Sean McBride

@bmribler not sure why you just tagged this as "approved"... it's a WIP and in fact it appears to be conflicted now...

OK, I removed the commit that was WIP, leaving only the 1 commit that modernizes the dtors. This is now ready for review & merge IMHO.

@bmribler @byrnHDF friendly ping...

I'm inclined to agree with @gnuoyd. It seems to me whatever compiler you are using needs to know intimate details of what it's compiling *for*, whether cross compiling or not....

> @seanm you meant for the copy assignment operators? No, those were moved to https://github.com/HDFGroup/hdf5/pull/3306 This PR is ready to merge IMHO.

>The old way doesn't need to prove anything, it already works correctly for all existing cases. Uh, no, the old way does ***not*** work for Universal Binary builds on macOS,...

There are a few PRs coming that should help with cross-compiling: https://github.com/HDFGroup/hdf5/pull/1400 https://github.com/HDFGroup/hdf5/pull/1426

I may have made a discovery... It seems like these settings: ```bash sysctl kern.sysv.shmmax=8388608 sysctl kern.sysv.shmseg=48 sysctl kern.sysv.shmall=98304 ``` Must be done ***before*** building AFL. They are not just important...

> I think we should make the build not require that much share mem.... I agree; it would probably also complicate building afl++ by something like homebrew. >Can you try...

> thank you! No, thank you! I'm happy to now actually be able to do some fuzzing of my code! I'll probably be back with more questions soon. :)