proof-of-concept decode XLS from ByteString

[olafklinke]

I added two functions to decode a ByteString. The function decodeXlsByteString is just a wrapper around decodeXlsIO as suggested by Erik Rybakken. The function decodeXlsByteString' makes use of the libxls function that can open buffers. The latter is experimental and when proven to work properly should replace the former. Since this is my first FFI project, someone please review the definition of Data.Xls.toCBuffer and the Storable XLSError instance. Specifically, how does one determine the correct sizeOf a C enum? On my machine, sizeof(xls_error_t) returns 4 with gcc, but that might be different with different compiler/architecture combinations.

[olafklinke]

My branch further changes the return type of decodeXlsIO to be more structured than list of lists of strings. Since while decoding there is type information on the cell available, there is no reason we should not expose that in the library. This ought to make downstream processing beyond xls2csv more convenient.

[harendra-kumar]

@olafklinke thanks for the PR. I will take a look at it on the weekend.

[olafklinke]

I neglected that the test suite needs adjustments, too. Will file new PR.