The module cowboy_req
provides functions to access, manipulate
and respond to requests.
There are four types of functions in this module. They can be differentiated by their name and their return type:
Type | Name pattern | Return type |
---|---|---|
access |
no verb, parse_*, match_* |
|
question |
has_* |
|
modification |
set_* |
|
action |
any other verb |
|
Any Req
returned must be used in place of the one passed as
argument. Functions that perform an action in particular write
state in the Req object to make sure you are using the function
correctly. For example, it’s only possible to send one response,
and to read the body once.
Raw request:
-
cowboy_req:method(3) - HTTP method
-
cowboy_req:version(3) - HTTP version
-
cowboy_req:scheme(3) - URI scheme
-
cowboy_req:host(3) - URI host name
-
cowboy_req:port(3) - URI port number
-
cowboy_req:path(3) - URI path
-
cowboy_req:qs(3) - URI query string
-
cowboy_req:uri(3) - Reconstructed URI
-
cowboy_req:header(3) - Raw HTTP header value
-
cowboy_req:headers(3) - Raw HTTP headers
-
cowboy_req:peer(3) - Peer address and port
Processed request:
-
cowboy_req:parse_qs(3) - Parse the query string
-
cowboy_req:match_qs(3) - Match the query string against constraints
-
cowboy_req:parse_header(3) - Parse the given HTTP header
-
cowboy_req:parse_cookies(3) - Parse cookie headers
-
cowboy_req:match_cookies(3) - Match cookies against constraints
-
cowboy_req:binding(3) - Access a value bound from the route
-
cowboy_req:bindings(3) - Access all values bound from the route
-
cowboy_req:host_info(3) - Access the route’s heading host segments
-
cowboy_req:path_info(3) - Access the route’s trailing path segments
Request body:
-
cowboy_req:has_body(3) - Is there a request body?
-
cowboy_req:body_length(3) - Body length
-
cowboy_req:read_body(3) - Read the request body
-
cowboy_req:read_urlencoded_body(3) - Read and parse a urlencoded request body
-
cowboy_req:read_part(3) - Read the next part of a multipart body
-
cowboy_req:read_part_body(3) - Read the current part’s body in a multipart body
Response:
-
cowboy_req:set_resp_cookie(3) - Set a cookie
-
cowboy_req:set_resp_header(3) - Set a response header
-
cowboy_req:has_resp_header(3) - Is the given response header set?
-
cowboy_req:delete_resp_header(3) - Delete a response header
-
cowboy_req:set_resp_body(3) - Set the response body
-
cowboy_req:has_resp_body(3) - Is there a response body?
-
cowboy_req:reply(3) - Send the response
-
cowboy_req:stream_reply(3) - Send the response and stream its body
-
cowboy_req:stream_body(3) - Send a chunk of the response body
-
cowboy_req:push(3) - Push a resource to the client
push_opts() :: #{
method => binary(), %% case sensitive
scheme => binary(), %% lowercase; case insensitive
host => binary(), %% lowercase; case insensitive
port => inet:port_number(),
qs => binary() %% case sensitive
}
Push options.
By default, Cowboy will use the GET method, an empty query string, and take the scheme, host and port directly from the current request’s URI.
read_body_opts() :: #{
length => non_neg_integer(),
period => non_neg_integer(),
timeout => timeout()
}
Body reading options.
The defaults are function-specific.
req() :: #{
method := binary(), %% case sensitive
version := cowboy:http_version() | atom(),
scheme := binary(), %% lowercase; case insensitive
host := binary(), %% lowercase; case insensitive
port := inet:port_number(),
path := binary(), %% case sensitive
qs := binary(), %% case sensitive
headers := cowboy:http_headers(),
peer := {inet:ip_address(), inet:port_number()}
}
The Req object.
Contains information about the request and response. While some fields are publicly documented, others aren’t and shouldn’t be used.
You may add custom fields if required. Make sure to namespace them by prepending an underscore and the name of your application:
Req#{_myapp_auth_method => pubkey}.
resp_body() :: iodata()
| {sendfile, Offset, Length, Filename}
Offset :: non_neg_integer()
Length :: pos_integer()
Filename :: file:name_all()
Response body.
It can take two forms: the actual data to be sent or a tuple indicating which file to send.
When sending data directly, the type is either a binary or an iolist. Iolists are an efficient way to build output. Instead of concatenating strings or binaries, you can simply build a list containing the fragments you want to send in the order they should be sent:
1> RespBody = ["Hello ", [<<"world">>, $!]].
["Hello ",[<<"world">>,33]]
2> io:format("~s~n", [RespBody]).
Hello world!
When using the sendfile tuple, the Length
value is mandatory
and must be higher than 0. It is sent with the response in the
content-length header.