This tool aids in the proper upkeep of copyright headers for changes contained within this repository. It will generate a new header, or replace the original one if one exists. As this task has often been forgotten by both developers and contributors, having a tool manage this will hopefully improve the situation.
The choice of Node.JS for this tool was deliberate, as many developers and CI solutions already have Node.JS in a reasonably up to date version installed. Additionally the versatility of Node.JS eliminates the need to create custom or platform specific solutions for tasks that are relatively simple. While the performance is not ideal, it still completes its task relatively quickly.
A full clean build is only required for releases, so opting for iterative builds will speed up CI builds significantly. As we can re-use caches from across branches, opting for the first matching key, this means that any branch should re-use the closest match from the main branch.
Downside is that this may break if there is a change to the CMakeLists.txt file, since the project configuration is now different. Might be fixable, but that's not the aim of these changes.
There's no need to split by configuration type on proper generators, which slightly speeds up the build. Downside is that we can't get builds for both Debug and RelWithDebInfo - though we only need the latter anyway.
This functionality broke at some point in the past without anyone noticing, resulting in most dropdowns that rely on this functionality being blank.
Fixes#1025
Many platforms (and/or kernels) don't zero memory before it is acquired, resulting in uninitialized memory being used to store critical content. This made the threadpool assume it had an infinite number of threads to work with, despite actually having spawned none.
Fixes#1017
Builds on Ubuntu incorrectly used gcc instead of g++ for C++ building, resulting in files that should never have worked. Switching this to g++ should drastically improve the stability of StreamFX on Linux.
Also on MacOS we use Homebrew to install the remaining tools, but failed to correctly install a specific version of it. As Homebrew only installs the latest version of a package to the global search path, this adjustment should fix the issue with clang-format.
Finally since I was already at it, I added clang-tidy integration and made the clang integration global. This way it does not need to be called on a per-project basis any more.
This should fix the case where comparing '0.12.0b164' with '0.12.0a169' results in it claiming the former is newer. Also should fix the case where stable releases would always be treated as older.
Fixes#1015
The rendering code here was one of the older code bases, which was riddled with bugs and leaks. The new code doesn't look pretty, but it works for the time being until a better solution is found or made. It should be able to handle HDR inputs now, but it may not be completely correct yet. This also fixes the double-free bug.
As an additional improvement, I've moved the allocation of the effect to a shared class which should reduce the memory usage slightly when multiple effects are in play. And now selecting nothing selects the filter target itself without infinitely adding references to the filter. Good enough in my eyes.
Fixes#819
* New translations en-US.ini (Danish)
* New translations en-US.ini (Spanish)
* New translations en-US.ini (Korean)
* New translations en-US.ini (Korean)
* New translations en-US.ini (Japanese)
While this had no actual effect due to the immediate ID3D11DeviceContext being tied to the ID3D11Device itself, it shouldn't have occured at all. With this there should now only be a single Release() call for every AddRef() call.
The location for System-wide and User-only installation depend on Windows-managed locations, which means that they will never require a user to select the correct location. Therefore it is best to hide the page in all modes except for Portable, which should still offer a path to extract to.
While the previous approach of a static thread pool worked, it was sub-optimal in its resource usage. Many of the threads would never see a single task, and simply permanently sleep. This seems like a good idea, except that sleeping threads still end up in the scheduler, and thus waste a tiny amount of resources.
It is better to instead dynamically spawn threads when needed and only keeping the bare minimum around all the time. These dynamically spawned threads are also explicitly set to background priority which further reduces scheduling overhead. Finally optimizing the memory layout to prevent unwanted false sharing should also keep sporadic wake ups at a minimum.
This new model should be able to handle many more tasks than ever before, but is still not as optimal as it could be.
