before
times, as far back as 1.0-stable it already wasn't used for anything.
So I'm removing it and moving the export icon generation to
`platform_methods`, where it makes more sense.
Fixes godotengine#76847, Fixes godotengine#74286
Reorders problematic steps from the DisplayServerWindows::window_set_mode function.
These were causing a resize with the dimensions of the previously used mode instead
of the current one.
This applies our existing style guide, and adds a new rule to that style
guide for modular components such as platform ports and modules:
Includes from the platform port or module should be included with relative
paths (relative to the root folder of the modular component, e.g.
`platform/linuxbsd/`), in their own section before Godot's "core" includes.
The `api` and `export` subfolders also need to be handled as self-contained
(and thus use relative paths for their "local" includes) as they are all
compiled for each editor platform, without necessarily having the api/export
matching platform folder in the include path.
E.g. the Linux editor build will compile `platform/android/{api,export}/*.cpp`
and those need to use relative includes for it to work.
The follow options were added to the (new) `run/window_placement/android_window` editor setting:
- `Same as Editor`: run the project in the same window as the editor
- `Side-by-side with Editor`: run the project in an adjacent window to the editor
- `Auto`: choose how to run the project based on the device screen size
The default behaviour for X11 is to crash even on non-fatal errors
when there is no error handler set. This change allows the window to
stay open and may enable users to save their work when things go
wrong.
This acts as a workaround for #65425 and #68471
plus minor static-related fixes
* linuxbsd: get_systemd_os_release_info_value() -> static breaks usage if used multiple times
* windows/linuxbsd: get_video_adapter_driver_info() writes info into static
* linuxbsd: get_distribution_name() + get_version() -> write bsd fallback into static variable
* windows/uwp/android: remove unnecessary use of static
- Fix issues with using a touchpad to click, drag, interact with the navigation controls, etc..
- Fix issues with panning with 2+ fingers
- Fix issues with using double-tap to magnify on the spatial editor
Move the benchmarking measuring methods from `Engine` to `OS` to allow for platform specific overrides (e.g: can be used to hook into platform specific benchmarking and tracing capabilities).
The `InputEvent` class currently supports the `pressed` and `released` states, which given the binary nature, is represented by a `bool` field.
This commit introduced the `CANCELED` state, which signals that an ongoing input event has been canceled.
To represent all the states, the `InputEventState` enum is added and the `InputEvent` logic is refactored accordingly.
Turns out the list of trusted root certificates contains disallowed
certificates (i.e. certificates which are no longer trusted or have been
revoked).
We need to check for the property `CERT_DISALLOWED_FILETIME_PROP_ID` to
check if and when the certificates should be distrusted.
udev doesn't work in sandboxes, notably the new Steam container runtime
as found notably on the Steam Deck, and in Flatpak/Snap packages.
Like SDL does, when we detect such a containerized environment, we fall
back to parsing `/dev/input` directly.
See smcv's comments in #76879 for details.
Fixes#76879.
Adds a new OS::get_system_ca_certs method which can be implemented by
platforms to retrieve the list of trusted CA certificates using OS
specific APIs.
The function should return the certificates in PEM format, and is
currently implemented for Windows/macOS/LinuxBSD(*)/Android.
mbedTLS will fall back to bundled certificates when the OS returns no
certificates.
(*) LinuxBSD does not have a standardized certificates store location.
The current implementation will test for common locations and may
return an empty string on some distributions (falling back to the
bundled certificates).
Fixes two related segfaults caused by running the DBus version check
unconditionally and potentially calling null function pointers.
This was clearly an oversight as all the other wrappers have proper
loading status checks.
Instead of reading files over the network, the new version uses a local file cache and only updates files when it changes.
The original remote filesystem was created 14 years ago, when ethernet was faster than hard drives or even flash. Also, mobile devices have a very small amount of storage.
Nowadays, this is no longer the case so the approach is changed to using a persistent cache in the target device.
Co-authored-by: m4gr3d
