You are looking at historical revision 12106 of this page. It may differ significantly from its current revision.
Intarweb is an advanced http library. It parses all headers into more useful Scheme values.
The intarweb egg is set up to be used from a variety of situations. For this reason, it does not try to be a full HTTP client or server. If you need that kind of functionality, see eggs like spiffy.
A request object (a defstruct-type record) can be created using the following procedure:[procedure] (make-request #!key uri port (method 'GET) (major 1) (minor 1) (headers (make-headers '())))
An existing request can be picked apart using the following procedures: <procedure>(request-uri REQUEST) => URI</procedure> <procedure>(request-port REQUEST) => PORT</procedure> <procedure>(request-method REQUEST) => SYMBOL</procedure> <procedure>(request-major REQUEST) => NUMBER</procedure> <procedure>(request-minor REQUEST) => NUMBER</procedure> <procedure>(request-headers REQUEST) => HEADERS</procedure>
The uri defines the entity to retrieve on the server, which should be a uri-generic-type URI object. The port is the port where the request is written to or read from. The method is a symbol that defines the HTTP method to use (case sensitive). major and minor identify the major and minor version of HTTP to use. Currently, 0.9, 1.0 and 1.1 are supported (but be careful with 0.9, it has some weird consequences and is not widely supported). Headers must be a headers object, which is described below.
The client will generally write requests, while the server will read them. To write a request, use the following procedure:[procedure] (write-request REQUEST) => REQUEST
This will write a request line with headers to the server. In case it is a request type that has any body data, this should be written to the the request's port. Beware that this port can be modified by write-request, so be sure to write to the port as it is returned by the write-request procedure![procedure] (read-request PORT) => REQUEST
Reads a request object from the given input-port. An optional request body can be read from the request-port after calling this procedure.
Requests are parsed using parse procedures, which can be customized by overriding this parameter:[parameter] (request-parsers [LIST])
The list is one of procedures which accept a request line string, which produce a request object from that, or #f if the request is not of the type handled by that procedure.
Requests are written using unparse procedures, which can be customized by overriding this parameter:[parameter] (request-unparsers [LIST])
The list is one of procedures which accept a request object and write to the request's output port and return the new, possibly updated request object. If the request object is not unparsed by this handler, it returns #f.
A response is also a defstruct-type record, much like a request:[procedure] (make-response #!key port (code 200) (reason "OK") (major 1) (minor 1) (headers (make-headers '())))
An existing response can be picked apart using the following procedures: <procedure>(response-port RESPONSE) => PORT</procedure> <procedure>(response-code RESPONSE) => NUMBER</procedure> <procedure>(response-reason RESPONSE) => STRING</procedure> <procedure>(response-major RESPONSE) => NUMBER</procedure> <procedure>(response-minor RESPONSE) => NUMBER</procedure> <procedure>(response-headers RESPONSE) => HEADERS</procedure>
The port, major, minor and headers are the same as for requests. code and reason are an integer status code and the short message that belongs to it, as defined in the spec (examples include: 200 OK, 301 Moved Permanently, etc).
A server will usually write a response, a client will read it. To write a response, use the following procedure:[procedure] (write-response RESPONSE) => RESPONSE
If there is a response body, this must be written to the response-port after sending the response headers.[procedure] (read-response PORT) => RESPONSE
Reads a response object from the port. An optional response body can be read from the response-port after calling this procedure.
Responses are parsed using parse procedures, which can be customized by overriding this parameter:[parameter] (response-parsers [LIST])
The list is one of procedures which accept a response line string, which produce a response object from that, or #f if the response is not of the type handled by that procedure.
Responses are written using unparse procedures, which can be customized by overriding this parameter:[parameter] (response-unparsers [LIST])
The list is one of procedures which accept a response object and write to the response's output port and return the new, possibly updated response object. If the response object is not unparsed by this handler, it returns #f.
Requests and responses contain HTTP headers wrapped in a special header-object to ensure they are properly normalized.[procedure] (headers ALIST [HEADERS]) => HEADERS
This creates headers based on an input list. This list has the header-name as a symbol key, and a list of values as value:
<example> <expr> (headers `((host ("example.com" . 8080))
(accept #(text/html ((q . 0.5))) #(text/xml ((q . 0.1))))) old-headers)
This adds the named headers to the existing headers in old-headers. The host header is either a string with the hostname or a pair of hostname/port. The accept header is a list of allowed mime-type symbols. As can be seen here, optional parameters or "attributes" can be added to a header value by wrapping the value in a vector of length 2. The first entry in the vector is the header value, the second is an alist of attribute name/value pairs.
The headers all have their own different types. Here follows a list of headers with their value types:
|Header name||Value type||Example value|
|accept||List of mime-types (symbols), with optional q attribute indicating "quality" (preference level)||(text/html #(text/xml ((q . 0.1))))|
|accept-charset||List of charset-names (symbols), with optional q attribute||(utf-8 #(iso-8859-5 ((q . 0.1))))|
|accept-encoding||List of encoding-names (symbols), with optional q attribute||(gzip #(identity ((q . 0))))|
|accept-language||List of language-names (symbols), with optional q attribute||(en-gb #(nl ((q . 0.5))))|
|accept-ranges||List of range types acceptable (symbols). The spec only defines bytes and none.||(bytes)|
|age||Age in seconds (number)||(3600)|
|allow||List of methods that are allowed (symbols).||(GET POST PUT DELETE)|
|authorization||Authorization information. This consists of a symbol identifying the authentication scheme, with scheme-specific attributes.||(digest #((username . "foo")))|
|cache-control||An alist of key/value pairs. If no value is applicable, it is #t||((public . #t) (max-stale . 10) (no-cache . (max-age set-cookie)))|
|connection||A list of connection options (symbols)||(close)|
|content-encoding||A list of encodings (symbols) applied to the entity-body.||(deflate gzip)|
|content-language||The natural language(s) of the "intended audience" (symbols)||(de nl en-gb)|
|content-length||The number of bytes (an exact number) in the entity-body||(10)|
|content-location||A location that the content can be retrieved from (a uri-generic object)||(<#uri-generic# ...>)|
|content-md5||The MD5 checksum (a string) of the entity-body||("12345ABCDEF")|
|content-range||Content range (pair with start- and endpoint) of the entity-body, if partially sent||((25 . 120))|
|content-type||The mime type of the entity-body (a symbol)||(text/html)|
|date||The date at which the message originated||TODO|
|etag||An entity-tag (pair, car being either the symbol weak or strong, cdr being a symbol) that uniquely identifies the resource contents.||((strong . foo123))|
|expect||Expectations of the server's behaviour (alist of symbol-string pairs), possibly with parameters.||(#(((100-continue . #t)) ()))|
|expires||Expiry timestamp for the entity||TODO|
|from||The e-mail address (a string) of the human user who controls the client||("email@example.com")|
|host||The host to use (for virtual hosting). This is a pair of hostname and port||(("example.com" . 80))|
|if-match||Entity-tags (pair, weak/strong symbol and unique entity identifier symbol) which must match.||((strong . foo123) (strong . bar123))|
|if-modified-since||Timestamp which indicates since when the entity must have been modified.||TODO|
|if-none-match||Entity tags (pair, weak/strong symbol and unique entity identifier symbol) which must not match.||((strong . foo123) (strong . bar123))|
|if-range||The range to request, if the entity was unchanged||TODO|
|if-unmodified-since||A timestamp since which the entity must not have been modified||TODO|
|last-modified||A timestamp when the entity was last modified||TODO|
|location||A location (an URI object) to which to redirect||(<#uri-object ...>)|
|max-forwards||The maximum number of proxies that can forward a request||(2)|
|pragma||An alist of symbols containing implementation-specific directives.||((no-cache . #t) (my-extension . my-value))|
|proxy-authenticate||Proxy authentication options (authentication scheme symbol, with parameters)||(digest #((username . "foo")))|
|proxy-authorization||Same as the above, only request-side instead of response-side||(digest #((username . "foo")))|
|range||The range of bytes (a pair of start and end) to request from the server.||((25 . 120))|
|referer||The referring URL (uri-generic object) that linked to this one.||(<#uri-object ...>)|
|retry-after||Timestamp after which to retry the request if unavailable now.||TODO|
|server||Information about the server (a string)||TODO|
|te||Allowed transfer-encodings (symbols, with optional q attribute) for the response||(deflate #(gzip ((q . 0.2))))|
|trailer||Names of header fields (symbols) available in the trailer/after body||(range etag)|
|transfer-encoding||The encodings (symbols) used in the body||(chunked)|
|upgrade||Product names to which must be upgraded (strings)||TODO|
|user-agent||Product name of the user agent being used (string)||TODO|
|vary||The names of headers that define variation in the resource body, to determine cachability (symbols)||(range etag)|
|via||The intermediate hops through which the message is forwarded (strings)||TODO|
|warning||Warning code for special status||TODO|
|www-authenticate||If unauthorized, a challenge to authenticate (symbol, with attributes)||(digest #((username . "foo")))|
|set-cookie||Cookies to set (name/value string pair, with attributes)||(#(("foo" . "bar") ((max-age . 10))))|
|cookie||Cookies that were set (name/value string pair, with attributes)||(#(("foo" . "bar") (($path . "/"))))|
Any unrecognised headers are assumed to be multi-headers, and the entire header lines are put unparsed into a list, one entry per line.
The parsers used to read and write header values can be customized with the following parameters:[parameter] (header-parsers [ALIST])
[parameter] (header-unparsers [ALIST])
These parsers are indexed with as key the header name (a symbol) and the value being a procedure which accepts three values: the name of the header (symbol), the contents of the header (a string, without the leading header name and colon) and the preceding headers. It should merge the new header with the preceding headers and return the resulting headers.
Header parsers are supposed to call these procedures to add headers:[procedure] (replace-header-contents NAME CONTENTS HEADERS) => HEADERS
[procedure] (replace-header-contents! NAME CONTENTS HEADERS) => HEADERS
[procedure] (update-header-contents NAME CONTENTS HEADERS) => HEADERS
[procedure] (update-header-contents! NAME CONTENTS HEADERS) => HEADERS
The replace procedures replace any existing contents of the named header with new ones, the update procedures add these contents to the existing header. The procedures with a name ending in bang are linear update variants of the ones without the bang. The header contents have to be normalized to be a 2-element vector, with the first element being the actual value and the second element being an alist (possibly empty) of parameters/attributes for that value.
The update procedures append the value to the existing header if it is a multi-header, and act as a simple replace in the case of a single-header.
Whether a header is allowed once or multiple times in a request or response is determined by this parameter:[parameter] (single-headers [LIST])
The value is a list of symbols that define header-names which are allowed to occur only once in a request/response.
- 0.1 Initial version
Copyright (c) 2008, Peter Bex All rights reserved. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. Neither the name of the author nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.