Add PCH support and 'CATCH_ENABLE_PCH' flag

[vittorioromeo]

This PR enables precompiled headers via the CATCH_ENABLE_PCH CMake option, aiming to significantly speed up compilation times.

Description

These are some benchmark results on my machine with the following specs:

• Intel Core i9-13900K
• 32GB DDR5-6400 RAM
• EVO Plus 900 NVMe SSD

#
#
# given this CMake command...
cmake .. -DCMAKE_BUILD_TYPE=Debug -GNinja 
         -DCMAKE_EXPORT_COMPILE_COMMANDS=1 -DCMAKE_CXX_COMPILER=clang++ 
         -DCMAKE_CXX_FLAGS="-ftime-trace -Wall -Wextra -Wpedantic -Wno-ignored-attributes" 
         -DCATCH_DEVELOPMENT_BUILD=1 -DCATCH_BUILD_EXAMPLES=1 -DCATCH_BUILD_EXTRA_TESTS=1 -DCATCH_BUILD_FUZZERS=0

#
#
# ...with PCH disabled:
hyperfine "ninja clean && ninja" -n5 --warmup 1

Benchmark 1: 5
  Time (mean ± σ):     14.559 s ±  0.220 s    [User: 245.178 s, System: 33.759 s]
  Range (min … max):   14.362 s … 15.001 s    10 runs

#
#
# ...with PCH enabled:
hyperfine "ninja clean && ninja" -n5 --warmup 1

Benchmark 1: 5
  Time (mean ± σ):      9.718 s ±  0.165 s    [User: 104.442 s, System: 30.406 s]
  Range (min … max):    9.584 s … 10.152 s    10 runs

I get a ~5.0s speed-up for a full recompilation of Catch. I expect the speed-up to be more significant for machines with less cores.

Let's see if CI is happy with these changes :)