post_process_shard_results needs to be more robust

[rorth]

When running a LLVM 18.1.0 build on FreeBSD 14.0/amd64 with test-release.sh, llvm-lit didn't properly summarize the test results. The build log shows

[...]
PASS: llvm-libc++-shared.cfg.in :: std/utilities/variant/variant.visit/visit_return_type.pass.cpp (97112 of 97112)
Traceback (most recent call last):
  File "/var/llvm/reltest/llvm-18.1.0/final/llvm-project/llvm/utils/lit/lit/formats/googletest.py", line 304, in post_process_shard_results
    testsuites = json.load(f)["testsuites"]
  File "/usr/local/lib/python3.9/json/__init__.py", line 293, in load
    return loads(fp.read(),
  File "/usr/local/lib/python3.9/json/__init__.py", line 346, in loads
    return _default_decoder.decode(s)
  File "/usr/local/lib/python3.9/json/decoder.py", line 337, in decode
    obj, end = self.raw_decode(s, idx=_w(s, 0).end())
  File "/usr/local/lib/python3.9/json/decoder.py", line 355, in raw_decode
    raise JSONDecodeError("Expecting value", s, err.value) from None
json.decoder.JSONDecodeError: Expecting value: line 1 column 1 (char 0)

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/var/llvm/reltest/llvm-18.1.0/final/Phase3/Release/llvmCore-18.1.0-final.obj/./bin/llvm-lit", line 65, in <module>
    main(builtin_parameters)
  File "/var/llvm/reltest/llvm-18.1.0/final/llvm-project/llvm/utils/lit/lit/main.py", line 129, in main
    selected_tests, discovered_tests = GoogleTest.post_process_shard_results(
  File "/var/llvm/reltest/llvm-18.1.0/final/llvm-project/llvm/utils/lit/lit/formats/googletest.py", line 306, in post_process_shard_results
    raise RuntimeError(
RuntimeError: Failed to parse json file: /var/llvm/reltest/llvm-18.1.0/final/Phase3/Release/llvmCore-18.1.0-final.obj/runtimes/runtimes-bins/compiler-rt/lib/msan/tests/./Msan-x86_64-Test-MemorySanitizer-Unit-37844-0-9.json

FAILED: CMakeFiles/check-all /var/llvm/reltest/llvm-18.1.0/final/Phase3/Release/llvmCore-18.1.0-final.obj/CMakeFiles/check-all

I could trace this to two msan tests producing empty *.json files (Issue #84482), which googletest.py chokes on. This must not happen: llvm-lit needs to be robust enough to cope with such input, invalid or no.