ggggggggg

I think this is the nature of the HDF5 library, not of the Julia wrapper. Try the same in Python (I think h5py is supported by the HDF5 group directly)....

Can we close this as wontfix if I'm correct that this is a limitation of the HDF5 library?

Can you be more specific, eg do you want to do something like this? ```julia s3_bucket_with_hdf5_file_in_it = S3.openbucket() h5 = HDF5.h5open(s3_bucket_with_hdf5_file_in_it, "rw") h5["a"]=4 x = h5["b"] ```

There is a tiny bit about resizable datasets in `docs/hdf5.md`, search for `extendible`. I think defining `H5S_UNLIMITED=-1` will work.

When external links are traversed, they are always opened with `H5F_CLOSE_WEAK`. Somehow it seems that despite closing both `source_file` and `target_file`, a reference to `target_file` still exists with `H5F_CLOSE_WEAK`. Then...

Ok thanks. I'm going to have to do more than just add a `p` to my code.

I'm not sure this is possible? Repeated calls to `datatype("this is a string")` don't return the same value.

I first looked for `haskey`. But I posted about `has` because that works for the plain `HDF5File`. I'm happy to try to add this. It seems like `JLDFile` (and `HDF5File`)...

I wrote a version of `message` with only one copy, which is a bit faster than the previous one. Then I benchmarked all 3 version (my two, plus ZMQ.Message) vs...

`send_multipart` + `recv_multipart` are clearly matched, while `send` vs `recv_multipart` are not. Also this is how pyzmq does it, which is a small plus. Regardless of the name for `send`,...