Update pytest framework

[wpbonelli]

Pytest framework changes include:

• use pytest-benchmark's builtin profiling capability instead of manual implementation
• update DEVELOPER.md to reflect above
• remove requires_exe("mf6") from test_mf6.py tests that don't run models/simulations
• add @requires_spatial_reference marker to conftest.py (for tests depending on spatialreference.org)
• split test_export.py::test_polygon_from_ij into network-bound and non-network-bound cases (and use marker above)
• try both importlib.import_module and pkg_resources.get_distribution in @requires_pkg marker
  • test: refine pytest skip framework, improve confest.py #1486 only used latter, which caused some tests to be unintentionally skipped due to differences between package name and import name, e.g. pymake vs mfpymake
• mark test_lgr.py::test_simple_lgr_model_from_scratch as flaky (occasional forrtl error (65): floating invalid)
• add comments to flaky tests with links to potentially similar issues
• remove unneeded markers from pytest.ini (profile, benchmark)
• match profiling/benchmarking test files in pytest.ini
  • profile_*.py
  • benchmark_*.py
  • *_profile*.py
  • *_benchmark*.py
• mark get-modflow script tests as flaky again due to internal github API egress issues
• various tidying/cleanup

CI changes include:

• add timeouts to CI jobs (10min for build, lint, & smoke, 45min for test, 90min for daily jobs)
• fix benchmark artifact names (previously Windows benchmark artifacts were saved as benchmark--<python version> because matrix.os was incorrectly used with no job matrix, now fixed with runner.os so that all benchmark artifacts are named as benchmark-<platform>-<python version>)
• cache benchmark results in daily CI and compare with prior runs (a performance regression failure condition is introduced such that if any benchmark sees an avg increase > 25%, the benchmark job is failed)