Neil Fortner

MPI_File_read_at_all doesn't read all values when certain datatypes are used

10

## Background information While testing a new feature in HDF5, we noticed that OpenMPI sometimes fails in a test that generates random I/O patterns in HDF5. I've tried to create...

Remove file "post open" VOL callback

1

Also add "get_wrap_ctx_pre_open" VOL callback and add "wrap_ctx" parameter to file create and file open VOL callbacks to allow this to work. The main aim of these changes is to...

For 2.0, we should consider changing the default library version bounds (H5Fset_libver_bounds) so the low bound is equal to the 2.0 library version. This will allow users that are trying...

For version 1 b-trees, knowing the "k" value (related to the node size) is necessary to parse the file format. In some circumstances, this k value is not stored in...

H5Dread_chunk() currently has the signature: `herr_t H5Dread_chunk(hid_t dset_id, hid_t dxpl_id, const hsize_t *offset, uint32_t *filters, void *buf);` This is potentially not memory safe because hte caller does not specify the...

The dataset chunk code uses the configurable "size of sizes" set by H5Pset_sizes for encoding various quantities. However, we do not currently test chunekd datasets with any value other than...

We should test using a filter to expand a chunk from under 4 GiB to 4 GiB or more. This will require the ability to turn the test off, possibly...