Scot Breitenfeld

Thanks, I'll ask Jordan and Neil. I've not seen that segmentation fault before. Though I've only run it on a local desktop.

BTW, even if I add an ESwait after the last exists, it still fails.

Yes, HDF5_ASYNC_EXE_FCLOSE fixes the issue.

It passes most of the time, but running it over and over, I can sometimes get it to fail with: ``` async_test: ../../src/H5Fint.c:631: H5F__get_objects_cb: Assertion `obj_ptr' failed. Program received signal...

That was my mistake. It got removed in my editing of the run script. Using that, all the test pass.

The documentation changes should note this is for an autotools build. For CMake, the logic in HDFMacros.cmake handles this correctly. https://github.com/LLNL/H5Z-ZFP/blob/491e69f1bbf59e1da1dac689d4fc78beab780ee0/cmake/HDFMacros.cmake#L50C1-L84C1

Did you do this for your application? https://h5z-zfp.readthedocs.io/en/latest/installation.html#including-h5z-zfp-filter-in-a-cmake-project

Would you like H5Z-ZFP added to our CI at HDF5? It would be similar to our current testing of CGNS with HDF5 1.8, 1.10, 1.12 on various platforms and compilers.

How it works now with CGNS is we have a script that downloads the latest version of CGNS (or older versions). It then uses the builds from Daily testing for...

It was issue #61, H5MM_xfree/free https://github.com/LLNL/H5Z-ZFP/issues/61