When using the url http://docs.python-requests.org/en/latest/, it redirects to https://2.python-requests.org//en/latest/ (notice the extra / before en). This causes a HTTP 404.

Expected Result

The redirect should be to https://2.python-requests.org/en/latest/

Actual Result

HTTP 404

Reproduction Steps

Try to visit the latest en documentation for requests using the

Description
Sending empty files in a multipart POST form is a proper use case due to web browsers supporting it.

I worked on a API layer using guzzle that had to work with an existing legacy codebase and trigger some actions (a proxy of some sort). There was a case where files had to be sent with empty content and would not work if you would omit them.

Example
_Currently no implem

If you're using proxies with requests-html and rendering JS sites is all good. Once you render a website pyppeteer don't know about this proxies and will expose your IP. This is an undesired behavior when scraping with proxies.

The idea is that whenever someone passes in proxies to the session object or any method call, make pyppeteer also use these proxies. #265

When using browser.get(url, stream=True) it does not in fact stream, but wait for the entire content to download and add it to the soup. One has to do browser.session.get(...).

This should be documented somewhere or there should be a get_without_adding_to_soup. I spent 3 hours debugging this yesterday, maybe this can save someone some time in the future.

https://github.com/MechanicalSo

Sample Code:

auto r = cpr::Get(cpr::Url{ "https://youtube.com" },cpr::VerifySsl(false));
std::cout << r.url << std::endl; 
std::cout << r.status_code << std::endl; // 200
std::cout << r.text << std::endl;

Output:

https://youtube.com/
0

The above output is identical even if spr::VerifySsl(false) is removed. Request works fine if using HTTP.

I'm running Ubuntu 16.

背景描述

最近才开始了解项目，正在看文档，中文的文档在访问的时候会出现无法访问情况。简单查了一下服务好像不在国内

重点描述遇到的问题：在国内访问文档不稳定

期望的功能特性

期望 文档服务器在国内可以有镜像

Currently grequests always sets a filename when creating a multipart post request https://github.com/levigross/grequests/blob/master/request.go#L312 . This leads to inconsistent behaviors with python's requests.

More specifically, lots of web servers (e.g., go's net/http, python's flask) rely on the existence of 'filename' when parsing a multipart form. If 'filename' exists, the field will be

PSR的RequestInterface没有规定实现exec方法，所以我设计组装一个http客户端无关的request的时候，我肯定不能绑定和客户端强相关的exec方法到Request上，因为每个客户端的异常类型、处理逻辑都不相同。

我设计一个composer组件，在组装请求部分，返回了个psr对象，本意是guzzle或者saber等支持PSR标准的HTTP客户端都可以按psr标准把这个对象代表的请求发送出去，现在Guzzle可以做到（$guzzleClient->send($PSRrequest)）而saber因为把PSR相关的处理逻辑绑定到他自定义的Request上，导致这样的设计没法实施。

Suggested by @Shanoir on Gitter:

One thing that I can suggest though: Overall the documentation of uplink is very nice. it clearly lays out how the API classes can be set up, and how to use the decorators. However, there is almost no examples available for the "use" of the methods. I think that across the board, providing these usage examples could really help newcomers get onboarded on the l

It would be great to have a changelog file where the changes between releases were detailed.
It would probably be too tedious to add all previous releases, but at least creating one just with the last release and updating it on future releases would be very helpful for people using this library.

Sending a POST-request to a site which is then redirecting back to itself results in reposting the request. This can result in an endless loop which eventually throws a StackOverflowException.

At least 303 Redirects should overwrite the request method to GET. Please read here:
https://developer.mozilla.org/en-US/docs/Web/HTTP/Redirections#Temporary_redirections