scipy.io.wavfile.read fails silently on files > 2**32 bytes

[pletnes]

I noticed that scipy.io.wavfile.read will happily read a wavefile larger than 2**32 bytes, even though these files are not supported by the standard. However, scipy seems to read the filesize % 2**32 first bytes.

Reproducing code example:

# if `filename` corresponds to a large file, this does not crash but returns partial data
fs, wavedata = scipy.io.wavfile.read(filename)

Question: should this raise e.g. an IOError('wavefile too large') or similar?

Scipy/Numpy/Python version information:

import sys, scipy, numpy; print(scipy.__version__, numpy.__version__, sys.version_info)
# 1.2.0 1.15.4 sys.version_info(major=3, minor=6, micro=8, releaselevel='final', serial=0)

[sethtroisi]

Makes senses from the code

I suspect that the header gets the lower bytes of filesize and we respect that when we read.
It's possible we could check the size of the file and read the whole file (especially if it's 2^32 + header size)