| Age | Commit message (Collapse) | Author |
|---|
|
Thanks Artem.
|
|
|
|
I will test it on a private project before settling on it.
|
|
If content-length is set in the response headers
we can skip chunked transfer-encoding.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Includes an experimental change to avoid deleting kerl
unnecessarily and/or incorrectly.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
We now check that the connection gets closed.
|
|
It was not working at all.
|
|
We don't implement CONNECT yet so we unconditionally return
a 501 for these requests and that makes this test fail.
|
|
|
|
|
|
It duplicated a test in misc_SUITE. Also fix a couple issues
in misc_SUITE including a test that was not being executed.
|
|
It seems to just be a duplicate of something we now have
in Cowlib.
|
|
|
|
Also fix the idle_timeout test which was producing
an extra crash log.
|
|
|
|
|
|
The feature was removed in Cowboy 2.0. An alternative to obtain
the same behavior will be introduced as cowboy_req:read_body_async/1.
|
|
|
|
|
|
|
|
|
|
|
|
This should reduce the amount of noise in RabbitMQ.
|
|
This fixes curl when fetching resources that use push.
|
|
|
|
|
|
Depend on Cowlib master for the moment.
|
|
Not much to test but heh.
|
|
|
|
[ci skip]
|
|
|
|
|
|
|
|
|
