hdf5set_format_compatibility.c: Update new function documentation.

[Dave-Allured]

• Comment change only. Add one paragraph. Explain the reason for the third argument to H5Pset_libver_bounds. Simplifies experimental code using advanced features.

This updates the original function added in PR #2176.

[WardF]

I am a bit puzzled how this is failing the checks given the nature of the change. Testing by hand to try to shed some light on it.

[Dave-Allured]

I have been noticing random CI failures here and there, for several days. In the single reported failure above, in the CI logs, I saw two timeouts during test suite execution, and nothing else I think.

I think these are probably CI problems unrelated to this PR. You must admit that the simple change in this PR is unlikely to cause any code execution problems.

[Dave-Allured]

Um, make that three timeouts. It appears that test no. 87, nc_test_run_mmap, may not be reporting its results properly. The fail was in the log, but not in the fail summary at the end.

[WardF]

@Dave-Allured Yeah, I'll take a look at this tomorrow and make sure it's something in the back-end infrastructure and isn't the result of something on our end that I'll need to track down/correct.