Bump Catch2 version due to GCC11.2 incompatibility

[pgrete]

PR Summary

I also stumbled across catchorg/Catch2#2178

[ 18%] Building CXX object tst/CMakeFiles/catch2_define.dir/catch2_define.cpp.o
In file included from /usr/include/signal.h:328,
                 from /home/pgrete/src/parthenon/external/Catch2/single_include/catch2/catch.hpp:7955,
                 from /home/pgrete/src/parthenon/tst/catch2_define.cpp:16:
/home/pgrete/src/parthenon/external/Catch2/single_include/catch2/catch.hpp:10735:58: error: call to non-‘constexpr’ function ‘long int sysconf(int)’
10735 |     static constexpr std::size_t sigStackSize = 32768 >= MINSIGSTKSZ ? 32768 : MINSIGSTKSZ;
      |                                                          ^~~~~~~~~~~
In file included from /usr/include/bits/sigstksz.h:24,
                 from /usr/include/signal.h:328,
                 from /home/pgrete/src/parthenon/external/Catch2/single_include/catch2/catch.hpp:7955,
                 from /home/pgrete/src/parthenon/tst/catch2_define.cpp:16:
/usr/include/unistd.h:640:17: note: ‘long int sysconf(int)’ declared here
  640 | extern long int sysconf (int __name) __THROW;
      |                 ^~~~~~~
In file included from /home/pgrete/src/parthenon/tst/catch2_define.cpp:16:
/home/pgrete/src/parthenon/external/Catch2/single_include/catch2/catch.hpp:10794:45: error: size of array ‘altStackMem’ is not an integral constant-expression
10794 |     char FatalConditionHandler::altStackMem[sigStackSize] = {};
      |                                             ^~~~~~~~~~~~
make[2]: *** [tst/CMakeFiles/catch2_define.dir/build.make:76: tst/CMakeFiles/catch2_define.dir/catch2_define.cpp.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:1510: tst/CMakeFiles/catch2_define.dir/all] Error 2
make: *** [Makefile:146: all] Error 2

Given this was fixed in upstream Catch2, this now includes the latest stable release.

PR Checklist

• Code passes cpplint
• New features are documented.
• Adds a test for any bugs fixed. Adds tests for new features.
• Code is formatted
• Changes are summarized in CHANGELOG.md
• CI has been triggered on Darwin for performance regression tests.
• (@lanl.gov employees) Update copyright on changed files

Still WIP because the discovery now gets confused with Kokkos info output, see first and last tests

$ ctest -N
Test project /home/pgrete/build-test
  Test  #1: Kokkos::OpenMP::initialize WARNING: OMP_PROC_BIND environment variable not set
  Test  #2:   In general, for best performance with OpenMP 4.0 or better set OMP_PROC_BIND=spread and OMP_PLACES=threads
  Test  #3:   For best performance with OpenMP 3.1 set OMP_PROC_BIND=true
  Test  #4:   For unit testing set OMP_PROC_BIND=false
  Test  #5: Adding MeshBlockData objects to a DataCollection
  Test  #6: Just check everything
  Test  #7: Task Object Lifecycle
  Test  #8: Can create a vector-valued face-variable
  Test  #9: Add, Get, and Update are called
  Test #10: reset is called
  Test #11: when hasKey is called
  Test #12: Physical constants
  Test #13: Checking IndexShape indices
  Test #14: Checking IndexShape cell counts
  Test #15: Sorting
  Test #16: par_for loops
  Test #17: nested par_for loops
  Test #18: Overlapping SpaceInstances
  Test #19: Device Object Allocation
  Test #20: Built-in flags are registered
  Test #21: A Metadata flag is allocated
  Test #22: A Metadata struct is created
  Test #23: ParArrayND
  Test #24: ParArrayND with LayoutLeft
  Test #25: ParArray resizing
  Test #26: Time simple stencil operations
  Test #27: Check registry pressure
  Test #28: Check many arrays
  Test #29: Can pull variables from containers based on Metadata
  Test #30: Coarse variable from meshblock_data for cell variable
  Test #31: Get the correct access pattern when using FlatIdx
  Test #32: MeshData works as expected for simple packs
  Test #33: Swarm memory management
  Test #34: Test required/desired checking from inputs
  Test #35: Parthenon Error Checking
  Test #36: Check that integer division, rounded up, works
  Test #37: Check that partitioning a container works
  Test #38: Test Add/Get in Packages_t
  Test #39: Test Associate in StateDescriptor
  Test #40: Test dependency resolution in StateDescriptor
  Test #41: Test SparsePool interface
  Test #42: Kokkos::OpenMP::initialize WARNING: OMP_PROC_BIND environment variable not set
  Test #43:   In general, for best performance with OpenMP 4.0 or better set OMP_PROC_BIND=spread and OMP_PLACES=threads
  Test #44:   For best performance with OpenMP 3.1 set OMP_PROC_BIND=true
  Test #45:   For unit testing set OMP_PROC_BIND=false
  Test #46: Catch2 Container Iterator Performance

[pgrete]

I now understand what's going on.
From https://github.com/catchorg/Catch2/blob/v2.x/docs/cmake-integration.md#catchcmake-and-catchaddtestscmake

Catch.cmake provides function catch_discover_tests to get tests from a target. This function works by running the resulting executable with --list-test-names-only flag, and then parsing the output to find all existing tests.

Therefore the Kokkos output is incorrectly parsed.
The fix I tried was to add EXTRA_ARGS "--kokkos-disable-warnings" to the discovery to suppress the Kokkos warning during parsing of the test names.
It didn't work as the argument is apparently not passed to the discovery call, see catchorg/Catch2#2382

[pgrete]

Alright, this is ready for review. I got around all the quirks of the other discovery mechanism by suppressing Kokkos warning during discovery and by manually adding ctest labels (as parsing catch tags as ctest _labels_is currently not supported -- though there's an open PR on Catch2 for that).