| Age | Commit message (Collapse) | Author |
|---|
|
|
|
This is a first step to improve the HTTP status codes returned
by Cowboy on crashes. We will tweak it over time.
Also fixes a small bug where two replies may have been sent
when using loop handlers under rare conditions.
|
|
|
|
|
|
Pointed out during the ConcuError tutorial by Kostis. Thanks! :-)
|
|
|
|
|
|
|
|
When something went wrong in a handler we used to report errors
and then terminate the process normally. This doesn't work so
well with links which won't detect failure.
Now we still catch the error, but throw another one with more
details on why it happened, including the Req object information
and the stacktrace. Ranch will then print an error message with
all this information.
Because we crash directly, this also means that we will not hog
resources unnecessarily for too long when something bad happens.
|
|
|
|
|
|
|
|
|
|
The unmasking logic was based on the length of inflated data instead
of the length of the deflated data. This meant data would get corrupted
when we receive a websocket frame split across multiple TCP packets.
|
|
|
|
|
|
|
|
We can change this call when we start supporting only R16B+,
which may happen when R17B is released or some time before.
|
|
|
|
It accepts iodata() and not binary() for the payload.
|
|
|
|
|
|
|
|
Worked until cccc0bc
|
|
Sends can be successful even if the connection is gone.
|
|
|
|
Also fixes a warning.
|
|
We now always send a failure reason (bad protocol, bad encoding, etc.)
unless the closure was initiated by the client and it didn't send a
close code.
We now check that the close frames have a payload that is valid UTF-8,
unless they don't have a payload at all.
We now do not crash the process anymore when bad opcodes are sent, or
when the opcode 0 is sent before fragmentation was initiated.
Overall this makes us closer to full compliance with the RFC.
|
|
The autobahntestsuite now passes 100% of the tests. We are
getting close to fully implementing the Websocket RFC.
|
|
No change in functionality, but this will allow us to validate
that text frames are utf8 without having to receive the frame
entirely.
|
|
Good in theory, but implementations may vary. If something stops
working after this commit we might need some tweaks to support
existing clients.
Please try it and give feedback.
|
|
It was only used by Safari 5.0.1 and possibly 5.1. Their market share
is dropping as we speak. It was also insecure (disabled in Firefox
for that reason).
This will allow us to make much more efficient and cleaner code for
the rest of the Websocket versions we support (drafts 7 to 17 + RFC),
which are pretty much all versions seen in the wild excluding the
one we're removing here.
|
|
It was added to help with response body streaming functions.
But it was a clumsy solution that we discarded in favor of
passing socket and transport to said function. It was also
very odd compared to the rest of the cowboy_req interface.
If you used this function before, worry not, here's its
proper equivalent.
[Socket, Transport] = cowboy_req:get([socket, transport], Req)
|
|
Middlewares allow customizing the request processing.
All existing Cowboy project are incompatible with this commit.
You need to change `{dispatch, Dispatch}` in the protocol options
to `{env, [{dispatch, Dispatch}]}` to fix your code.
|
|
This commit prevents erlang messages from keeping a websocket connection
alive. Previously, the timer was canceled upon any activity. Now, the
timeout is only canceled when actual data is sent from the client. The
handler_loop_timeout/1 function is called from websocket_data/4 instead
of handler_before_loop/4. It is also called after every successful reply
in handler_call/4.
|
|
The purpose of this patch is to make the arguments cowboy passes to
error_logger more consistent. With this patch there's only 3 variations
on the error_logger argument list; a 5 element list, an 8 element list
and a 10 element list. In all cases, the first 3 arguments are the
Module, Function and Arity of the function being called and the
second-to-last argument is always the Request. Additionally, for lists
longer than 5 elements, the last argument is always the stack-trace.
The added consistency of the argument ordering makes it much easier to
write code in lager's error_logger handler to catch these messages and
write a pretty one-liner (while writing the full message to the
crash.log).
|
|
|
|
|
|
|
|
|
|
We can now reply empty close, ping and pong frames, or close
frames with a payload.
This means that we can send a frame and then close the connection
in a single operation.
If a close packet is sent, the connection is closed immediately,
even if there was frames that remained to be sent. Cowboy will
silently drop any extra frames in the list given as a reply.
|
|
|
|
In some situations, the underlying socket might become "half-open" in
which case the websocket will stay in a waiting state indefinitely. The
detection of this state requires checking for errors when calling send.
|
|
Instead of returning {text, Data}, you can now return
[{text, Data}, {text, Data2}, ...].
|
|
As suggested by @prof3ta.
|
|
Header names are now binaries. Since header names are case insensitive
they are all converted to lowercase. For example: <<"content-length">>.
The max_line_length option was removed. Three new options have been
added instead:
* max_request_line_length (defaults to 4096)
* max_header_name_length (defaults to 64)
* max_header_value_length (defaults to 4096)
|
|
The signature of parse_header, body_qs, multipart_data and
the set_resp_* functions has changed.
See the cowboy_req module edoc for more details.
|
|
|
|
|
|
Use it in cowboy_websocket for hixie76, replacing http by ws.
|
