Turns out that for Qt to properly handle plurals in English a
translation needs to be provided, otherwise the user is left with
messages such as "Building: 2 shader(s)"
Plurals for other all other languages are handled on transifex.
I wrote the README.md to just refer to it as a translation
collaboration site just in case we ever switch.
These translations being out of date won't pose any technical problems
so I believe it is fine to handle them manually on a "best effort"
basis.
The files are generated into the source directory so that the
relative filenames are correct. The generated file is added to
.gitignore
[REUSE] is a specification that aims at making file copyright
information consistent, so that it can be both human and machine
readable. It basically requires that all files have a header containing
copyright and licensing information. When this isn't possible, like
when dealing with binary assets, generated files or embedded third-party
dependencies, it is permitted to insert copyright information in the
`.reuse/dep5` file.
Oh, and it also requires that all the licenses used in the project are
present in the `LICENSES` folder, that's why the diff is so huge.
This can be done automatically with `reuse download --all`.
The `reuse` tool also contains a handy subcommand that analyzes the
project and tells whether or not the project is (still) compliant,
`reuse lint`.
Following REUSE has a few advantages over the current approach:
- Copyright information is easy to access for users / downstream
- Files like `dist/license.md` do not need to exist anymore, as
`.reuse/dep5` is used instead
- `reuse lint` makes it easy to ensure that copyright information of
files like binary assets / images is always accurate and up to date
To add copyright information of files that didn't have it I looked up
who committed what and when, for each file. As yuzu contributors do not
have to sign a CLA or similar I couldn't assume that copyright ownership
was of the "yuzu Emulator Project", so I used the name and/or email of
the commit author instead.
[REUSE]: https://reuse.software
Follow-up to 01cf05bc75
There was a bug where, when using the numeric keyboard, moving between buttons resulted in an infinite loop, resulting in a stuck state.
This was due to prev_button being the only one enabled in that row or column, causing the condition in the while loop to always be true.
To fix this, detect whether we have returned to that initial row/column and break out of the loop.
Between packages breaking, Conan always being a moving target for
minimum required CMake support, and now their moves to Conan 2.0 causing
existing packages to break, I suppose this was a long time coming. vcpkg
isn't without its drawbacks, but at the moment it seems easier on the
project to use for external packages.
Mostly removes the logic for Conan from the root CMakeLists file,
leaving basic find_package()'s in its place. Sets only the
find_package()'s that require CONFIG mode as necessary. clang and linux
CI now use the vcpkg toolchain file configured in the Docker container
when possible.
mingw CI turns off YUZU_TESTS because there's no way on the container to
run Windows executables on a Linux host anyway, and it's not easy to get
Catch2 there.
- Avoids new GCC 12 warnings when Type is of form std::optional<T>
- Makes more sense this way, because ranged is not a property which would change over time
Button inputs were broken as button was assumed to be the bit position of NpadButton prior to the input rewrite. Since this was changed to use NpadButton directly, we should count the number of trailing zeros to determine the bit position.
Currently to access the SP register, RegRead and RegWrite rely on a
out-of-bounds array access to reach the next element in a struct. As
of writing only git versions of GCC catch this error.
Specify the SP register when we want to access it in these functions.
The latest git version of GCC has issues with my diamond inheritance
shenanigans. Since that's now two compilers that don't like it I thought
it'd be best to just axe all of it and just have the two templates like
before.
This rolls the features of BasicRangedSetting into BasicSetting, and
likewise RangedSetting into Setting. It also renames them from
BasicSetting and Setting to Setting and SwitchableSetting respectively.
Now longer name corresponds to more complex thing.
In testing future versions of Qt I forgot to compile with `YUZU_USE_QT_WEB_ENGINE`, so with that flag enabled there are two issues that cropped up.
1. yuzu currently uses setRequestInterceptor, added in Qt 5.6, deprecated in 5.13 with this explaination at https://doc.qt.io/qt-5/qwebengineprofile-obsolete.html
Interceptors installed with this method will call QWebEngineUrlRequestInterceptor::interceptRequest on the I/O thread. Therefore the user has to provide thread-safe interaction with the other user classes. For a duration of this call ui thread is blocked. Use setUrlRequestInterceptor instead.
2. QWebEngineSettings::globalSettings() pointer no longer exists in later versions of Qt
From what I can tell, QtNXWebEngineView doesn't need to set these globally,
when we make changes to settings(), QtWebEngineView::page() creates the page
object if it doesn't exist yet. I don't see the page object being destroyed
or otherwise replaced, except via destroying the QtNXWebEngineView object.
The globalSettings() make sense if Pages or Views objects are being
created outside of yuzu's control.
To test this I've compared what BrowseNX and Odyssey's Action guide do in mainline 1049 and this PR.
For now we're going to go up the chain to QWebEngineProfile::defaultProfile()->settings()
Because logging infrastructure initializes before the loading of the
config, it reads the default setting for log_filter and ignores the one
set in config. To change log_filter after logging initialization some
additional calls need to be made.
Currently this just stops all the emulation
This works under assumption that only application will try to use
ExitProcess, with services not touching it
If application exits - it quite makes sense to end the emulation
While this is the primary change, we also:
- Remove the mpsc namespace and rename Queue to MPSCQueue
- Make Slot a private struct within MPSCQueue
- Remove the AlignedAllocator template argument, as we use std::allocator
- Replace instances of mask + 1 with capacity, and mask + 2 with capacity + 1
According to the standard, a narrowing conversion is an implicit conversion from an integer or unscoped enumeration type to an integer type that cannot represent all the values of the original type, except when the value is a literal or constant expression.
MSVC, unlike GCC or Clang, determines this to be a narrowing conversion despite the enumeration exclusively containing values that fit within the range of a 32 bit integer, emitting a warning since designated initializers prohibit narrowing conversions.
To solve this, explicitly cast to the type we are initializing.
This seems to be unsupported in newer libstdc++ versions due to
Flow::Block's base class being a non-literal type. It's not clear to me
why this was permitted in earlier versions.
Removing this as we don't enforce unused parameter warnings elsewhere in the project, and explicitly specify -Wno-unused-parameter in the main CMakeLists.
Now that the entire project is free of variable shadowing, we can enforce this as a compile time error to prevent any further introduction of this logic bug.
GCC/Clang treats variables within lambdas as potentially shadowing those outside the lambda, despite them not being captured inside the lambda's capture list.
This is the "Double-click to add a new folder to the game list" message
that shows up when users first launch yuzu and is most likely never seen
again. Previously this message was not re-translated.
Idea works as follows, while going fullscreen we compare the current window geometry with
available screens and ask for an intersection rectangle, we go fullscreen where most of
the window is located
GuessCurrentScreen could also potentially be used to see which screen
the window is on for dynamic DPI handling
For people not used to the Yuzu UI it's not always clear if the emulated
console is docked or not. The other items update their text when clicked,
this PR brings the DOCK button in line with this.
DOCK -> DOCKED or HANDHELD
Sometimes when yuzu crashes, it restarts with the games list in fullscreen,
which would be fine, except there isn't an easy way to exit this.
It also doesn't occur often enough for qt-config.ini files to be in good supply.
UILayout\geometry value in qt-config.ini is the culprit,
at least for the one provided.
Proposed fix is to simply check isFullScreen when yuzu is starting up,
and take it out of full screen immediately
This does a few things in order to make the default setting Vulkan
workable.
- When yuzu boots, it just opens the Vulkan library.
- If it works, all good and we continue with Vulkan as the default.
- If something breaks, a new file in the config directory will be left
behind (this is deleted normally).
- If Vulkan is not working, has_broken_vulkan is set to true.
- The first time this happens, a warning is displayed to notify the
user.
- This forces use of OpenGL, and Vulkan cannot be selected.
- The Shader Backend selector is made accessible for use in custom
configurations.
- To disable has_broken_vulkan, the user needs to press a button in
Graphics Configuration to manually run the Vulkan device
enumeration.
Two reasons for this:
1. Out of 7 connections, 6 are in ConfigureMotionTouch::ConnectEvents,
this is the outlier.
2. Qt6 doesn't moc the connection properly
There was some discussion about updating to Qt6 and I figured I would
work on some smaller parts. For Windows platform the WinMain function has moved
from the Qt5::WinMain to a new one called Qt6::EntryPointPrivate
Also Qt5 supports versionless CMake targets
https://www.qt.io/blog/versionless-cmake-targets-qt-5.15
These other changes in this commit are to support Qt6, but in ways that don't mess with Qt5.
src/yuzu/bootmanager.cpp: Qt6 complains about not being able to know to use QPoint or QPointF, picking QPoint
src/yuzu/bootmanager.h: Qt6 prefers that QStringList.h be included rather than an empty class definition
src/yuzu/configuration/configure_system.cpp: toULongLong intends to return unsigned 64 bit integer, but
Settings::values.rng_seed is only 32 bits wide
src/yuzu/game_list.cpp: Qt6 returns a different datatype for QStringList.length than Qt5,
it used to be int, but in Qt6 its now qsizetype
src/yuzu/loading_screen.cpp: Qt5's for QStyleOption.init say to switch to initFrom.
The QStyleOption.init doesn't exist in Qt6
src/yuzu/main.cpp: Another QPointer and QStringList.size, lets standardize on size()
Qt5 and Qt6 don't really do a good job of reporting Windows versions past the 2004 version.
Current: Windows 10 Version 2009
This Patch: Windows 10 Version 21H1 (Build 19043.1706)
Also: Windows 11 Version 21H2 (Build 22000.675)
Fixes: #8362
-mwindows doesn't work with Clang. tpoechtrager/wclang resolves this by
just using MinGW-GCC to link the executable, however this prevents us
from using LLVM-exclusive tools when building yuzu.
Solution is to send the linker argument we need from -mwindows directly
to the linker.
From https://gcc-help.gcc.gnu.narkive.com/FogklN5J/gcc-wl-subsystem-windows-mwindows-options
Clang (rightfully) warns that we are checking for the existence of
pointer to something just allocated on the stack, which is always true.
Instead, check whether GetModuleFileNameW failed.
Co-authored-by: Mai M <mathew1800@gmail.com>
A bug occurs in yuzu when VK_KHR_workgroup_memory_explicit_layout is
available but 16-bit integers are not supported in the host driver.
Disable usage of the extension when this case arises.
Recent AMD Vulkan drivers (22.5.2 or 2.0.226 for specifically Vulkan)
have a broken VK_KHR_push_descriptor implementation that causes a crash
in yuzu. Disable it for the time being.
