- For now everything imports multithreaded by default (should work I guess, let's test).
- Controllable per importer
Early test benchmark. 64 large textures (importing as lossless, _not_ as vram) on a mobile i7, 12 threads:
Importing goes down from 46 to 7 seconds.
For VRAM I will change the logic to use a compressing thread in a subsequent PR, as well as implementing Betsy.
-Decode/Encode functions for u8,s8,u16,s16,u32,s32,u64,s64,half,float,double,variant
-Improved binder template to allow this
Given in Godot 4.0 PackedByteArray is passed as reference, it is now possible to have these functions there, which makes the most sense.
* Added option for importers to show an Advanced settings dialog
* Created advanced settings dialog for Scene Importer
* Cleaned up importers (remove many old/unused options)
* Added the ability to customize every node, material, mesh and animation individually
* Saving to animations and meshes to files is now a manual process, making it more predictable
* Added the ability for materials to be replaced by external files (or to be made external, up to you).
* When doubleclicking an impoted scene in the filesystem dock, it automatically shows the import settings instead of asking to open it.
WARNING: Lightmap UV unwrap is not working, it needs to be re-made.
This ensure that typed arrays are properly checked when setting an
element.
Moved the macro to a straight declaration since the macro was only used
for Array and it now is quite specific to the Array class.
The array should just assimilate the type of the other one since
assignment in this case means a change in the reference.
This also adds a `typed_assign` function for the cases where type
validation is wanted.
Make sure that RemoteDebuggerPeer wait at least 100us between polls
(effectively forcing a min tick of 100 microseconds).
This greatly improve performances (the call to poll was useless since
during low traffic, writes would always be available, and during high
traffic, reads would always be available, effectively making it a
busy-waiting loop). We could further improve this, by separating the two
polls, and adjust the min tick based on load, but this is most likely
more than enough already without sacrificing too much on high loads.
* Properly exposed, including validated and variant call
* Bound static functions in String and Color
* Did not add support for scripting languages, will have to be added manually.
Other functions in the same file validate parameters using the ERR_FAIL
macros. This validation was missing for Variant::construct resulting in
a crash when called with invalid data (p_type < 0).
fix#46067
The problem happened because `poll` assumed that when the SSL flag was
true, the `connection` would be a subclass of StreamPeerSSL. However
that invariant could be broken by calling HTTPClient::set_connection
with a `connection` that is not a subclass of StreamPeerSSL.
Fixes#46138
Removes the error message when the network peer is not valid and returns false instead.
This makes it simpler to make games that are both on/offline by replacing server checks of
'''
if is_instance_valid(get_tree().network_peer) and get_tree().is_network_server():
# Do server things
'''
with
'''
if get_tree().is_network_server():
# Do server things
'''
Requires no changes to the docs because both the MultiplayerAPI and SceneTree docs don't mention the error.
Thow errors if requesting an unexisting InputMap action.
Makes `Input.is_action_*` methods consistents with `Event.is_action_*` which already throw errors.
fixes#33303
That's the most common one we've been using for the general editor, and while the script editor
also supports Ctrl+Y, it should have lower priority.
In theory this code should make both be supported the same but for some reason the general editor
only seems to use the first entry (the script editor does support both).
- Based on C++11's `atomic`
- Reworked `SafeRefCount` (based on the rewrite by @hpvb)
- Replaced free atomic functions by the new `SafeNumeric<T>`
- Replaced wrong cases of `volatile bool` by the new `SafeFlag`
- Platform-specific implementations no longer needed
Co-authored-by: Hein-Pieter van Braam-Stewart <hp@tmm.cx>
This removes hardcoded actions from things like LineEdit and TextEdit.
Previously, things like copy, paste, etc were all hardcoded to Ctrl+C, Ctrl+V, etc. They could not be changed. This allows the possibility of them being changed, by making them use the action map. This has the added benefit of greatly simplifying the input handling logic in those controls. The logic which was previously in a huge and hard to follow switch statement has been extracted to individual methods.
This adds the ability to add overrides for built-in actions (i.e. ui_*) in the editor. Also added a number of additional built-in actions for various text-related actions, gui-generic actions (like copy and paste) and graph-related actions (duplicate nodes), etc. Moved the definition of input actions to input_map, rather than in project_settings so the editor can make use of these actions as well.
-Advanced Settings toggle also hides advanced properties when disabled
-Simplified Advanced Bar (errors were just plain redundant)
-Reorganized rendering quality settings.
-Reorganized miscelaneous settings for clean up.
* Add support for decimal numeric entities to String::xml_unescape
* Add more error checks to String::xml_unescape
* Refactor XMLParser to use String::xml_unescape instead of an internal
implementation
This can be used to ensure a file has its contents saved
even if the project crashes or is killed by the user
(among other use cases).
See discussion in #29075.
-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.
Cylinder collision detection uses a mix of SAT and GJKEPA.
GJKEPA is used to find the best separation axis in cases where finding
it analytically is too complex.
Changes in SAT solver:
Added support for generating separation axes for cylinder shape.
Added support for generating contact points with circle feature.
Changes in GJKEPA solver:
Updated from latest Bullet version which includes EPA fixes in some
scenarios.
Setting a lower EPA_ACCURACY to fix accuracy problems with cylinder vs.
cylinder in some cases.
-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).
Channels that are inactive -or when playback has not started yet- will report -200 dB as their peak value (which is also the lowest value possible during playback).
- 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)
Named color constants renamed to UPPERCASE. Unlike #41019, this PR
is complete and implements these changes in the simplest way possible.
Co-authored-by: Shivam Mukherjee <mshivam98@gmail.com>
- Based on C++14's `shared_time_mutex`
- No more need to allocate-deallocate or check for null
- No pointer anymore, just a member variable
- Platform-specific implementations no longer needed
- Simpler for `NO_THREADS`
No longer use emscripten functions for gamepads, implement them as
library functions in library_godot_display.js instead.
This allows us to do a better job at "guessing" vendorId, productId, OS,
etc. thus allowing us to better find the remapping for the controller.
In the core input handling code we have checks to make sure that if axis
rapidly change sign we inject mid-points to release any pending inputmap
action.
The function though, did not correctly insert the mid-point causing
dpads mapped to an axis that behaves like tri-state buttons (-1,0,1) to
not be released correctly.
This commit fixes that by including in the check the case where the axis
swtiches from abs(1) to 0.
Happy new year to the wonderful Godot community!
2020 has been a tough year for most of us personally, but a good year for
Godot development nonetheless with a huge amount of work done towards Godot
4.0 and great improvements backported to the long-lived 3.2 branch.
We've had close to 400 contributors to engine code this year, authoring near
7,000 commit! (And that's only for the `master` branch and for the engine code,
there's a lot more when counting docs, demos and other first-party repos.)
Here's to a great year 2021 for all Godot users 🎆
This can be used to print thread IDs in logs. This can make it easier
to debug multi-threaded applications.
Co-authored-by: Khaos <khaos@khaos-coders.org>
To get counts of items before getting the list, which is useful for
GDNative so users can pre-allocate the buffer with the correct size
without having to get the list twice.
The calls to the BVH need to use the world space AABB, rather than local space for it to work. Also, update was not being called which is required to update the AABB as objects move.
-Much greater pairing/unpairing performance
-For now, using it for culling too, but this will change in a couple of days.
-Added a paged allocator, to efficiently alloc/free some types of objects.
As a bonus, to have consistency between use Beziers and create insert tracks, use Beziers also gets a default via editor settings that is used when the confirmation dialog is disabled, instead of just falling back to creating non-Bezier tracks.
-Based on Bullet Dbvh, has style and functional changes.
-Provides efficient pairing
-Needed to optimize rendering
-Needed to optimize physics
This PR is up for others to review the implementation.
-Happens on import by default for all models
-Just works (tm)
-Biasing can be later adjusted per node or per viewport (as well as globally)
-Disabled AABB.get_support test because its broken
Since we clone the environments to build thirdparty code, we don't get an
explicit dependency on the build objects produced by that environment.
So when we update thirdparty code, Godot code using it is not necessarily
rebuilt (I think it is for changed headers, but not for changed .c/.cpp files),
which can lead to an invalid compilation output (linking old Godot .o files
with a newer, potentially ABI breaking version of thirdparty code).
This was only seen as really problematic with bullet updates (leading to
crashes when rebuilding Godot after a bullet update without cleaning .o files),
but it's safer to fix it everywhere, even if it's a LOT of hacky boilerplate.
Added additional param to action related methods to test for exactness.
If "p_exact_match" is true, then the action will only be "matched" if the provided input event *exactly* matches with the action event.
Before:
* Action Event = KEY_S
* Input Event = KEY_CONTROL + KEY_S
* Is Action Pressed = True
Now:
You can still do the above, however you can optionally check that the input is exactly what the action event is:
* Action Event = KEY_S
* Input Event = KEY_CONTROL + KEY_S
* p_exact_match = True
* Is Action Pressed = False
* If the Input Event was only KEY_S, then the result would be true.
Usage:
```gdscript
Input.is_action_pressed(action_name: String, exact_match: bool)
Input.is_action_pressed("my_action", true)
InputMap.event_is_action(p_event, "my_action", true)
func _input(event: InputEvent):
event.is_action_pressed("my_action", false, true) # false = "allow_echo", true = "exact_match"
event.is_action("my_action", true)
```
We haven't had a proper implementation for COMPRESS_PVRTC2 (which is PVRTC1 2-bpp) in years,
so let's drop it instead of keeping a compress type which doesn't work.
The other enum values were renamed to clarify that our PVRTC2 and PVRTC4 are respectively
PVRTC1 2-bpp and PVRTC1 4-bpp. PVRTC2 2-bpp and 4-bpp are not implemented yet.
Instead of references. This is needed because those function pointers
are used in GDNative which needs to work with plain C, which doesn't
support passing parameters by reference.
Alternative to `_change_notify()` to be called from within C++ classes.
Achieves low-level consistency with scripting, where this method is
exposed for updating the editor (inspector) with new values.
- added `state` as a property to restore internal state of RNG;
- `get_seed()` returns last seed used to initialize the state rather than the current state.
Co-authored-by: MidZik <matt.idzik1@gmail.com>
- Avoid spaces in Mono log file names.
- Use a `.log` extension for Mono logs, just like non-Mono logs.
- Use periods to separate hours/minutes/seconds for non-Mono logs.
-Changed how mesh data is organized, hoping to make it more efficient on Vulkan and GLES.
-Removed compression, it now always uses the most efficient format.
-Added support for custom arrays (up to 8 custom formats)
-Added support for 8 weights in skeleton data.
-Added a simple optional versioning system for imported assets, to reimport if binary is newer
-Fixes #43979 (I needed to test)
WARNING:
-NOT backwards compatible with previous 4.x-devel, will most likely never be, but it will force reimport scenes due to version change.
-NOT backwards compatible with 3.x scenes, this will be eventually re-added.
-Skeletons not working any longer, will fix in next PR.
- Initialize Object pointer to nullptr so it's not used by mistake.
- When setting an Object check if it's a reference so refcounting works
as intended.
- ClassDoc added to GDScript and property reflection data were extracted
from parse tree
- GDScript comments are collected from tokenizer for documentation and
applied to the ClassDoc by the GDScript compiler
- private docs were excluded (name with underscore prefix and doesn't
have any doc comments)
- default values (of non exported vars), arguments are extraced from the
parser
- Integrated with GDScript 2.0 and new enums were added.
- merge conflicts fixed
Made InputEvent as_text() return a readable and presentable string. Added to_string() overrides for each which returns a 'debug-friendly' version which is not as presentable but provides more information and in a more structured fashion. Use as_text() for UI display scenarions and to_string() for debug cases
ptrcall is now also used to optimize calls in GDScript, on top of the existing
use by the GDNative and Mono modules.
It no longer makes sense to make it optional.
- Allow getting an opaque pointer, no matter the type (for ptrcall).
- Allow setting object pointer and id directly.
- Allow initializing the data given a type, to allow properly setting
return types on ptrcalls.
The validated getters were only setting the value without changing the
type, leading to wrong results. This uses the same path used for
methods to the same purpose.
Made serialization of Command optional. If command is serialized, Control (On Win/Linux) or Meta (on Mac) are not.
Example use case: You are on Windows and you set a shortcut to be Control + E. This would serialize as Command=true and Control=true. If you then run this project on Mac, you would need to press Command AND Control to activate the shortcut - which is not what is intended. Now, you can set store_command to true, and it will only serialize to Command = true (no Control serialized). On Windows, this means Control. On Mac, it means only command.
Copied relevant documentation from the original `@GDScript` built-ins,
which will likely be removed in a future commit.
Various fixups to `variant_utility.cpp` while working on this.
* Using C-style function pointers now, InternalMethod is gone.
* This ensures much better performance in typed code.
* Renamed builtin_funcs to utility_funcs, to avoid naming confusion
-Moved Expression to use this, removed its own.
-Eventually GDScript/VisualScript/GDNative need to be moved to this.
-Given the JSON functions were hacked-in, removed them and created a new JSONParser class
-Made sure these functions appear properly in documentation, since they will be removed from GDScript
The underscore prefix was used to avoid the conflict between the `RID` class
name and the matching enum value in `Variant::Type`.
This can be fixed differently by prefixing uses of the `RID` class in `Variant`
with the scope resolution operator, as done already for `AABB`.
-Discern between named, indexed and keyed
-Get direct access to functions for typed GDScript and GDNative bindings
-Small changes to some classes in order to work with the new setget binder
It's non-deterministic so it's better to show a fixed value like 0 instead of
having it potentially change whenever `randomize()` is called.
Fixes#43317.
Also separated Light2D in PointLight2D and DirectionalLight2D.
Used PointLight2D because its more of a point, and it does not work
the same as OmniLight (as shape depends on texture).
Added a few utility methods to Rect2D I needed.
Returning the most contrasting color isn't a trivial task, as there
are often many possible choices. It's usually best left for the user
to implement using a script.
They're now disabled by default, and can be enabled with the command line
argument `--vk-layers`.
When enabled, the errors about them being missing are now warnings, as
users were confused and thought this meant Vulkan is broken for them.
Fix crash in `~VulkanContext` when validation layers are disabled (exposed by
this PR since before they could not be disabled without source modification).
Also moved VulkanContext member initializations to header.
Fixes#37102.
Moved to a system using variadic templates, shared with CallableBind.
New code is cleaner, faster and allows for much better optimization of core
type functions from GDScript and GDNative.
Added Variant::InternalMethod function for direct call access.
This patch fixes two related issues. One is the race condition in issue #42107..
The other is a crash which happens when the reader is lapped near the end of the buffer.
It was added in 3e20391bf6 but it doesn't seem
particularly useful, and it was only implemented for the custom splash branch
and not the default one, so it could return an uninitialized int.
1024 KB was low enough that many users seem to hit it, which can lead to the
editor freezing.
The proper fixed as described in #35653 would be to implement a page allocator
to prevent this overflow, but as a stop-gap measure, we can increase the
default value to a more lenient 4096 KB which should be high enough for the
vast majority of use cases.
The default size can be brought down again if/when #35653 is properly fixed,
and if it's actually relevant from a memory point of view.
is_equal_approx is able to handle values of any size, and is_equal_approx_ratio is no longer used in any exposed APIs, so we don't need is_equal_approx_ratio anymore. Also, add #ifdef MATH_CHECKS for a method that is only used when MATH_CHECKS is defined.
Add __NetBSD__ to `platform_config.h` so that it can find `alloca`
and use the proper `pthread_setname_np` format.
Rename RANDOM_MAX to avoid conflict with NetBSD stdlib.
Fixes#42145.
- immediately return if A == B;
- first and last elements (front, back) are updated upon relinking;
- handles a special case of forward and backward adjacent elements.
updated variables to use size_t
removed line break to make code style more consistent
added conditional check to return an error if offset field is used when loading a ZIP package
fixed typo
formatted file
added commit regarding self contained exe files
handled error loging for load zip file with offset
spelling tweak
updated conditional statement for magic check
udpated error message when load Zip file with offset is called
fix CI
Trying to fix CI
fix CI done
Added error message for loading self-contained exe with offset.
Updated documentation.
Fix indent
final fix indent
Updated documentation.
fix indents
Updated doc based on suggestion
Final fix
fixed format
This is used to get a direct pointer for the data inside a Variant.
Added as another class instead of in Variant directly because this
should only be used in special cases, so use must be explicit.
Changes:
* EngineDebugger is exposed to gdscript. Game side of communication can be implemented through it.
* EditorDebuggerPlugin is added which handles the editor side of communication.
And revert follow-up regression fix "Remap script path when registering class."
After the regression fix, the original issue is valid again so it's better
to go back to the previous state.
This reverts commits e264ae20d2 and
fceb64827e.
This method starting being used in 079ca220e1,
which now triggers this warning from GCC 10:
```
./core/error_macros.h:151:25: error: comparison of unsigned expression in '< 0' is always false [-Werror=type-limits]
```
New contributors added to AUTHORS:
@hinlopen, @naithar, @rrcore, @SkyLucilfer, @TwistedTwigleg
Thanks to all contributors and donors for making Godot possible!
---
Also changes to relevant code that parses the DONORS.md to match
the new tiers.
A new `env.Run` method is added which allows to control the verbosity
of builders output automatically depending on whether the "verbose"
option is set. It also allows to optionally run any SCons commands in a
subprocess using the existing `run_in_subprocess` method, unifying
the interface. `Action` objects wrap all builder functions to include a
short build message associated with any action.
Notably, this removes quite verbose output generated by `make_doc_header`
and `make_editor_icons_action` builders.
- Enhance directory API
- Fix `FileAccess::exists()` not checking for PackedData being disabled
- Fix moving to the parent directory (`..`)
- Allow absolute paths in existence checks
The input to smoothstep is not actually a weight, and the decscription
of smoothstep was pretty hard to understand and easy to misinterpret.
Clarified what it means to be approximately equal.
nearest_po2 does not do what the descriptions says it does. For one,
it returns the same power if the input is a power of 2. Second, it
returns 0 if the input is negative or 0, while the smallest possible
integral power of 2 actually is 1 (2^0 = 1). Due to the implementation
and how it is used in a lot of places, it does not seem wise to change
such a core function however, and I decided it is better to alter the
description of the built-in.
Added a few examples/clarifications/edge-cases.
In general they are more confusing to users because they expect
inheritance to fully override parent methods. This behavior can be
enabled by script writers using a simple super() call.
Implements exit codes into the engine so tests can return their statuses.
Ideally we don't do this, and we use FIXUP logic to 'begin' and 'end' the engine execution for tests specifically.
Since realistically we're initialising the engine here we don't want to do that, since String should not require an engine startup to test a single header.
This lowers the complexity of running the unit tests and even for
physics should be possible to implement such a fix.
UDPServer now uses a single socket which is shared with the
PacketPeerUDP it creates and has a new `poll` function to read incoming
packets on that socket and delivers them to the appropriate peer.
PacketPeerUDP created this way never reads from the socket, but are
allowed to write on it using sendto.
This is needed because Windows (unlike Linux/BSD) does not support
packet routing when multiple sockets are bound on the same address/port.
- Extacted all syntax highlighting code from text edit
- Removed enable syntax highlighting from text edit
- Added line_edited_from signal to text_edit
- Renamed get/set_syntax_highlighting to get/set_syntax_highlighter
- Added EditorSyntaxHighligher
- Use the `.log` file extension (recognized on Windows out of the box)
to better hint that generated files are logs. Some editors provide
dedicated syntax highlighting for those files.
- Use an underscore to separate the basename from the date and
the date from the time in log filenames. This makes the filename
easier to read.
- Keep only 5 log files by default to decrease disk usage in case
messages are spammed.
So places that need to look into it can use the list instead of parsing
ProjectSettings details (like checking "*" in path for testing if it's
singleton).
Fix errors when comparing strings with large numbers (> INT64_MAX).
Comparisons now occur by comparing individual digits, instead of
converting to int64_t first.
Current error checks for to_int and to_int64 do not issue overflow error
messages for INT64_MAX + 1, INT64_MAX + 2, and other numbers close to the
integer limits. Likewise, error checks for hex_to_int, hex_to_int64 and
bin_to_int64 issue false positive error messages for INT64_MIN or INT32_MIN.
This commit fixes these error checks.
This change makes the behavior consistent when the value
is not found between erasing from an empty list
(no error, just returning false) and erasing from a non-empty list
(previously displaying triggering an error and returning false).
Error message previously triggered:
ERROR: erase: Condition ' !p_I ' is true. returned: false
At: ./core/list.h:157.
Custom monitors can be added/removed/checked using `Performance.add_custom_monitor`/`Performance.remove_custom_monitor`/`Performance.has_custom_monitor`
The value can be viewed in the `Monitor` tab of Debugger.
Text before `/` is used to categorize the custom monitor.
`EditorPerformanceProfiler` class is created to separate logic from `ScriptEditorDebugger`
User can click on the graph of monitors to read the value at that point.
Graph includes intermediate base lines.
Fix calculation for negative times to ensure Sundays are wrapped around to '0'
instead of '7', making it consistent with the output for positive times.
Rémi Verschelde
kobewi
Matheus Lima Cunha
Rémi Verschelde
Juan Linietsky
Hugo Locurcio
Marcel Admiraal
George Marques
Bartłomiej T. Listwon
Julien Nguyen
Fabio Alessandrelli
dam
Johannes
reduz
Daniel Kříž
Danil Alexeev
jmb462
rafallus
mashumafi
Aaron Franke
bruvzg
Alex Hirsch
Fredia Huya-Kouadio
Anshul7sp1
Aitor Cereceto
Gilles Roudière
Ev1lbl0w
Rafał Mikrut
Pedro Rodrigues
Kyle
Mateo Kuruk Miccino
Carlos Cabello
Pedro J. Estébanez
kleonc
Emmanuel Leblond
Bastiaan Olij
abaire
Eric M
Liz Haas
Delf Neumärker
andybarcia
JestemStefan
Henry Conklin
PouleyKetchoupp
Sergey Minakov
Yuri Roubinsky
Vedat Günel
Andrii Doroshenko (Xrayez)
AndreaCatania
Joshua Dahl
Mai Lavelle
Maganty Rushyendra
Marcus Brummer
lawnjelly
Fredia Huya-Kouadio
Marco Cognetta
Theogen Ratkin
Jordan Schidlowsky
Brian Semrau
Andy Savage
Alf Kraus
Thakee Nathees
jcgollnick
Jon Bonazza
Rafał Mikrut
Zae
Haoyu Qiu
Michael Alexsander
Marios Staikopoulos
Lyuma
Serhat
Wilson E. Alvarez
Craig-Stoneham
Adam Brown
Yilin Ma
strank
Yakov Borevich
simpu
SkyJJ
Stijn Hinlopen
Vivraan
Meriipu
RevoluPowered
Adam Bates
Andy Maloney
Nathan Franke
Paulb23
Ignacio Etcheverry
Jummit