Reserve Console_Init() for actual console init, if code only needs colors (e.g. a console application that already has a console), then they should call Console_ColorInit().
The terminal window isn't always necessary, especially on Linux systems. In fact, it causes issues on some Wine environments. Allow user to disable it with -noconsole. This option only exists on the dedicated server. The client builds have the console disabled by default, and can enable them with -wconsole.
- Don't free console if the process is being closed from the console, this will cause the process to freeze within the FreeConsole() call
- Properly check for CPU features in order, and moved all checks to a single function utilizing the CPUInformation struct.
- Made SDK_Init() and SDK_Shutdown() more resilient against mistakes with new checks and error messages (added since they are exported now).
Moved implementation to dedicated class 'CImguiSystem' with the following changes:
* Added mutex for ImGui_ImplWin32_WndProcHandler(), the window proc is ran from a different thread than CImguiSystem::SampleFrame() is being called from. Needs a mutex as both functions interact with the input queue list.
* Added init stages for debugging and preventing the system from running certain parts of the code if a certain stage hasn't been reached yet.
* NULL out ImGuiContext::ConfigNavWindowingKeyNext , we don't need any nav window shortcuts as users don't need then in-game, and it also interferes with certain key binds some users have.
* CImguiSystem::SampleFrame() (previously ImguiSystem_SampleFrame()), is now being called from CMaterialSystem::SwapBuffers(), as that function is called from the main thread while also getting updated during level loads/etc, so instead of switching between 2 static buffers for as long as the levels are being loaded, we imgui updates properly too and directly swap afterwards.
* Removed hack in CGame::WindowProc() where we would only run windowproc if we aren't shutting the game down. This was needed as there was a chance of a crash at some rare cases, but this was simply due to a missing mutex which has been addressed and fixed as of this commit.
* Lock mutex during init and shutdown.
ImGui drawing code now takes place in the main thread, a snapshot of the render data is created in CMaterialSystem::SwapBuffers(), and is being rendered in the render thread right before SpinPresent().
The reason why this was necessary, is because ConVar::GetString() isn't thread safe if its not marked FCVAR_MATERIAL_SYSTEM_THREAD or FCVAR_ACCESSIBLE_FROM_THREADS, and we used it for the console suggestions window, which iterates over every ConVar, accessible from threads or not.
Fully implemented ConVar class so we could statically construct all SDK convars, this avoids a level of indirection, and allows for creating ConVar's everywhere in the project.
This patch also removed the settings tab of the ImGui server browser, as it has threading issues, while it technically never caused a crash yet, it has been removed as there was no point keeping it vs the work required to make it thread save (it only managed 2 convars which are perfectly manageable through cfg's or the in-game console).
Also temporarily disabled the creation of ConVar's in the mod system due to a memory leak, we would allocate and register a convar based on details parsed out of a mod file definition, but never unregister and free it.
Avoid heap memory allocation and a level of indirection. This allows the compiler to optimize the program even more. No logic has been changed in this patch.
Originally, we store the search results in a CMemory instance which we then assign to the actual function pointer. CMemory is just a pointer class; we can assign the results directly to the actual function pointer. This commit reduces a lot of code verbosity, and also reduced roughly 2KiB worth of static pointers in the resulting executable. This commit also officially deprecates the support for any GameDLL's below S3 (Season 3), since it makes more sense to port the assets from earlier/later games back to the version this SDK supports.
DrawImGui should be called BEFORE ID3D11DeviceContext::ClearRenderTargetView() or ID3D11DeviceContext::End(). Moved drawing code to SpinPresent which is just before the clear/end calls are being made. Console now draws in the 4:3/5:4 aspect ratio video modes.
Testing revealed better latency reduction when moved directly here as we also take into account the additional buffy copy's the game does in SpinPresent, our new render thread frame limiter and and ImGui. The difference is very large when the render thread frame limiter is used. The code still passes NVIDIA's reflex testing utility after this patch.
Utilize the new IDetour::DetourSetup() code, IDetour::Attach and IDetour::Detach have been removed in favor of this (significantly reduces chance of user error). Since the template check happens in the idetour header, it is much more aggressive on type mismatches, such as a difference in parameter types, between the function and detour, will now raise a compile time error. As a result, some type mismatches have been fixed in this commit as well.
Some parts of the engine have to be rebuild in order to implement this correctly, therefore, it has been removed for now to avoid potential performance problems.
fps_max_rt and fps_max_gfx have been limited to 295 to avoid a contest with the engine's hard limit causing huge performance hits.
Propagated latency markers in an attempt to further reduce input latency. Code is pending rigorous testing on several systems, including systems powered by an AMD GPU to make sure we are not impacting performance on systems that does NOT support NVIDIA Reflex.
Limits frames at a much higher level of precision than 'fps_max' and 'fps_max_gfx', probably ideal to reduce input latency even more. Also changed the logic of the NVIDIA Reflex frame limiter, to which it would use the desktop's refresh rate if set to '-1'. The new render thread frame limiter has a similar behavior. Using desktop refresh rates on the render thread or NVIDIA Reflex frame limiter requires 'fps_max' to be set to 0 (unlimited), as it would otherwise result in a major performance drop due to a contest if fps_max_(gfx/rt) is set to a similar number as fps_max.
This commit fixes an issue where the console window would shrink with 1 unit when calling 'SetConsoleBackgroundColor'. The bottom should be incremented by 1 each time its getting called, as it appears to be decremented again in WINAPI 'SetConsoleScreenBufferInfoEx'.
Find regex pattern:
inline auto ([a-zA-Z0-9_]+) = ([a-zA-Z0-9_]+)\.RCast<([a-zA-Z0-9_:<>*]+) *\(\*\)\(([^)]*)\)>\(\);
Replace regex pattern:
inline $3(*$1)($4);
This commit also removes the unnecessary initialization (which was required to type the auto variables),
and therefore removed 6kb of unnecessary dynamic initialization code.
Before, we had to do a hack of capturing the command line using GetCommandLineA, and then checking if a certain argument is present. This was required due to how early the GameSDK dll was loaded (the g_CmdLine object was far from initialized in the engine). Due to the loader refactor, the commandline can be used directly after creation in the game's entry point (which is the time the SDK is getting loaded). Therefore, no copies of the command line are required anymore.
This commit contains the following changes:
- Correctly ordered the initialization, and deinitialization of systems (first init = last shutdown).
- Factored out command line string copy in favor of game's implementation.
- Factored the R5Reloaded emblem print into its own function.
- Removed 'SpdLog_PostInit()', we can now directly call DevMsg() once SpdLog_Init() has been called, the logger callback sink deals with the formatting of the output.
- Fixed a bug where the logger did not print the correct color for 'SYSTEM_WARNING' and 'SYSTEM_ERROR' in the external console.
- Fixed a bug where the command line did not work when the game wasn't launched with the '-launcher' parameter.
- Logs now equally appear on the external, and in-game console windows.
When and after 'Host_Shutdown()' is called, the window handle is still valid, thus the handler still gets called. This gets destroyed in 'OnShutdown()' called from 'CEngineAPI::RunListenServer()'. But we do shutdown the ImGui implementation in 'Host_Shutdown()'... Therefore, we should check if the implementation is initialized or not before running the window procedure handler.
The 'DirectX_Init()' call was performed late in code, shortly after the window has been created (at this point all device objects and window handles are valid), but the 'DirectX_Shutdown()' call was performed on DLL_DETACH, which was way too late, as the objects were already destroyed at this point. This wasn't an issue before, as we created our own objects in the old DX code. But due to optimizations, we were using the same pointers as the game (noticeable performance boost), but did not adjust the shutdown to accommodate the changes. The shutdown is now performed while the device objects and window handles are valid. Code has been tested on Nvidia and AMD systems, and has confirmed to fix the aforementioned issues.
The shutdown patch from commit '48c2401c' created another bug where all inputs get collated, and once drawn, emitted to the ImGui interfaces. The patch has been undone, and the 'ImGui_ImplDX11_Shutdown()' call has been placed before 'ImGui_ImplWin32_Shutdown()', as this was how it was performed according to the official documentation and examples provided by Dear ImGui. The call 'ImGui::DestroyContext()' has also been added (taken from the examples). Removed redundant static global bool.
Fix a rare crash that occurs in AMD driver code, when ImGui shutdown was called. The crash did not occur if the library was shutdown after having rendered one of the ImGui panels for one frame. The fix is to just never call 'ImGui_ImplDX11_NewFrame()', 'ImGui_ImplWin32_NewFrame()' and 'ImGui::NewFrame()', 'ImGui::EndFrame()', 'ImGui::Render()' if none of the windows are visible. code has been tested on a system that would trigger the crash, and after the patch, the crash no longer happened.
This commit significantly reduces output code size, and a noticeable increase in performance. Changes are:
- Removed all extraneous std::string and std::vector copy constructions; use raw pointers instead to boost performance.
- Marked simple getters in CModule inline.
- Marked several functions in CModule const.
- Slightly reordered CModule class.
- 'CMemory::CheckOpCodes' and 'CMemory::Patch' now take a const reference.
Moved, and renamed to 'CreateTextureResource'. Reason for move was that the rtech libraries is used by server and client, and using this on dedicated requires linking directx libraries, as it has to be hooked (even when not used). Moved to client only code to avoid having to hook it. Material system is no longer linked to the dedicated server module, as nothing from it was getting used.
* All libraries have been isolated from each other, and build into separate artifacts.
* Project has been restructured to support isolating libraries.
* CCrashHandler now calls a callback on crash (setup from core/dllmain.cpp, this can be setup in any way for any project. This callback is getting called when the apllication crashes. Useful for flushing buffers before closing handles to logging files for example).
* Tier0 'CoreMsgV' function now calls a callback sink, which could be set by the user (currently setup to the SDK's internal logger in core/dllmain.cpp).
TODO:
* Add a batch file to autogenerate all projects.
* Add support for dedicated server.
* Add support for client dll.
Bugs:
* Game crashes on the title screen after the UI script compiler has finished (root cause unknown).
* Curl error messages are getting logged twice for the dedicated server due to the removal of all "DEDICATED" preprocessor directives to support isolating projects. This has to be fixed properly!
RCON upgrade with additional logging system improvements:
* Netconsole's can now log received messages in color, even when the RCON server has ANSI colors disabled; logs are fully composed locally.
* RCON server now also sends the log type over the wire, along with the (already existing) context.
* SDK logging code is now shared with the standalone netconsole application.
* Improved logging readability for the standalone netconsole application.
Initialize console and spdlog before anything else that is using the console to log errors to (in this case, winsock init errors were never getting logged).
