mutantmonkey

The documentation makes it sound like it should, but that's not the case unfortunately. If you trace things back through urllib3's underlying `stream` and `read` functions, through `http.client.HTTPResponse.read`, you ultimately...

Using a chunk size of one causes unnecessarily high CPU usage because each time a byte is received, it has to be processed by the Python code instead of just...

> Alternatively get rid of the short read functionality altogether. Who knows what other issues with it haven't been discovered yet. That sounds like a problem that should be solved...

I have a fix that I believe will work waiting at https://github.com/mutantmonkey/sseclient/tree/use_urllib3_decode, but I haven't had a chance to set up a SSE server that uses chunked encoding to confirm...

@TheSandDoctor I have not; the situation is unchanged right now. I have a fix that I believe will work but I need way to reproduce the issue to confirm this....

I'm using Go 1.5.3.

I have the server set to `https://textsecure-service.whispersystems.org:443`. This is the output I get when I run `textsecure`: ``` DEBU[2016/01/21 22:26:51] PUT /v1/directory/tokens 200 tls: either ServerName or InsecureSkipVerify must be...