Neil Fortner

We will make sure to document this limitation

This doesn't seem to be the right place to put this check. Generally file consistency checks should be done at the point where the values are read from the file....

Assertions at read time would be fine but it's probably not worth passing around a bunch of extra data to do it.

The bug seems to be fixed in versions of HDF5 newer than 1.10.7. Why is this change necessary? Is there still a case where we could hit this problem? Is...

We've decided not to go this route, at least for 1.14. I'll leave this up for a bit longer as a draft incase anyone wants to weigh in on the...

Unfortunately I think fixing this would require a file format change, since the width of the encoded chunk size isn't stored in the file, but calculated as (in the filtered...

I think the plan is to modify the file format to always encode chunk length using 64 bits in the next major release.

I'm still unsure about making the H5FD API calls reentrant-only. Is this change needed to fix something or is it only to remove the overhead associated with the init check...

> > I'm still unsure about making the H5FD API calls reentrant-only. Is this change needed to fix something or is it only to remove the overhead associated with the...

> Following up on our discussion at the community meeting earlier today, I reviewed the difference between the FUNC___API and FUNC___API_REENTER macro pairs. It's not just the API context that...