Fixed content size parsing issue

[g3rb3n]

Added extra check for request_headers['content_length']

[prosperocu]

Hi g3rb3n. Your solution don't work on Django 2.1.7 because request_headers["content_length"] can be an empty string too. One possible solution can be don't parsing to int this header:

    req_size = request_headers["content_length"]

Im looking on the code, and this var is not needed for other function, is only for info. But, in other future case this can be help:

    req_size = 0
    if "content_length" in request_headers and request_headers["content_length"]:
        req_size = int(request_headers["content_length"])

[g3rb3n]

If request_headers["content_length"] is an empty string
"content_length" in request_headers will evaluate to True
request_headers["content_length"] will evaluate to False

So int(request_headers["content_length"]) won't be called and req_size will become 0 from the else clause.

Your proposal is logical equivalent. Or am I missing something?

[prosperocu]

Well, sorry for my bad English. Im cuban and, i don't a top python programmer. This solution works for me on my project using your library. Im using python 3.4.7 and Django 2.1.7 like a say with venv. In this case: int(request_headers["content_length"]) fails because request_headers["content_length"] is an empty string when i debug code with Pycharm. Thanks for reply. 2019-03-01 6:39 GMT-05:00, G3RB3N <notifications@github.com>:

…

If request_headers["content_length"] is an empty string "content_length" in request_headers will evaluate to True request_headers["content_length"] will evaluate to False So int(request_headers["content_length"]) won't be called and req_size will become 0 from the else clause. Your proposal is logical equivalent. Or am I missing something? -- You are receiving this because you commented. Reply to this email directly or view it on GitHub: #4 (comment)

[g3rb3n]

I had the same problem when request_headers["content_length"] was returning an empty string. This PR fixes this issue. It is still not merged though by @elafarge, so you should indeed use my master version if you have the same issue.