Fix parseMultipartForm to accept unquoted strings

[tchaloupka]

Content-Disposition fields can be passed as an unquoted string in which case the current parser fails.

https://tools.ietf.org/html/rfc6266#section-4.1

So I adjusted the parser to depend on semicolon as a separator or to use the rest of the input if the semicolon is not found.

For example .Net can send requests like this:

POST https://192.168.0.78:8082/api/v1/foos HTTP/1.1
User-Agent: bar
Accept-Encoding: gzip, deflate
Accept-Language: en-US
Content-Type: multipart/form-data; boundary="7518e894-589d-44be-9a04-9f45c6124429"
Host: 192.168.0.78:8082
Content-Length: 55839
Expect: 100-continue
Connection: Keep-Alive

--7518e894-589d-44be-9a04-9f45c6124429
Content-Type: text/plain; charset=utf-8
Content-Disposition: form-data; name=name

Some test
--7518e894-589d-44be-9a04-9f45c6124429
Content-Type: text/plain; charset=utf-8
Content-Disposition: form-data; name=desc

Some test description
--7518e894-589d-44be-9a04-9f45c6124429
Content-Disposition: form-data; name=file; filename=CadEx_schema.png; filename*=utf-8''CadEx_schema.png

...file content...
--7518e894-589d-44be-9a04-9f45c6124429--

[tchaloupka]

This fixes #1562

[s-ludwig]

To be fully correct, shouldn't there be two branches, one that searches for " in case the value begins with " and one that searches for ; if it doesn't?

Otherwise LGTM.

[tchaloupka]

I've made the branch but also it is actually needed to skip escaped quotes inside so I changed that too.

[s-ludwig]

Probably would be the best idea to throw an exception here. Otherwise this would result in an out-of-bounds error in parseValue, which terminates the application.

[s-ludwig]

Thanks! I indeed missed the escaping, but this looks good now.