differenciate between form_data and files based on filename headder

[theelous3]

>>> from io import BytesIO
>>> from multipart import MultipartParser
>>> 
>>> m = '--8banana133744910kmmr13a56!102!2405\r\nContent-Disposition: form-data; name="file_1";\r\n\r\nCompooper\r\n--8banana133744910kmmr13a56!102!2405--\r\n'
>>> 
>>> boundary = '8banana133744910kmmr13a56!102!2405'.encode()
>>> 
>>> parser = MultipartParser(BytesIO(m.encode()), boundary)
>>> 
>>> parser.parts()[0]
<multipart.MultipartPart object at 0xb716d4ec>
>>> parser.parts()[0].form_data
True
>>> parser.parts()[0].file
False

---

>>> m = '--8banana133744910kmmr13a56!102!2405\r\nContent-Disposition: form-data; name="file_1"; filename="test_file1.txt"; Content-Type: application/octet-stream\r\n\r\nCompooper\r\n--8banana133744910kmmr13a56!102!2405--\r\n'
>>> 
>>> boundary = '8banana133744910kmmr13a56!102!2405'.encode()
>>> 
>>> parser = MultipartParser(BytesIO(m.encode()), boundary)
>>> 
>>> parser.parts()[0]
<multipart.MultipartPart object at 0xb723678c>
>>> parser.parts()[0].file
True
>>> parser.parts()[0].form_data
False
>>> 

[theelous3]

Note that this is breaking tests. (Things expect to use the bytesio object at .file, but .file is now just a bool.)

I foresee a fair bit of work going on in this lib over the next short while, and keeping tests up to date will likely be fruitless work in the mean time.

[theelous3]

Actually, who cares. Realising my problem wasn't that the type of the data wasn't being catagorised, but only that everything was catagorised under file. Easier to just chuck that whole thing.