Matt L

Already fixed in #2101/#2114

From further testing, this error doesn't have anything to do with UTF-8 encoding. It happens when attempting to write a sequence of variable-length strings of any type, because HSDS's current...

> Doesn't testPutVLenCompoundBinary in vlen_test.py have vlen and fixed types in the same compound type? Yes, but that test doesn't cover a variable-length sequence of variable length strings. That said,...

> But currently vlen data isn't supported with H5T_ARRAY - see https://github.com/HDFGroup/hsds/blob/master/hsds/util/hdf5dtype.py, line 663 H5T_ARRAY doesn't allow its base type to be a variable length sequence (type of class `H5T_VLEN`),...

This behavior is now properly tested as an expected failure (since vlen sequences of vlen data aren't supported) by `testPutVlenVlenError`

Looks like it's correct when HSDS returns the variable lengths types in binary instead of JSON. Changing the VOL to use binary instead of JSON should be faster for this...

`H5Tdecode()` doesn't take a parameter specifying the size of the buffer, so it's not possible for the library to do normal bounds checking on it - instead, bounds checking has...

There isn't currently any way to pre-allocate chunks before expanding a dataset. Extending and then shrinking the dataset won't leave anything allocated - if its extent is decreased, chunks outside...

Chunk preallocation fits within the design of HDF5. I'll leave this open for now as an indication of interest in the feature, though it's not planned for implementation in the...