Improvement to multipart parser

[mgaillard]

I noticed the parsing of multipart forms had two issues:

1. The regex would fail if attributes are not in the following order: name, filename, filename*
2. Filename* attributes are not supported

This PR is my attempt at solving the problem. I noticed two issues are related to this piece of code: #1250 and #1212.

[mgaillard]

I see that the build is failing because my tests are not compatible with the .h .cpp split script. I will look at it tomorrow.

[mgaillard]

When running split.py the class MultipartFormDataParser ends up in the httplib.cc file, which is not visible in the test.cc
For writing tests it is much easier to access directly MultipartFormDataParser instead of going through a webserver, etc.
Would it be fine to split declaration and definition of MultipartFormDataParser so that it becomes public?

[mgaillard]

I added:

bool parse_multipart_response(const Response &res, MultipartFormDataItems& files);

To allow testing the multipart form data parser without making it public.
It also allows the client code to parse multipart responses from the server.

[mgaillard]

I fixed the code so that tests pass, could I get a review?

[yhirose]

@mgaillard I fixed this issue in a simpler way. Could you try the latest httplib.h in the master branch? Thanks!

https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Content-Disposition
https://httpwg.org/specs/rfc6266.html#header.field.definition
https://www.rfc-editor.org/rfc/rfc5987.html#section-3.2
https://stackoverflow.com/questions/93551/how-to-encode-the-filename-parameter-of-content-disposition-header-in-http