Ctrl + arrow should move to word end, not next word. This also fixes a bug where it would index into into the second character when the line end gets reached.
This code originated from an imgui code editor, where it would make sense to have per-character colors. But for the console, we only ever do 1 color per line. Just store the string and one color instance, this also allows us to just take the string instead of manually reconstructing it which is very slow. The code now also only stores 32bits for color (per line) instead of 128 bits (ImVec4) as we only ever need to convert it once.
As a result, a bunch of rendering code could be dropped, and the pass filter is now also very fast as we don't need to reconstruct the string to check for the filter, we just pass a char* in which doesn't cost any perf.
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.
The KeyValues class belongs here. Also reimplemented most loading methods for KeyValues, and adjusted the VPK building code to account for it. Pointers to the engine's implementation of KeyValues have been moved to a separate header ('keyvalues_iface.h'), as this allows external tools code to utilize the standalone KeyValues class implementation. Playlist utilities are completely separated from the KeyValues header; these have nothing to do with KeyValues other than manipulating a global KeyValues object for the playlists, and thus have been named as such and moved to rtech/playlists.
Merge recastnavitagion/recastnavitagion@c393777d26
Instead of using a distance check which can fail at large magnitudes due
to low precision we can check whether the edges are actually on the
hull.
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.
This will deprecate IDetour::Attach and IDetour::Detatch. This should reduce the chances of potential errors caused by using the wrong function in the wrong method (e.g. using DetourAttach in IDetour::Detatch).
In the engine, m_query gets memset to size '0x58', and it passes this field as reference to other functions which then access the 3 pointers. Moving the 3 pointers there made the struct size 0x58, and showed a much better disassembly. Moved these type here to reflect the changes made in the engine.
Enabling LTCG caused the emplacement order to be different, which is most likely due to additional optimizations for the binary search logic. Reverted to fix this bug.
- Use std::map for mapping vtables to interface objects, previously done with a set and a vector.
- Objects are no longer inline, which significantly reduced output code size as all redundant dynamic initializers (created for each translation unit) have been pruned.
The game's memalloc does not support memory expansion, the base expansion function couldn't be shadowed properly without linker errors. However we want to make sure no instances of it get compiled into the module, ever. Therefore, the duplicate symbol is kept in memoverride (_expand_base), and the only usage has been removed from LZHAM's 'lzham_mem'. If a linkage is attempted to anything utilizing '_expand', a linker error will be emitted and no image will be generated.
- Disable function inlining entirely, this allows for much easier hooking between engine and SDK code (the compiled code is now identical between engine and SDK, except that the SDK has all relevant security problems patched as per https://curl.se/docs/vuln-7.54.0.html).
- Enable buffer security checks to avoid potential remote attacks.
Previously, it was all controlled from the global init (applied to all projects), but some projects need different options. With these changes, you can disable the common options applied in the 'add_module' macro, and set your own if desired.
