python 3.5 requests 2.9.1 can not upload Chinese filename

[pc10201]

source code

`import requests

headers = { 'Origin': 'http://home.ctfile.com', 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/49.0.2623.75 Safari/537.36', 'Accept': '/', 'Referer': 'http://home.ctfile.com/', 'Accept-Encoding': 'gzip, deflate', 'Accept-Language': 'zh-CN,zh;q=0.8', }

files = {'file': open('中文.txt', 'rb')} values = {'name': r'中文.txt', 'filesize': '101'}

url = 'xxxx' r = requests.post(url, files=files, data=values, headers=headers)

print(r.content)`

I use fiddler to capture browser data.

The right data is: Content-Disposition: form-data; name="file"; filename="中文.txt"

The python script data is Content-Disposition: form-data; name="file"; filename*=中文.txt

so I change

Python35\Lib\site-packages\requests\packages\urllib3\fields.py

line 46-47 to

# value = email.utils.encode_rfc2231(value, 'utf-8') value = '%s="%s"' % (name, value)

It is OK now.

[Lukasa]

The right data is: Content-Disposition: form-data; name="file"; filename="中文.txt"

That is not the right data. This is:

Content-Disposition: form-data; name="file"; filename*=utf-8''%E4%B8%AD%E6%96%87.txt

This uses RFC 2231 already. To achieve it, change your code to this:

files = {'file': open(u'中文.txt', 'rb')}
values = {'name': r'中文.txt', 'filesize': '101'}

url = 'xxxx'
r = requests.post(url, files=files, data=values, headers=headers)

[pc10201]

The fiddler right data screenshot

not like filename*=utf-8''%E4%B8%AD%E6%96%87.txt

[Lukasa]

@pc10201 Did you change your code to match mine, including the unicode string for opening the file?

[pc10201]

I will try tomorrow.I will tell you result.

[pc10201]

I tried your code. It do not work

If you use windows,you can use fiddler to capture http data. if you use linux or mac os,you can use wireshark.

[Lukasa]

@pc10201 In what sense is that not working? I see the filename field formatted correctly.

[pc10201]

The server response is not right. The right data

The bad data

Please check the third line with each image.

[Lukasa]

Assuming for a moment that the different filenames are unrelated to this problem, the issue here is that the "right" case is not using RFC 2231 encoding. This is common, but wrong: the server needs to be able to handle the RFC 2231-encoded filename, which is what we're providing.

[pc10201]

You are right.The server may cause this problem.Please close this issue.