Description

I have a hardware abstraction layer interface, as well as a driver interface. I have multiple drivers utilizing the HAL interface, while implementing the driver interface. This plugin approach causes me to have identically named tests across multiple test files.

For example:

TEST_CASE("init() returns `E_HAL_INIT`, when HardwareAbstractionLayer::init() returns an er

Version of OpenTTD

9c5a7d3a572d02813e6bab83004334b9a154a62b

Expected result

The function IConsoleError behaves exactly the same as using IConsolePrintF(CC_ERROR, "ERROR: %s", string);, but is shorter and more concise.
https://github.com/OpenTTD/OpenTTD/blob/9c5a7d3a572d02813e6bab83004334b9a154a62b/src/console.cpp#L168-L171

Therefore, IConsoleError(string) should in general be

EDIT, from last post:
Latest non-DCH drivers (465.89) work fine for all generations (Pascal, Volta, Turing and Ampere): https://www.nvidia.com/download/driverResults.aspx/172060/en-us

No memory leak, decent speeds, complete compatibility.

DCH ONLY!
It looks like latest NVIDIA drivers (461.92 and 491.75, perhaps others) have massive memory leak bugs in NVML library. This memor