NOT COMPLIANT WITH RFC FORMAT

[cheneydeng]

hi all,
I found that you guys have not understand what the format of multipart/form-data, the boundary is not as the same as it show in the body.Say we got the boundary in the HTTP header,as "--xxyy",then the "boundary part" in the body is
--boundary =====> ----xxyy.
and the last "boundary" =====> --boundary-- ======> --xxyy--
you misunderstand this and it will cause your code parse the data incorrectly.

[jamesamcl]

Are you sure? Isn't that what s_part_data_final_hyphen is for?

[andrzej-bieniek]

I verified with curl, and "boundary part" in the body is additionally prefixed with "--".
@udp s_part_data_final_hyphen handles "--" suffix but not prefix.

$ curl -X POST -F name=John http://localhost:2000
Sends:

POST / HTTP/1.1
User-Agent: curl/7.35.0
Host: localhost:2000
Accept: */*
Content-Length: 143
Expect: 100-continue
Content-Type: multipart/form-data; boundary=------------------------adcf844310cc278d

--------------------------adcf844310cc278d
Content-Disposition: form-data; name="name"

John
--------------------------adcf844310cc278d--

Comparing boundary values:

  ------------------------adcf844310cc278d    (from Content-Type)
--------------------------adcf844310cc278d    (from body)
--------------------------adcf844310cc278d--  (from body)

[c1tru55]

Don't know how it is

Tested as part of Cosmonaut HTTP server.

but code logic is wrong and @cheneydeng absolutely right. Surprisingly that such serious bug still not fixed...

it will cause your code parse the data incorrectly.

in my case it doesn't work at all...

[iafonov]

@c1tru55 feel free to fork and fix. I don't use it anymore and Cosmonaut was pet project never used in real application

[twxjyg]

I found this bug yet ...

[MrBuddyCasino]

Just found this bug as well. Please at least add a hint to your readme.

[zhengw1024]

not rfc standard

[ghost]

Hi!
Is this commit actual?