aboutsummaryrefslogtreecommitdiffstats
path: root/src/cowboy_http.erl
AgeCommit message (Collapse)Author
2018-11-09Add sendfile support to cowboy_req:stream_bodyLoïc Hoguin
It is now possible to stream one or more sendfile tuples. A simple example of what can now be done would be for example to build a tar file on the fly using the sendfile syscall for sending the files, or to support Range requests with more than one range with the sendfile syscall. When using cowboy_compress_h unfortunately we have to read the file in order to send it. More options will be added at a later time to make sure users don't read too much into memory. This is a new feature however so existing code is not affected. Also rework cowboy_http's data sending to be flatter.
2018-11-03Add an option to disable sendfile for a listenerLoïc Hoguin
2018-10-31Add compress_threshold protocol optionSteve Domin
Currently the compression threshold is set to 300 and hardcoded in the codebase. There are cases where it make sense to allow this to be configured, for instance when you want to enforce all responses to be compressed regarldess of their size.
2018-10-31Make sure we don't send error_responses on the wrong streamLoïc Hoguin
2018-10-31Improve a few types, including cowboy_req:req()Loïc Hoguin
2018-10-30Initial support for the PROXY protocol headerLoïc Hoguin
Depend on Ranch master for now since it isn't in any release yet.
2018-10-28Exit gracefully on parent exit/sys:terminate/2,3Loïc Hoguin
2018-09-12Do not send a 101 after a final response in switch_protocolLoïc Hoguin
2018-06-28Introduce undocumented option loggerLoïc Hoguin
This commit reworks the logging that Cowboy does via error_logger to make the module that will do the actual logging configurable. The logger module interface must be the same as logger and lager: a separate function per log level with the same log levels they support. The default behavior remains to call error_logger, although some messages were downgraded to warnings instead of errors. Since error_logger only supports three different log levels, some messages may get downgraded/upgraded depending on what the original log level was to make them compatible with error_logger. The {log, Level, Format, Args} command was also added to stream handlers. Stream handlers should use this command to log messages because it allows writing a stream handler to intercept some of those messages and extract information or block them as necessary. The logger option only applies to Cowboy itself, not to the messages Ranch logs, so more work remains to be done in that area.
2018-06-27Disable warnings for erlang:get_stacktrace/0 in OTP-21+Loïc Hoguin
2018-06-25Add streaming without chunking for HTTP/1.1Eric Meadows-Jönsson
If content-length is set in the response headers we can skip chunked transfer-encoding.
2018-05-18Remove the trailer header from HTTP/1.1 response if no TELoïc Hoguin
2018-05-18Don't send transfer-encoding when streaming 204 responsesLoïc Hoguin
2018-05-18Do not process HTTP/1.1 requests coming in after the lastLoïc Hoguin
2018-05-18Add more validation of absolute-form request targetsLoïc Hoguin
2018-05-16Add option linger_timeout to cowboy_httpLoïc Hoguin
2018-05-16Wrap the sendfile call in a try/catch for HTTPLoïc Hoguin
This should reduce the amount of noise in RabbitMQ.
2018-03-26Fix the flushing of messages when switching to WebsocketLoïc Hoguin
We now flush messages that are specific to cowboy_http only. Stream handlers should also flush their own specific messages if necessary, although timeouts will be flushed regardless of where they originate from. Also renames the http_SUITE to old_http_SUITE to distinguish new tests from old tests. Most old tests need to be removed or converted eventually as they're legacy tests from Cowboy 1.0.
2018-03-14Add case for handling infinity for idle/request_timeoutBartek Walkowicz
Currently cowboy assumes that idle_timeout or request_timeout is a number and always starts timers. Similar situation takes place in case of preface_timeout for http2. This commit adds case for handling infinity as a timeout, allowing to not start mentioned timers.
2018-03-13Handle system messages in cowboy_websocketLoïc Hoguin
2018-03-13Handle supervisor calls properly everywhereLoïc Hoguin
2017-12-13Fix Dialyzer warningsLoïc Hoguin
2017-12-12Fix HTTP/1.1 pipeliningLoïc Hoguin
Cases where a request body was involved could sometimes fail depending on timing. Also fix all of the old http_SUITE tests.
2017-12-07Add more rfc7231 tests and a new max_skip_body_length optionLoïc Hoguin
The option controls how much body we accept to skip for HTTP/1.1 connections when the user code did not consume the body fully. It defaults to 1MB.
2017-12-06Fix 408 not sending connection: close for HTTP/1.1Loïc Hoguin
Also make sure the header is sent for all types of early_error that result in the closing of the connection.
2017-12-06Also disable the TRACE method entirelyLoïc Hoguin
2017-12-06Disable the CONNECT method completelyLoïc Hoguin
It's safer than allow it with the wrong behavior.
2017-11-29Don't send chunked bodies when the method is HEADLoïc Hoguin
2017-11-29Reject absolute URIs with no authority componentsLoïc Hoguin
2017-11-29Reject absolute URIs with userinfo components in HTTP/1.1Loïc Hoguin
2017-11-29Reject whitespace before/after HTTP/1.1 header names properlyLoïc Hoguin
2017-11-22Add more rfc7230 tests and better handle bad chunk sizesLoïc Hoguin
Bad chunk sizes used to be accepted and could result in a badly parsed body or a timeout. They are now properly rejected. Chunk extensions now have a hard limit of 129 characters. I haven't heard of anyone using them and Cowboy does not provide an interface for them, but we can always increase or make configurable if it ever becomes necessary (but I honestly doubt it). Also a test from the old http suite could be removed. Yay!
2017-11-20Add more rfc7230 tests and improve transfer-encodingLoïc Hoguin
It's worth noting that transfer-encoding now takes precedence over content-length as recommended by the RFC, so that when both headers are sent we only care about transfer-encoding and explicitly remove content-length from the headers.
2017-11-20Add a test for early errors that occur on the request-lineLoïc Hoguin
2017-11-20Remove a useless todoLoïc Hoguin
2017-11-20Add more rfc7230 testsLoïc Hoguin
Also fixes the handling of the max_headers option for HTTP/1.1. It is now a strict limit and not dependent on whether data is already in the buffer.
2017-11-15Add preliminary support for trailers in responsesLoïc Hoguin
This depends on changes in Cowlib that are only available on master.
2017-11-13Fix packet being dropped when using switch_protocolLoïc Hoguin
This only happens if the switch takes too long, and should not happen unless a spawned process refuses to shut down immediately.
2017-11-01Fix two edge cases for cowboy_req:stream_bodyLoïc Hoguin
Sending data of size 0 with the fin flag set resulted in nothing being sent to the client and still considering the response to be finished for HTTP/1.1. For both HTTP/1.1 and HTTP/2, the final chunk of body that is sent automatically by Cowboy at the end of a response that the user did not properly terminate was not passing through stream handlers. This resulted in issues like compression being incorrect. Some tests still fail under 20.1.3. They are due to recent zlib changes and should be fixed in a future patch release. Unfortunately it does not seem to be any 20.1 version that is safe to use for Cowboy, although some will work better than others.
2017-10-31Fix another warningLoïc Hoguin
2017-10-31Add informational responses to metricsLoïc Hoguin
2017-10-25Introduce cowboy_req:sock/1 and cowboy_req:cert/1Loïc Hoguin
To obtain the local socket ip/port and the client TLS certificate, respectively.
2017-10-22Ensure stream terminate is called when switching protocolsLoïc Hoguin
2017-10-21Fix sending of final chunk in HTTP/1.1Loïc Hoguin
I broke this when fixing stream handlers earlier.
2017-10-21Fix stream handler state being discarded on terminateLoïc Hoguin
When we have to send a response before terminating a stream, we call info. The state returned by this info call was discarded when we called terminate after that. This commit fixes it. There are no tests for this, however the new metrics test in the next commit requires the correct behavior so this is ultimately covered.
2017-10-20Fix HTTP/1.1 stopping streams too earlyLoïc Hoguin
It is possible in some cases to move on to the next request without waiting, but that can be done as an optimization later on if necessary.
2017-09-27Ensure the behavior on stream handler crash is consistentLoïc Hoguin
Also corrects the lack of error response when HTTP/1.1 is used.
2017-09-25Cleanup various commentsLoïc Hoguin
2017-09-25Move body length count to cowboy_stream_h instead of protocolsLoïc Hoguin
The documentation was correct, the code was not. This should make it easier to implement new protocols. Note that for HTTP/2 we will need to add some form of counting later on to check for malformed requests, but we can do simpler and just reduce from the expected length and then check if that's 0 when IsFin=fin.
2017-09-21Centralize stream handler error reporting in cowboy_streamLoïc Hoguin