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.
Moved adapter vendor check to its own func, added a global which could be used to fully disable the low latency system. Also added groundwork for the (future) PCL stats implementation.
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.
Hook 'CMaterialSystem::Init()', and only load the 'startup.rpak' file, as that's all we need. This pak file might also not even be needed on the server, but removing this requires rebuilding 'common_early.rpak' to drop all texture/material assets and their references. The dedicated server is now also linked to a newly added 'No-DirectX' version of the materialsystem for patching.
This fixes the exceptions that were caused by loading ported bsps with replaced models
CStaticProp::Init was using static prop lump data to request non-existent skins due to the replaced models, causing invalid material pointers to be fetched.
When encountered, these skins will be forced to 0 and an error will be printed.
Ideally these issues are fixed from within the bsp file before they are shipped.
This change was planned for a long time. This moves all REGISTER calls to a single translation unit, this is required as we currently added a very dirty workaround for not registering duplicates by checking if VFTable pointer was already present in the vector... Registering from single translation unit prevents duplicate instances that gets created if header is included by more cpp files.
Reworking this reduced 100kb+ of compiled code. This commit also reworked the way functions/variables/constant gets logged with their addresses; the new code formats them on the fly, and allows for resize at any time. Formatting is no longer required by programmer.
TODO: currently there are some compile errors for dedicated and client dll's. These will be resolved very soon as they need to be properly worked out still (server & client only stuff needs to be properly split). Use the 'main' (stable) branch for the time being if you need to compile these dll's.
This is required as we still have a very hacky approach for running 'custom' BSP files. Even though we are pretty much ready to move over the proper approach, we haven't been able to update all files yet. This is kept in place but will have to be deprecated and removed among with the MaterialSystem hack before the next release.
Previously we used Json, however the rest of the game features KeyValues. I changed the system to feature KeyValues instead, to maintain consistency. Also improved the logic behind loading/parsing so we don't reparse the same level settings file if the pointer is still valid and we are on the same level.
The behavior has been changed; when a material is missing and bComplain is false, nothing gets printed to the console unless 'mat_alwaysComplain' is nonzero.
This commit implements missing material error logging.
When a material does not exist, it will get replaced with an error material, we log which material is missing, and with which error material it gets replaced with.
This commit also comes with some updates regarding the CMaterialGlue class, and some new types used by this class from the RePak project.
CMaterial and CMaterialGlue derive from the same abstract classes. The function that builds the culling map uses IMaterialInternal, which is used by CMaterial, CMaterialInternal and CMaterialSubRect. Added a check for CMaterialGlue as these are valid too.
Some CMaterial instances are allocated on the heap, and we only check for bounds in .data/.pdata. Checking bounds is cheap, checking valid ptr's is not. However we are lucky most of the data is within the segments, so running this isn't too costly.
This fixes the problem where the large leaf model in worlds edge was missing.
* Fix all compiler error for GameSDK on S0 and S1.
* Remove some unused patterns and globals (launch performance gains of 100ms).
* Remove most duplicate patterns.
* Relocate globals from engine to a more appropriate location.
* Renamed some unknown vars to a more appropriate name (most of these where obtained a while back, but as we reversed more parts of the engine, some of these vars become known).
* Renamed some vars to fit the naming convention used in the SDK.
* Fixed pattern debug logging using the wrong pointer, thus displaying the wrong address.
* Use c++ methods as much as possible.
* Use enum types for accessing NavMesh objects from array.
* Use size_t for for loops when testing against size types.
* Don't compute strlen twice of more on the same string.
* Don't use unnecessary c string casts if there is a method with a std::string overload.
* Don't create string objects from string pointers if we could use them directly.
* Don't initialize RCON password twice on each change, and don't set if the new password equals the old.
Moved logging functions to dbg.h (tier0) and export them from the dll.
Added additional functions for checking bad pointers (debug only!).
Reduced output code size.
This crash happens when the BSP is missing a lot of shaders and materials.
The call originates from 'R_DrawWorldMeshesDepthOnly()', but it is unclear which missing shader is causing the nullptr. checking for a nullptr here should be sufficient to deal with this kind of missing shader
