Features:
- Debug-only tracking of objects by type. See
get_driver_allocs_by_object_type et al.
- Debug-only Breadcrumb info for debugging GPU crashes and device lost
- Performance report per frame from get_perf_report
- Some VMA calls had to be modified in order to insert the necessary
memory callbacks
Functionality marked as "debug-only" is only available in debug or dev
builds.
Misc fixes:
- Early break optimization in RenderingDevice::uniform_set_create
============================
The work was performed by collaboration of TheForge and Google. I am
merely splitting it up into smaller PRs and cleaning it up.
Before, they would always complain even if there was no attempt at
initializing them (e.g. because there's no X11 display).
While we're at it, this patch also adds a specific message for OpenGL ES
and rewords "OpenGLES" to "OpenGL ES" in an error, for consistency
(AFAIK we either say "GLES" or "OpenGL ES").
Before, multiple capability events would instantiate the same object
over and over as long as its bit was set. This caused issues with
hotplug and device suspension.
On Windows this allows to avoid having to change the owner of the window
after it has been created, which in rare circumstances may cause the
window to bug out.
Allocated XImages are improperly free'd with XFree.
The X11 documentation says that XImage should use
XDestroyImage to free both the image structure and
the data pointed to by the image structure.
Also fix a potential use-after-free bug.
This reduces even further the amount of work we have to do when scaling
and potentially improves input accuracy as now the input code is free
from any form of rounding.
Before of this patch, as explained in the usual
commented-wall-of-text-longer-than-the-actual-patch-itself™, due to the
multithreaded nature of the Wayland thread, it was possible to commit a
surface while the renderer was doing stuff, which was _very_ wrong.
Initially the consequences of such a sin weren't obvious but, now that
explicit synchronization is becoming more and more common, we can't
commit a buffer randomly without basically guaranteeing a nasty, nasty
crash (and we should have avoided commits altogether in the first place
to ensure atomic surface updates).
We now only trigger a commit _in the main thread_ when low processor usage
mode is on _and_ if we know that we won't be rendering anything as, due to
its intermittent nature, it makes "legacy" (pre xdg_wm_base v6) frame
callback based suspension quite annoying.
- Make warnings print only once per session.
- Tweak the message to be less confusing, and mention that the issue
most likely stems from a graphics driver limitation.
Input events go to the tooltip because it's added to `popup_list` in
DisplayServer `popup_open`. I think there's no harm in tooltips being omitted
from the list, so this commit blocks non-popup windows from being added if they
have `FLAG_NO_FOCUS` and `FLAG_MOUSE_PASSTHROUGH`.
I'm not happy with this way of detecting tooltips. It'll also catch other
windows where this behavior may or may not be wanted.
I thought about adding `FLAG_TOOLTIP`, but went with the smaller change for
now.
Fixes#79500.
This avoids any assumption from the driver, which would otherwise select
a specific platform and potentially mess up everything, resulting
usually in a display server failure.
usleep(3) was declared obsolete in POSIX.1-2001 and removed in POSIX.1-2008.
nanosleep(2) was recommended to be used instead.
`OS::delay_usec()` internally uses `nanosleep()`.
This also uses large number separators for improved readability.
* Replaces `find(...) != -1` with `contains` for `String`
* Replaces `find(...) == -1` with `!contains` for `String`
* Replaces `find(...) != -1` with `has` for containers
* Replaces `find(...) == -1` with `!has` for containers
Mainly, this fixes auto UI scaling with _single-monitor_ fractional
setups (see the comment in `display_server_wayland.cpp` for more info).
This is the result of a bunch of current limitations, mainly the fact
that the UI scale is static (it's probed at startup) and the fact that
Wayland exposes fractional scales only at the window-level, by design.
The `screen_get_scale` special case should help in 99% of cases, while
the auto UI scale part will unfortunately only help with single-screen
situations, as multi-screen fractional scaling requires dynamic UI
scale changing.
Random-access access to `List` when iterating is `O(n^2)` (`O(n)` when
accessing a single element)
* Removed subscript operator, in favor of a more explicit `get`
* Added conversion from `Iterator` to `ConstIterator`
* Remade existing operations into other solutions when applicable
Previously we pretty much hardcoded most of the globals we requested,
causing compatibility issues with certain compositors like Weston, which
support only some pretty old versions or miss some more advanced
protocols.
To put fuel on the fire, we also errored out when certain protocols
weren't available, despite us being able to boot a game just fine (but
obviously with a degraded featureset).
The solution is to simply allow all the way from version 1 to the
current latest, adding some compatibility code (such as for older
`wl_output`s or newer `wl_pointer`s).
While we're at it, this commit also fixes a few typos and naming inconsistencies
I found.
The OS module get_unique_id and get_processor_name rely
on linux files which don't exist on a standard FreeBSD install,
make sysctl calls to get the required data.
Before, the cursor kept updating for no good reason really.
It's also a bit neater and it ever-so-slightly makes `WAYLAND_DEBUG`
logs easier to read, although they're still spammed by the window's
frame logic (which is needed).
- `Main::setup` early exits (failure or `--help`/`--version`) now
consistently return `EXIT_FAILURE` or `EXIT_SUCCESS` on all platforms,
instead of 255 on some and a Godot Error code on others.
- `Main::start` now returns the exit code, simplifying the handling of early
failures.
- `Main::iteration` needs to explicit set the exit code in OS if it errors
out.
- Web and iOS now properly return `OS::get_exit_code()` instead of 0.
This code was already partially there, although heavily incomplete and
nowadays commented out.
It got broken after the `WaylandThread` refactor and I didn't bother to
bring it over, preferring to `#if 0` it into oblivion for the time
being as I don't have a tablet/pen which support an eraser and tilt
reporting.
This commit brings it back and adds proper multi-tool support (needed
for eraser detection) thanks to winston-yallow, who could test this code
with their more capable tablet.
This is a pretty popular approach that took a while for me to wrap my
head around and which only recently got "official" support through an
update (xdg_shell version 6), so I think that this is all-in-all a
better option than the overkill 2000Hz ticking we have now :P
Basically, we wait for a frame event and, if either too much time passes
or we get the new `suspended` state, we consider the window as "hidden"
and stop drawing, ticking by the low usage rate.
This should work great for KDE and Mutter, which support the new state,
but not yet for sway, which is still stuck at a very old xdg_shell
version and thus falls back to the timeout approach.
Be aware that if we rely on timing out the engine will have to stall for
the whole timeout, which _could_ be problematic but doensn't seem like
it. Further testing is needed.
Special thanks go to the guys over at #wayland on OFTC, who very
patiently explained me this approach way too many times.
Instead of hardcoding platform names that support C#, let platforms
set a flag indicating if they support it. All public platforms
except web already support it, and it's a pain to maintain a patch
for this list just to add additional names of proprietary console
platforms.
This makes adding new platforms or variants or existing platforms
much easier, as the platform can signal what it supports/doesn't
support directly, and we can avoid harcoding platform names.
Using 2.2.7.dev115+g0eb441d6.
Had to add `cancelled` to the ignore list, as it's a Wayland signal which
we're handling in our code, so we don't want codespell to fix that "typo".
Also includes the typo fix from #87927.
Co-authored-by: Divyanshu Shekhar <61140213+divshekhar@users.noreply.github.com>
This prevents a wayland-scanner message from appearing every build
when `wayland=yes` is used (the default). The error message when
wayland-scanner is still printed as it's not printed by
wayland-scanner itself.
This is a workaround for the most critical portion of the WM focus bug
described in #68305. On some specific X11 WM configurations, the
editor's main window and any popups it creates will fight for focus,
which causes a total system lockup due to mouse and keyboard input being
stolen as well. Getting out of this infinite loop requires force
restarting the system.
It can be tested with the following shell script:
```bash
!#/bin/sh
godot4 &
sleep 30
pkill -x godot4
```
The workaround identified in #68305 is to remove the call to
XSetInputFocus in the ConfigureNotify event handler, so I have removed
the conditional block that calls this as well as the setup code above it
since there is no need to allocate the memory for the variables if they
won't be used in that call anymore.
This is just a hack and is not a complete fix for #68305. Multiple
developers are collaborating on a proper fix in the discussion in that
issue, but time is a valuable resource that no one has enough of, so I
am committing this workaround as a stop-gap to prevent the most critical
problem while we work on a full solution for the underlying cause.
This allows previous X11-only setups to still build Godot with default
settings. Note that compilation will still abort if wayland-scanner is
present but not the various Wayland libraries.
Not everything is yet implemented, either for Godot or personal
limitations (I don't have all hardware in the world). A brief list of
the most important issues follows:
- Single-window only: the `DisplayServer` API doesn't expose enough
information for properly creating XDG shell windows.
- Very dumb rendering loop: this is very complicated, just know that
the low consumption mode is forced to 2000 Hz and some clever hacks are
in place to overcome a specific Wayland limitation. This will be
improved to the extent possible both downstream and upstream.
- Features to implement yet: IME, touch input, native file dialog,
drawing tablet (commented out due to a refactor), screen recording.
- Mouse passthrough can't be implement through a poly API, we need a
rect-based one.
- The cursor doesn't yet support fractional scaling.
- Auto scale is rounded up when using fractional scaling as we don't
have a per-window scale query API (basically we need
`DisplayServer::window_get_scale`).
- Building with `x11=no wayland=yes opengl=yes openxr=yes` fails.
This also adds a new project property and editor setting for selecting the
default DisplayServer to start, to allow this backend to start first in
exported projects (X11 is still the default for now). The editor setting
always overrides the project setting.
Special thanks to Drew Devault, toger5, Sebastian Krzyszkowiak, Leandro
Benedet Garcia, Subhransu, Yury Zhuravlev and Mara Huldra.
This adds a new enum `KeyLocation` and associated property
`InputEventKey.location`, which indicates the left/right location of key
events which may come from one of two physical keys, eg. Shift, Ctrl.
It also adds simulation of missing Shift KEYUP events for Windows.
When multiple Shifts are held down at the same time, Windows natively
only sends a KEYUP for the last one to be released.
This change introduces a new EditorThemeManager class
to abstract theme generatio and its subroutines.
Logic related to EditorTheme, EditorColorMap, and editor
icons has been extracted into their respective files with
includes cleaned up.
All related files have been moved to a separate folder to
better scope them in the project. This includes relevant
generated files as well.
This intends to be the correct way to handle non-child windows becoming covered by the current window when becoming focused.
Enabling this property on select windows, they will become transient to the currently focused one when becoming visible.
This deprecates the "unparent_when_invisible" function introduced by #76025.
Rémi Verschelde
Rémi Verschelde
Matias N. Goldberg
Riteo
bruvzg
alvinhochun
Alvin Wong
Ronald Casili
Hugo Locurcio
Pedro J. Estébanez
David Snopek
Thaddeus Crews
Anni Ryynänen
kleonc
Joel Winarske
Paulo Poiati
Jakub Marcowski
A Thousand Ships
Jakub Marcowski
sambler
Roman Vasilyev
Alistair Leslie-Hughes
kobewi
LinuxUserGD
Muller-Castro
Andreia Gaita
Dario
Jakub Marcowski
Dalton Lang
Markus Sauermann
Mel Collins
Yuri Sizov
Juan Linietsky
mequam