The `net' collection contains libraries that provide access to the
following _Internet_ (quasi-)protocols:

  URL parsing
  CGI backends
  sendmail
  SMTP
  NNTP
  POP-3
  IMAP
  Mail header reading and writing
  DNS

==========================================================================
_URL_ posting, _web clients_, _WWW_
==========================================================================

To load directly: (require (lib "url.ss" "net"))
Module files: _url.ss_ - provides the procedures documented below
              _url-unit.ss_ - provides unit net:url@
              _url-sig.ss_ - provides signature net:url^
              _url-struct.ss_ - provides the url and path/param structs


ABSTRACT -------------------------------------------------------------

The url package helps programmers use URIs as specified in RFC 2396:
http://www.ietf.org/rfc/rfc2396.txt

TYPES ----------------------------------------------------------------

_url struct_
 (define-struct url (scheme user host port path-absolute? path query fragment))
> url-scheme : url -> (union false/c string?) 
> url-user : url -> (union false/c string?)
> url-host : url -> (union false/c string?)
> url-port : url -> (union false/c number?)
> url-path-absolute? : url -> boolean?
> url-path : url -> (listof path/param?)
> url-query : url -> (listof (cons/c symbol? string?))
> url-fragment : url -> (union false/c string?)
> url? : any -> boolean
> make-url : ...as-above.. -> url

  The basic structure for all URLs, as explained in rfc3986
  http://www.ietf.org/rfc/rfc3986.txt

For example, this url:

  http://sky@www.cs.brown.edu:801/cgi-bin/finger;xyz?name=shriram;host=nw#top
  {-1}   {2} {----3---------} {4}{---5-------------}{----7-------------} {8}
                                                 {6}

  1 = scheme, 2 = user, 3 = host, 4 = port,
  5 = path,  6 = param (or last path segment),
  7 = query, 8 = fragment

  The strings inside the fields user, path, query, and fragment are
  represented directly as Scheme strings, ie without
  url-syntax-specific quoting. The procedures `string->url' and
  `url->string' translate things like %20 into spaces and back again.

  By default, query associations are parsed with either ";" or "&" as
  a separator, and they are generated with ";" as a separator. The
  `current-alist-separator-mode' parameter for "uri-codec.ss" adjusts
  this default.

  An empty string at the end of the list of paths
  corresponds to a url that ends in a slash. For example,
  this url: http://www.drscheme.org/a/ has a path field with
  strings "a" and "" and this url: http://www.drscheme.org/a
  has a path field with only the string "a".

_ path/param struct_
  (define-struct path/param (path param))

> path/param-path : path/param -> (union string? (symbols 'up 'same))
> path/param-param : path/param -> (listof string)
> path/param? : any -> boolean
> make-path/param : (union string? (symbols 'up 'same)) (listof string) -> path/param

  A pair, that joins a path segment with its params in a
  url.

> pure-port

  A pure port is one from which the MIME headers have been removed, so
  that what remains is purely the first content fragment.

> impure-port

  An impure port is one that still has its MIME headers.  Contrast
  with a pure-port.

PROCEDURES -----------------------------------------------------------

> unixpath->path

  This procedure, which used to be provided by this library, is no
  longer supported.

> (string->url string) -> url

  Parses the url specified by the string into a url struct. The
  `string->url' procedure uses `form-urlencoded->alist' when parsing
  the query, so it is sensitive to the `current-alist-separator-mode'
  parameter for determining the association separator.

> (combine-url/relative url string) -> url

  Given a base URL and a relative path, combines the two and returns a
  new URL as per the URL combination specification.  Call the
  arguments base and relative.  They are combined according to the
  rules in rfc3986 (above).

  This function does not raise any exceptions.

> (netscape/string->url string) -> url

  Turns a string into a URL, applying (what appear to be) Netscape's
  conventions on automatically specifying the scheme: a string
  starting with a slash gets the scheme "file", while all others get
  the scheme "http".

> (url->string url) -> string

  Generates a string corresponding to the contents of the url struct.
  For a "file:" URL, empty strings in the path list are treated as
  'same for `build-path'.

  The `url->string' procedure uses `alist->form-urlencoded' when
  formatting the query, so it it sensitive to the
  `current-alist-separator-mode' parameter for determining the
  association separator. In particular, the default is to separate
  associations with ";" (based on modern recommendations) instead of
  "&".

> (decode-some-url-parts url) -> url

  This procedure, which used to be provided by this library, is no
  longer supported or necessary. A `url' struct not contains properly
  decoded information.

> (get-pure-port url [list-of-strings]) -> input-port

  Takes a URL and returns a pure port corresponding to it.  Writes the
  optional RFC 822 header-line strings to the server.  For the "file"
  scheme, uses open-input-file, does not handle exceptions, and
  ignores the optional strings.

> (post-pure-port url post-byte-string [list-of-strings]) -> input-port

  Like `get-pure-port', but issues a POST request (instead of a GET
  request) to a web server using the given post data byte string.

> (get-impure-port url [list-of-strings]) -> input-port

  Takes a URL and returns an impure port corresponding to it.  Writes
  the optional RFC 822 header-line strings to the server.  There are
  no impure ports with scheme "file".

> (post-impure-port url post-byte-string [list-of-strings]) -> input-port

  Like `get-impure-port', but issues a POST request (instead of a GET
  request) to a web server using the given post data string.

> (display-pure-port input-port) -> void

  Writes the output of a pure port.  For debugging purposes.

> (purify-port input-port) -> string

  Purifies a port, returning the MIME headers, plus a leading line
  for the form

   HTTP/<vers> <code> <message>

  where <vers> is something like 1.0 or 1.1, <code> is an exact
  integer for the response code, and <message> is arbitrary text
  without a return or newline.

  The "head.ss" library provides procedures, such as `extract-field,
  for manipulating the header.

  Since web server sometimes return mis-formatted replies,
  `purify-port' is liberal in what it accepts as a header. as a
  result, the result string may be ill formed, but it will either be
  the empty string, or it will be a string matching the following
  regexp:

      #rx"^HTTP/.*?((\r\n\r\n)|(\n\n)|(\r\r))"

> (call/input-url url url->port-proc port->X-proc [list-of-strings]) -> X

  First argument is the URL to open.  Second is a procedure that takes
  a URL and turns it into a (pure or impure) port.  The third takes
  the (pure or impure) port and handles its contents.  The optional
  fourth argument is a set of strings to send to the server. The result
  of `call/input-url' is the result of the third argument procedure.

> (current-proxy-servers) -> (list (list scheme string port-number) ...)
> (current-proxy-servers server-mapping) -> void

  The `current-proxy-server' parameter determines a mapping of proxy
  servers used for connections. The mapping is of the form

    (list (list proxiable-scheme server-string port-number) ...)

  where, currently, the only proxiable scheme is "http". An omitted
  mapping indicates that no proxy should be used. The default mapping
  is the empty list (i.e., no proxies).

PRAGMATICS -----------------------------------------------------------

_web documents_

  To access the text of a document from the Web, first obtain its URL
  as a string; convert this into a url structure using string->url;
  then open the document using get-pure-port or get-impure-port,
  depending on whether or not you wish to examine its MIME headers.
  At this point you have a regular input port with which to process
  the document like any other file.  Note that currently the only
  supported protocol is "http".  The "news" protocol is not supported
  directly by this library, but the news library in this collection
  can be used to read NNTP documents.

EXAMPLE --------------------------------------------------------------

 (require (lib "url.ss" "net"))
 (define url:cs (string->url "http://www.cs.rice.edu/"))
 (define (test url)
   (call/input-url url get-pure-port display-pure-port))
 (test url:cs)

UNITS ----------------------------------------------------------------

The _url-sig.ss_ module exports the _net:url^_ signature which
contains the names documented above. the _url-unit.ss_ module exports
_url@_, a unit that imports net:tcp^ (see "tcp-sig.ss", below) and
exports the names in net:url^.

==========================================================================
_URL viewing_
==========================================================================

To load a module for accessing a Web _browser_ either
  (require (lib "external.ss" "browser"))
or
  (require (lib "sendurl.ss" "net"))

The version in "browser" will prompt the user for a browser preference
if not already set.  The version in "sendurl.ss" lacks this capability,
but does not depend on MrEd's graphical toolbox.

These modules define one main function:

> (send-url str [separate-window? #t]) -
  opens the url in a platform-specific manner,
  sending a message to the local browser to display the url named
  by `str'. The separate-window? parameter determines
  if the browser creates a new window to display the url
  or not.

  Under Windows, `send-url' uses the registry to get a
  browser-launching command line. The registry can be read only from
  MrEd (i.e., not MzScheme).

  Under Mac OS, `send-url' uses the URL apple event library.

  Under Unix, `send-url' uses the value of a the _external-browser_
  parameter to select a browser.  The parameter is initialized to the
  value of the _'external-browser preference_, which can be either
  'opera, 'galeon, 'netscape, 'mozilla, 'dillo, #f, or a pair of
  strings.  If #f, `send-url' uses the first of these browsers found
  in the path.  If the parameter is a pair of strings, then the first
  string, the url requested, followed by the second string is
  interpreted as a command line used to launch a browser.  If the
  preferred or default browser can't be launched, `send-url'
  fails. See the file.ss library for details on setting preferences.

> (browser-preference? v) -> bool
  return #t if v is a valid browser preference

==========================================================================
_CGI_ backends, _WWW_
==========================================================================

To load directly: (require (lib "cgi.ss" "net"))
Module files: _cgi.ss_ - provides the procedures documented below
              _cgi-unit.ss_ - provides unit net:cgi@
              _cgi-sig.ss_ - provides signature net:cgi^

ABSTRACT -------------------------------------------------------------

The cgi package helps programmers write scripts that follow the Common
Gateway Interface (CGI/1.1) protocol of the World-Wide Web:
http://hoohoo.ncsa.uiuc.edu/cgi/

INTERACTIONS ---------------------------------------------------------

The CGI library expects to be run in a certain context as defined by
the CGI standard.  This means, for instance, that certain environment
variables will be bound.  Unfortunately, not all CGI environments
provide this.  For instance, the FastCGI library, despite its name,
does not bind the environment variables required of the standard.
Users of FastCGI will need to bind REQUEST_METHOD and possibly also
QUERY_STRING to successfully employ the CGI library.  The FastCGI
library ought to provide a way to extract the values bound to these
variables; the user can then put these into the CGI program's
environment using PLT Scheme's `putenv' function.

TYPES ----------------------------------------------------------------

binding:

  A binding is an association of a form item with its value.  Some form
  items (such as checkboxes) may correspond to multiple bindings.  A
  binding is a tag-string pair, where a tag is a symbol or a string.

bindings:

  A list of `binding's.

html-string:

  A text string that has been escaped according to HTML conventions.

EXCEPTIONS -----------------------------------------------------------

> cgi-error
  struct cgi-error ()

  cgi-error is a super-structure for all exceptions thrown by this
  library.

> incomplete-%-suffix
  struct (incomplete-%-suffix cgi-error) (chars)
  chars : list of chars

  Used when a % in a query is followed by an incomplete suffix.  The
  characters of the suffix -- excluding the "%" -- are provided by the
  exception.

> invalid-%-suffix
  struct (invalid-%-suffix cgi-error) (char)
  char : char

  Used when the character immediately following a % in a query is
  invalid.

PROCEDURES -----------------------------------------------------------

> (get-bindings) -> bindings
> (get-bindings/post) -> bindings
> (get-bindings/get) -> bindings

  Returns the bindings that corresponding to the options specified by
  the user.  The /post and /get forms work only when POST and GET
  forms are used, respectively, while get-bindings determines the kind
  of form that was used and invokes the appropriate function.

> (extract-bindings symbol-or-string bindings) -> list of strings

  Given a key and a set of bindings, extract-bindings determines which
  ones correspond to a given key.  There may be zero, one, or many
  associations for a given key.

> (extract-binding/single symbol-or-string bindings) -> string

  Given a key and a set of bindings, extract-binding/single ensures
  that the key has exactly one association, and returns it.

> (output-http-headers) -> void

  Outputs all the http headers needed for a normal html response.
  Only call this function if you are not using generate-html-output
  or generate-error-output.

> (generate-html-output html-string list-of-html-strings [color color color color color]) -> void

  The first argument is the title.  The second is a list of strings
  that consist of the body.  The last five arguments are each strings
  representing a HTML color; in order, they represent the color of the
  text, the background, un-visited links, visited links, and a link
  being selected.

> (string->html string) -> html-string

  Converts a string into an html-string by applying the appropriate
  HTML quoting conventions.

> (generate-link-text string html-string) -> html-string

  Takes a string representing a URL, a html-string for the anchor
  text, and generates HTML corresponding to an anchor.

> (generate-error-output list-of-html-strings) -> <exit>

  The procedure takes a series of strings representing the body,
  prints them with the subject line "Internal error", and forces the
  script to exit.

> (get-cgi-method) -> string

  Returns either "GET" or "POST".  Always returns a string when
  invoked inside a CGI script.  Unpredictable otherwise.

> (bindings-as-html bindings) -> list of html-strings

  Converts a set of bindings into a list of html-string's.  Useful for
  debugging.

==========================================================================
_sending mail_, _sendmail_
==========================================================================

To load directly: (require (lib "sendmail.ss" "net"))
Module files: _sendmail.ss_ - provides the procedures documented below
              _sendmail-unit.ss_ - provides unit net:sendmail@
              _sendmail-sig.ss_ - provides signature net:sendmail^


ABSTRACT -------------------------------------------------------------

The sendmail package helps programmers write programs that need to
send electronic mail messages.  The package assumes the existence of a
conformant sendmail program on the local system; see also the SMTP
package, below (and, in most cases, use it instead).

TYPES ----------------------------------------------------------------

  All strings used in mail messages are assumed to conform to their
  corresponding SMTP specifications, except as noted otherwise.

EXCEPTIONS -----------------------------------------------------------

> no-mail-recipients
  struct (no-mail-recipients exn) ()

  Raised when no mail recipients were specified.

PROCEDURES -----------------------------------------------------------

> (send-mail-message/port from-string subject-string to-list-of-strings cc-list-of-strings bcc-list-of-string) -> output-port

  The first argument is the header for the sender, the second is the
  subject line, the third a list of To: recipients, the fourth a list
  of CC: recipients, and the fifth a list of BCC: recipients.  The
  optional sixth argument is used for other mail headers, which must
  be specified completely formatted.

  The return value is an output port into which the client must write
  the message.  Clients are urged to use close-output-port on the
  return value as soon as the necessary text has been written, so that
  the sendmail process can complete.

  The sender can hold any value, though of course spoofing should be
  used with care.

> (send-mail-message from-string subject-string to-list-of-strings cc-list-of-strings bcc-list-of-string body-list-of-strings [extra-headers-list-of-strings]) -> void

  The arguments are the same as that for send-mail-message/port except
  that there is one extra input, the list of strings corresponding to
  the mail message (followed by the optional additional headers, if
  present).  There is no interesting return value.

  Lines that contain a single period do not need to be quoted.

==========================================================================
_sending mail_, _SMTP_
==========================================================================

To load directly: (require (lib "smtp.ss" "net"))
Module files: _smtp.ss_ - provides the procedures documented below
              _smtp-unit.ss_ - provides unit net:smtp@
              _smtp-sig.ss_ - provides signature net:smtp^

ABSTRACT -------------------------------------------------------------

The SMTP package helps programmers write programs that need to send
electronic mail messages using SMTP. The client must provide the
address of an SMTP server; in contrast, the mail package (see above)
uses a pre-configured sendmail on the local system.

TYPES ----------------------------------------------------------------

  The head package defines the format of a `header' string, which is
  used by `send-smtp-message'. The head package also provides
  utilities to verify the formatting of a mail address. The procedures
  of the SMTP package assume that the given string arguments are
  well-formed.

EXCEPTIONS -----------------------------------------------------------

  Communication errors are signaled via exn:fail structure instances.

PROCEDURES -----------------------------------------------------------

> (smtp-send-message server-string from-string to-list-of-strings header
                     message-list-of-strings/bytes 
		     [#:port-no k] 
		     [#:auth-user user-string-or-#f]
		     [#:auth-passwd pw-string-or-#f]
		     [#:tcp-connect proc]
		     [port-no]) -> void

  The first argument is the IP address of the SMTP server. The
  `from-string' argument specifies the mail address of the sender, and
  `to-listof-strings' is a list of recipient addresses (including
  "To", "CC", and "BCC" recipients). The `header' argument is the
  complete message header, which should already include "From", "To",
  and "CC" fields consistent with the given sender and recipients.
  the `message-list-of-strings/bytes' argument is the body of the
  message, where each string or byte string in the list corresponds to
  a single line of message text; no string in
  `message-list-of-strings' should contain a carriage return or
  newline characters.

  The optional `port-no' argument --- which can be specified either
  with the #:port-no keyword or, for backward compatibility, as an
  extra argument after keywords --- specifies the IP port to use in
  contacting the SMTP server; the default is 25.

  The optional #:auth-user and #:auth-passwd keyword argument supply a
  username and password for authenticated SMTP (using the AUTH PLAIN
  protocol).

  The optional #:tcp-connect keyword argument supplies a connection
  procedure to be used in place of `tcp-connect'. For example, use
  `ssl-connect' from `(lib "mzssl.ss" "openssl")' to connect to the
  server via SSL.

  See the "head.ss" library for utilities that construct a message
  headers and validate mail address strings.

> (smtp-sending-end-of-message [proc])

  Parameter that determines a send-done procedure to be called after
  `smtp-send-message' has completely sent the message. Before the
  send-done procedure is called, breaking the thread that is executing
  `smtp-send-message' cancels the send. After the send-done procedure
  is called, breaking may or may not cancel the send (and probably
  won't).

==========================================================================
_NNTP_, _newsgroups_
==========================================================================

To load directly: (require (lib "nntp.ss" "net"))
Module files: _nntp.ss_ - provides the procedures documented below
              _nntp-unit.ss_ - provides unit net:nntp@
              _nntp-sig.ss_ - provides signature net:nntp^

ABSTRACT -------------------------------------------------------------

The nntp package helps programmers access Usenet groups via the NNTP
protocols:
http://www.ietf.org/rfc/rfc0977.txt

TYPES ----------------------------------------------------------------

> communicator
  struct communicator (sender receiver server port)
  sender : oport
  receiver : iport
  server : string
  port : number

  Once a connection to a Usenet server has been established, its state
  is stored in a communicator, and other procedures take communicators
  as an argument.

> message-index

  A number or string (the number is common, but a string is useful for
  specifying the entire Message-ID).

> desired

  A regular expression that matches against a Usenet header.

EXCEPTIONS -----------------------------------------------------------

> nntp
  struct (nntp exn) ()

  The super-struct of all subsequent exceptions.

> unexpected-response
  struct (unexpected-response nntp) (code text)
  code : number
  text : string

  Thrown whenever an unexpected response code is received.  The text
  holds the response text sent by the server.

> bad-status-line
  struct (bad-status-line nntp) (line)
  line : string

  Mal-formed status lines.

> premature-close
  struct (premature-close nntp) (communicator)
  communicator : communicator

  Thrown when a remote server closes its connection unexpectedly.

> bad-newsgroup-line
  struct (bad-newsgroup-line nntp) (line)
  line : string

  When the newsgroup line is improperly formatted.

> non-existent-group
  struct (non-existent-group nntp) (group)
  group : string

  When the server does not recognize the name of the requested group.

> article-not-in-group
  struct (article-not-in-group nntp) (article)
  article : number

  When an article is outside the server's range for that group.

> no-group-selected
  struct (no-group-selected nntp) ()

  When an article operation is used before a group has been selected.

> article-not-found
  struct (article-not-found nntp) (article)
  article : number

  When the server is unable to locate the article.

> authentication-rejected
  struct (authentication-rejected nntp) ()

  When the server reject an authentication attempt.

PROCEDURES -----------------------------------------------------------

> (connect-to-server server-string [port-number]) -> communicator

  Connects to the name server.  The second argument, if provided, must
  be a port number; otherwise the default NNTP port is used.

> (disconnect-from-server communicator) -> void

  Disconnects a communicator.

> (open-news-group communicator newsgroup-string) -> three values: number number number

  The second argument is the name of a newsgroup.  The returned values
  are the total number of articles in that group, the first available
  article, and the last available article.

> (authenticate-user communicator username password) -> void

  Tries to authenticate a user.  This is using the original authinfo command
  (uses cleartext).  The password argument is ignored if the server does not
  ask for it.

> (head-of-message communicator message-index) -> list of strings

  Given a message number, returns its headers.

> (body-of-message communicator message-index) -> list of strings

  Given a message number, returns the body of the message.

> (newnews-since communicator message-index) -> list of strings

  Implements the NEWNEWS command (often disabled on servers).

> (generic-message-command communicator message-index) -> list of strings

  Useful primitive for implementing head-of-message, body-of-message
  and other similar commands.

> (make-desired-header tag-string) -> desired

  Takes the header's tag and returns a desired regexp for that header.

> (extract-desired-headers list-of-header-strings list-of-desireds) -> list of strings

  Given a list of headers and of desired's, returns the header lines
  that match any of the desired's.

==========================================================================
_POP-3_, _reading mail_
==========================================================================

To load directly: (require (lib "pop3.ss" "net"))
Module files: _pop3.ss_ - provides the procedures documented below
              _pop3-unit.ss_ - provides unit net:pop3@
              _pop3-sig.ss_ - provides signature net:pop3^

ABSTRACT -------------------------------------------------------------

Implements RFC 1939, Post Office Protocol - Version 3, Myers & Rose.
http://www.ietf.org/rfc/rfc1939.txt

TYPES ----------------------------------------------------------------

> communicator
  struct communicator (sender receiver server port state)
  sender : oport
  receiver : iport
  server : string
  port : number
  state : symbol = (disconnected, authorization, transaction)

  Once a connection to a POP-3 server has been established, its state
  is stored in a communicator, and other procedures take communicators
  as an argument.

> desired

  A regular expression that matches against a mail header.

EXCEPTIONS -----------------------------------------------------------

> pop3
  struct (pop3 exn) ()

  The super-struct used for all other package exceptions.

> cannot-connect
  struct (cannot-connect pop3) ()

  When a connection to a server cannot be established.

> username-rejected
  struct (username-rejected pop3) ()

  If the username is rejected.

> password-rejected
  struct (password-rejected pop3) ()

  If the password is rejected.

> not-ready-for-transaction
  struct (not-ready-for-transaction pop3) (communicator)
  communicator : communicator

  When the communicator is not in transaction mode.

> not-given-headers
  struct (not-given-headers pop3) (communicator message)
  communicator : communicator
  message : number

  When the server does not respond with headers for a message as
  requested.

> illegal-message-number
  struct (illegal-message-number pop3) (communicator message)
  communicator : communicator
  message : number

  When the user specifies an illegal message number.

> cannot-delete-message
  struct (cannot-delete-message exn) (communicator message)
  communicator : communicator
  message : number

  When the server is unable to delete a message.

> disconnect-not-quiet
  struct (disconnect-not-quiet pop3) (communicator)
  communicator : communicator

  When the server does not gracefully disconnect.

> malformed-server-response
  struct (malformed-server-response pop3) (communicator)
  communicator : communicator

  When the server produces a mal-formed response.

PROCEDURES -----------------------------------------------------------

> (connect-to-server server-string [port-number]) -> communicator

  Connects to a server.  Uses the default port number if none is
  provided.

> (disconnect-from-server communicator) -> void

  Disconnects from as server.  Sets the communicator state to
  disconnected.

> (authenticate/plain-text user-string passwd-string communicator) -> void

  Takes a username and password string and, if successful, changes the
  communicator's state to transaction.

> (get-mailbox-status communicator) -> two values: count-number octet-number

  Returns the number of messages and the number of octets.

> (get-message/complete communicator message-number) -> two lists of strings

  Given a message number, returns a list of headers and list of
  strings for the body.

> (get-message/headers communicator message-number) -> list of strings

  Given a message number, returns the list of headers.

> (get-message/body communicator message-number) -> list of strings

  Given a message number, returns the list of strings for the body.

> (delete-message communicator message-number) -> void

  Deletes the specified message.

> (get-unique-id/single communicator message-number) -> string

  Gets the server's unique id for a particular message.

> (get-unique-id/all communicator) -> list of (cons message-number id-string)

  Gets a list of unique id's from the server for all the messages in
  the mailbox.

> (make-desired-header tag-string) -> desired

  Takes the header's tag and returns a desired regexp for that header.

> (extract-desired-headers list-of-strings list-of-desireds) -> list of strings

  Given a list of headers and of desired's, returns the header lines
  that match any of the desired's.

EXAMPLE --------------------------------------------------------------

 > (require (lib "pop3.ss" "net"))
 > (define c (connect-to-server "cs.rice.edu"))
 > (authenticate/plain-text "scheme" "********" c)
 > (get-mailbox-status c)
 196
 816400
 > (get-message/headers c 100)
 ("Date: Thu, 6 Nov 1997 12:34:18 -0600 (CST)"
  "Message-Id: <199711061834.MAA11961@new-world.cs.rice.edu>"
  "From: Shriram Krishnamurthi <shriram@cs.rice.edu>"
  ...
  "Status: RO")
 > (get-message/complete  c 100)
 ("Date: Thu, 6 Nov 1997 12:34:18 -0600 (CST)"
  "Message-Id: <199711061834.MAA11961@new-world.cs.rice.edu>"
  "From: Shriram Krishnamurthi <shriram@cs.rice.edu>"
  ...
  "Status: RO")
 ("some body" "text" "goes" "." "here" "." "")
 > (get-unique-id/single c 205)
 no message numbered 205 available for unique id
 > (list-tail (get-unique-id/all c) 194)
 ((195 . "e24d13c7ef050000") (196 . "3ad2767070050000"))
 > (get-unique-id/single c 196)
 "3ad2767070050000"
 > (disconnect-from-server c)

==========================================================================
_IMAP_, _reading mail_
==========================================================================

To load directly: (require (lib "imap.ss" "net"))
Module files: _imap.ss_ - provides the procedures documented below
              _imap-unit.ss_ - provides unit net:imap@
              _imap-sig.ss_ - provides signature net:imap^

ABSTRACT -------------------------------------------------------------

Implements portions of client-side RFC 2060, Internet Message Access
Protocol - Version 4rev1, Crispin
http://www.ietf.org/rfc/rfc2060.txt

TYPES ----------------------------------------------------------------

> imap

  An opaque record representing an IMAP connection.

> imap-flag

  A symbol, but generally not a convenient one to use within a Scheme
  program. The `imap-flag->symbol' and `symbol->imap-flag' procedures
  convert IMAP flags to convenient symbols and vice-versa.

EXCEPTIONS -----------------------------------------------------------

  Communication errors are signaled via exn:fail structure instances.

PROCEDURES -----------------------------------------------------------

- - - - - - Connecting and selecting mailboxes - - - - -

> (imap-connect server-string username-str/bytes password-str/bytes mailbox-str/bytes)
   -> three values: imap, message count, recent message count

  Establishes an IMAP connection to the given server using the given
  username and password, and selects the specified mailbox. The second
  and third return values indicate the total number of messages in the
  mailbox and the number of recent messages (i.e., messages received
  since the mailbox was last selected), respectively.

  See also `imap-port-number', below.

  A user's primary mailbox is always called "INBOX". (Capitalization
  doesn't matter for that keyword.)

  Updated message-count and recent-count values are available through
  `imap-messages' and `imap-recent'. See also `imap-new?' and
  `imap-reset-new!'.

> (imap-port-number [k])

  A parameter that determines the server port number. The initial
  value is 143.

> (imap-connect* in-port out-port username-str/bytes password-str/bytes mailbox-str/bytes)
   -> three values: imap, message count, recent message count

  Like `imap-connect', but given input and output ports instead of a
  server address.

> (imap-disconnect imap) -> void

  Closes an IMAP connection. The close may fail due to a communication
  error.

> (imap-force-disconnect imap) -> void

  Closes an IMAP connection forcefully (i.e., without send a close
  message to the server). A forced disconnect never fails.

> (imap-reselect imap mailbox-str/bytes)
  -> two values: message count and recent message count

  De-selects the mailbox currently selected by the connection and
  selects the specified mailbox, returning the total and recent
  message counts for the new mailbox. Expunge and message-state
  information is removed.

  Do not use this procedure to poll a mailbox to see whether there are
  any new messages. Use `imap-noop', `imap-new?', and
  `imap-reset-new?' instead.

> (imap-examine imap mailbox-str/bytes)
  -> two values: message count and recent message count

  Like `imap-reselect', but the mailbox is selected as read-only.

- - - - - - Selected mailbox state - - - - -

> (imap-noop imap mailbox-str/bytes)
  -> two values: message count and recent message count

  Sends a "no-op" message to the server, typically to keep the session
  alive. As for many commands, the server may report message-state
  updates or expunges.

  The return information is the same as for `imap-reselect'.

> (imap-poll imap) -> void

  Does not send a request to the server, but checks for asynchronous
  messages from the server that update the message count, to report
  expunges, etc.


> (imap-messages imap) -> number

  Returns the number of messages in the selected mailbox. The server
  can update this count during most any interaction.

  This operation does not communicate with the server. It merely
  reports the result of previous communication.

> (imap-recent imap) -> number

  Returns the number of "recent" messages in the currently selected
  mailbox, as most recently reported by the server. The server can
  update this count during most any interaction.

  This operation does not communicate with the server. It merely
  reports the result of previous communication.

> (imap-unseen imap) -> number-or-false

  Returns the number of "unseen" messages in the currently selected
  mailbox, as most recently reported by the server. The server can
  update this count during most any interaction. Old IMAP servers
  might not report this value, in which case the result is #f.

  This operation does not communicate with the server. It merely
  reports the result of previous communication.

> (imap-uidnext imap) -> number-or-false

  Returns the predicted next uid for a message in the currently
  selected mailbox, as most recently reported by the server. The
  server can update this count during most any interaction. Old IMAP
  servers might not report this value, in which case the result is #f.

  This operation does not communicate with the server. It merely
  reports the result of previous communication.

> (imap-uidvalidity imap) -> number-or-false

  Returns an id number that changes when all uids become invalid. The
  server *cannot* update this number during a session. Old IMAP
  servers might not report this value, in which case the result is #f.

  This operation does not communicate with the server. It merely
  reports the result of previous communication.

> (imap-new? imap) -> boolean

  Returns #t if the server has reported an increase in the message
  count for the currently mailbox since the last call to
  `imap-reset-new!'. Selecting a mailbox implicitly calls
  `imap-reset-new!'.

  This operation does not communicate with the server. It merely
  reports the result of previous communication.

> (imap-reset-new! imap) -> void

  Resets the new flag for the session; see `imap-new?', above.
  This operation does not communicate with the server.

> (imap-get-expunges imap) -> msg-num-list

  Returns pending expunge notifications from the server for the
  selected mailbox in terms of message positions (not uids), and
  clears the pending notifications. The result list is sorted,
  ascending.

  This operation does not communicate with the server. It merely
  reports the result of previous communication.

  The server can notify the client of newly deleted messages during
  most other commands, but not asynchronously between
  commands. Furthermore, the server cannot report new deletions during
  `imap-get-messages' or `imap-store' operations.

  Before calling any IMAP operation that works in terms of message
  numbers, pending expunge notifications must be handled by calling
  `imap-get-expunges'.

> (imap-pending-expunges? imap) -> boolean

  Returns #f if `imap-get-expunges' would return an empty list, #t
  otherwise.

> (imap-get-updates imap) -> list of (cons msg-num assoc-list)

  Returns information must like `imap-get-messages', but includes
  information reported asynchronously by the server (e.g., to notify a
  client with some other client changes a message attribute).  Instead
  of reporting specific requested information for specific messages,
  the result is associates message positions to field-value
  association lists. The result list is sorted by message position,
  ascending.

  This operation does not communicate with the server; it merely
  reports the result of previous communication. It also clears the
  update information from the connection after reporting it.

  When a server reports information that supersedes old reported
  information for a message, or if the server reports that a message
  has been deleted, then old information for the message is
  dropped. Similarly, if `imap-get-messages' is used to explicitly
  obtain information, any redundant (or out-of-date) information is
  dropped.

  A client need not use `imap-get-updates' ever, but accumulated
  information for the connection consumes space.

> (imap-pending-updates? imap) -> boolean

  Returns #f if `imap-get-updates' would return an empty list, #t
  otherwise.

- - - - - - Manipulating messages - - - - -

> (imap-get-messages imap msg-num-list field-list)
  -> list of field-value lists

  Downloads information for a set of messages. The `msg-num-list'
  argument specifies a set of messages by their message positions (not
  their uids). The `field-list' argument specifies the type of
  information to download for each message. The available fields are:

    * 'uid - value is an integer
    * 'header - value is a header (string; see the head package)
    * 'body - value is a byte string (with CRLF-separated lines)
    * 'flags - value is a list of imap flags

  The return value is a list of entry items in parallel to
  `msg-num-list'. Each entry is itself a list containing value items
  in parallel to `field-list'.

  Pending expunges must be handled before calling this function; see
  `imap-get-expunges'.

  Example:
    (imap-get-message imap '(1 3 5) '(uid header))
    ; => ((107 #"From: larry@stooges.com ...")
          (110 #"From: moe@stooges.com ...")
          (112 #"From: curly@stooges.com ..."))

> (imap-flag->symbol imap-flag) -> symbol
> (symbol->imap-flag symbol) -> imap-flag

  An imap flag is a symbol, but it is generally not a convenient one
  to use within a Scheme program, because it usually starts with a
  backslash and flag comparisons are case-insensitive. The
  `imap-flag->symbol' and `symbol->imap-flag' procedures convert IMAP
  flags to convenient symbols and vice-versa:

      symbol       imap flag
      ------       ----------
     'seen         '|\Seen|            \
     'answered     '|\Answered|        |
     'flagged      '|\Flagged|          > message flags
     'deleted      '|\Deleted|         |
     'draft        '|\Draft|           |
     'recent       '|\Recent|          /

     'noinferiors   '|\Noinferiors|    \
     'noselect      '|\Noselect|       |
     'marked        '|\Marked|          > mailbox flags
     'unmarked      '|\Unmarked|       |
     'hasnochildren '|\HasNoChildren|  |
     'haschildren   '|\HasChildren|    /

  `imap-flag->symbol' and `symbol->imap-flag' act like the identity
  function when any other symbol/flag is provided.

> (imap-store imap mode msg-num-list imap-flags) -> void

  Sets flags for a set of messages. The mode argument specifies how
  flags are set:

       * '+ - add the given flags to each message
       * '- - remove the given flags from each message
       * '! - set each message's flags to the given set

  The `msg-num-list' argument specifies a set of messages by their
  message positions (not their uids). The `flags' argument specifies
  the imap flags to add/remove/install.

  Pending expunges must be handled before calling this function; see
  `imap-get-expunges'. The server will not report back message-state
  changes (so they will not show up through `imap-get-updates').

  Example:
    (imap-store imap '+ '(1 2 3) (list (symbol->imap-flag 'deleted)))
    ; marks the first three messages to be deleted
    (imap-expunge imap)
    ; permanently removes the first three messages (and possibly others)
    ;  from the currently-selected mailbox

> (imap-expunge imap) -> void

  Purges every message currently marked with the '|\Deleted| flag from
  the mailbox.

- - - - - - Querying and changing (other) mailboxes - - - - -

> (imap-copy imap msg-num-list dest-mailbox-str/bytes) -> void

  Copies the specified messages from the currently selected mailbox to
  the specified mailbox.

  Pending expunges must be handled before calling this function; see
  `imap-get-expunges'.

> (imap-append imap mailbox-string message-str/bytes) -> void

  Adds a new message (containing message-string) to the given mailbox

> (imap-status imap mailbox-str/bytes status-symbol-list)
  -> list of status values

  Requests information about a mailbox from the server, typically
  *not* the currently selected mailbox.

  The status-symbol-list specifies the request, and the return value
  includes one value for each symbol in status-symbol-list. The
  allowed status symbols are:

    'messages - number of messages
    'recent - number of recent messages
    'unseen - number of unseen messages
    'uidnext - uid for next received message
    'uidvalidity - id that changes when all uids are changed

  Use `imap-messages' to get the message count for the currently
  selected mailbox, etc. Use `imap-new?' and `imap-reset-new!' to
  detect when new messages are available in the currently selected
  mailbox.

> (imap-mailbox-exists? imap mailbox-str/bytes) -> bool

  Returns #t if the specified mailbox exists, #f otherwise.

> (imap-create-mailbox imap mailbox-str/bytes) -> void

  Creates the specified mailbox. (It must not exist already.)

> (imap-list-child-mailboxes imap mailbox-str/bytes [delimiter-str/bytes])
  -> list of mailbox-info lists

  Returns information about sub-mailboxes of the given mailbox.  If
  mailbox-string is #f, information about all top-level mailboxes is
  returned. The optional `delimiter-string' is determined
  automatically (via `imap-get-hierarchy-delimiter') if it is not
  provided.

  The return value is a list of mailbox-information lists. Each
  mailbox-information list contains two items:
    * a list of imap flags for the mailbox
    * the mailbox's name

> (imap-get-hierarchy-delimiter imap) -> bytes

  Returns the server-specific string that is used as a separator in
  mailbox path names.

> (imap-mailbox-flags imap mailbox-str/bytes)
  -> list of imap flags

  Returns a list of imap flags for the given mailbox.

==========================================================================
_mime headers_, _mail headers_, _http headers_
==========================================================================

To load directly: (require (lib "head.ss" "net"))
Module files: _head.ss_ - provides the procedures documented below
              _head-unit.ss_ - provides unit net:head@
              _head-sig.ss_ - provides signature net:head^

ABSTRACT -------------------------------------------------------------

Implements utilities for RFC 822 headers and mail addresses.

TYPES ----------------------------------------------------------------

> header

  A string that is an RFC-882-compliant header. A header string
  contains a series of CRLF-delimited fields, and ends with two CRLFs
  (the first one terminates the last field, and the second terminates
  the header).

  OR

  A byte-string that is an RFC-822-compliant header. Format is as
  prior paragraph.  Only specifically noted functions below
  work with byte-string headers.

PROCEDURES -----------------------------------------------------------

> empty-header

  A string corresponding to the empty header, useful for building up
  headers with `insert-field' and `append-headers'.

> (validate-header candidate-header-string) -> void

  If the format of `candidate-header-string' matches RFC 822, void is
  returned, otherwise an exception is raised.

  This function works when called with a byte-string.

> (extract-field field-string header) -> string or #f

  Returns the header content for the specified field, or #f if the
  field is not in the header. `field-string' should not end with ":",
  and it is used case-insensitively. The returned string will not
  contain the field name, color separator, of CRLF terminator for the
  field; however, if the field spans multiple lines, the CRLFs
  separating the lines will be intact.

  Example:
    (extract-field "TO" (insert-field "to" "me@localhost" empty-header))
    ; => "me@localhost"

  This function works when called with byte-strings for both field-string
  and header.

> (extract-all-fields header) -> (list (cons string string))

  Returns an association-list version of the header; the case of the
  field names is preserved, as well as the order and duplicate uses of
  a field name.

  This function works with a byte-string header. In this case, it returns
  a value of the form (list (cons bytes bytes)).

> (remove-field field-string header) -> header

  Creates a new header by removing the specified field from `header'
  (or the first instance of the field, if it occurs multiple
  times). If the field is not in `header', then the return value is
  `header'.

  This function works when called with byte-strings for both
  field-string and header

> (insert-field field-string value-string header) -> header

  Creates a new header by prefixing the given header with the given
  field-value pair. `value-string' should not contain a terminating
  CRLF, but a multi-line value (perhaps created with
  `data-lines->data') may contain separator CRLFs.

  This function works when all arguments are byte-strings.

> (replace-field field-string value-string header) -> header

  Replaces or inserts the given field. If value-string is #f, then
  the field is simply removed.

  This function works when all arguments are byte-strings.

> (append-headers a-header another-header) -> header

  This function works when both arguments are byte-strings.

> (standard-message-header from-string to-list-of-strings cc-list-of-strings bcc-list-of-strings subject-string) -> header

  Creates a standard mail header given the sender, various lists of
  recipients, a subject, and the current date. (The BCC recipients do
  not actually appear in the header, but they're accepted anyway to
  complete the abstraction.)

> (data-lines->data list-of-strings) -> string

  Merges multiple lines for a single field value into one string,
  adding CRLF-TAB separators.

> (extract-addresses string kind) -> list of strings or
                                     list of list of strings

  Parses `string' as a list of comma-delimited mail addresses, raising
  an exception if the list is ill-formed. This procedure can be used
  for single-address strings, in which case the returned list should
  contain only one address.

  The `kind' argument specifies which portion of an address should be
  returned:

     * 'name - the free-form name in the address, or the address
               itself if no name is available:
        "John Doe <doe@localhost>" => "Jon Doe"
        "doe@localhost (Johnny Doe)" => "Johnny Doe"
        "doe@localhost" => "doe@localhost"

     * 'address - just the mailing address, without any free-form
               names:
        "Jon Doe <doe@localhost>" => "doe@localhost"
        "doe@localhost (Johnny Doe)" => "doe@localhost"
        "doe@localhost" => "doe@localhost"

     * 'full - the full address, essentially as it appears in the
               input, but normalized:
        "Jon Doe   < doe@localhost >" => "Jon Doe <doe@localhost>"
        "  doe@localhost  (Johnny Doe)" => "doe@localhost (Johnny Doe)"
        "doe@localhost" => "doe@localhost"

     * 'all - a list containing each of the three possibilities:
              free-form name, address, and full address (in that
              order)

  Example:
    (extract-addresses " \"Doe, John\" <doe@localhost>, john" 'address)
    ; => ("doe@localhost" "john")

> (assemble-address-field list-of-address-strings) -> string

  Creates a header field value from a list of addresses. The addresses
  are comma-separated, and possibly broken into multiple lines.

==========================================================================
_DNS_, _domain name service_
==========================================================================

To load directly: (require (lib "dns.ss" "net"))
Module files: _dns.ss_ - provides the procedures documented below
              _dns-unit.ss_ - provides unit net:dns@
              _dns-sig.ss_ - provides signature net:dns^

ABSTRACT -------------------------------------------------------------

Implements part of a DNS client, based on RFC 1035. The name cache
does not properly time out entries, however.

Thanks to Eduardo Cavazos and Jason Crowe for repairs and
improvements.

PROCEDURES -----------------------------------------------------------

> (dns-get-address nameserver-string address-string) -> address-string

  Consults the specified nameserver (normally a numerical address like
  "128.42.1.30") to obtain a numerical address for the given Internet
  address.

  The query record sent to the DNS server includes the "recursive"
  bit, but `dns-get-address' also implements a recursive search itself
  in case the server does not provide this optional feature.

> (dns-get-name nameserver-string address-string) -> address-string

  Consults the specified nameserver (normally a numerical address like
  "128.42.1.30") to obtain a name for the given numerical address.

> (dns-get-mail-exchanger nameserver-string address-string) -> address-string

  Consults the specified nameserver to obtain the address for a mail
  exchanger the given mail host address. For example, the mail
  exchanger for "ollie.cs.rice.edu" is currently "cs.rice.edu".

> (dns-find-nameserver) -> address-string or #f

  Attempts to find the address of a nameserver on the present system.
  Under Unix, this procedure parses /etc/resolv.conf to extract the
  first nameserver address. Under Windows, it runs "nslookup.exe".

==========================================================================
_MIME support__
==========================================================================

To load directly: (require (lib "mime.ss" "net"))
Module files: _mime.ss_ - provides the procedures documented below
              _mime-unit.ss_ - provides unit net:mime@
                               imports net:base64^ from "base64-sig.ss"
                               and net:qp^ from "qp-sig.ss"
              _mime-sig.ss_ - provides signature net:mime^

ABSTRACT -------------------------------------------------------------

Author: Francisco Solsona <solsona@acm.org>

MIME stands for Multipurpose Internet Mail Extensions. the mime
library provides support for MIME as described in: rfc2045 through
rfc2049.

This library is supposed to be used by PLT Scheme applications, or
libraries to make the MIME-aware.  For instance, MUAs, (Fast)CGI
libraries to support multipart/form-data extensions, etc.

EXCEPTIONS -----------------------------------------------------------

   unexpected-termination

   This exception is raised when eof is reached while parsing the
   headers of a MIME entity.  And it usually means, that the message
   does not conform to rfc2045, and friends. It has a message
   parameter, `msg', that gives exact information on the problem:
   unexpected-termination-msg.

   missing-multipart-boundary-parameter

   A multipart type is signaled, but no `boundary' parameter is
   given, also an error on the producer end of the message.

   malformed-multipart-entity

   Similar to `unexpected-termination', but it is used only while
   scanning parts of a multipart message.

   empty-mechanism

   No transport encoding mechanism was provided with the
   Content-Transfer-Encoding field.

   empty-type

   No type specified for Content-Type, or incorrectly formated.

   empty-subtype

   No subtype specified for Content-Type, or incorrectly formated.

   empty-disposition-type

   No type specified for the Content-Disposition field, or incorrectly
   formated.


   mime-error

   A catch-all exception.

PROCEDURES -----------------------------------------------------------

> (mime-analyze message)

   Message must conform to the RFC 2045, 2046, 2047, 2048, and 2049.

   This is basically the only function provided by the MIME support
   library.  It returns a `message' structure, having the following
   fields: version, entity, and fields.

> (message-version message-struct)

   the value of the MIME-version header.  It has a default 1.0 value
   when message does not contain a MIME-version header.

> (message-entity message-struct)

   another structure, `entity'.  According to rfc2045, and entity
   refers to the MIME-defined header fields and contents of either a
   message or one of the parts in the body of a multipart entity.

   See below, for the fields of an entity-structure.  All the
   MIME-headers, parts (sub-entities), and/or contents of message are
   stored into this structure.

> (message-fields message-struct)

   a list of all non-MIME headers in message.  If message is an
   rfc822, for instance, this list could contain 'From', 'Subject',
   etc.

struct entity
-------------

   These kind of structure contains all the headers, parameters, and
   contents of a given entity, and its parts (sub-entities).  These
   are the fields of the entity structure:

   type, subtype, charset, encoding, disposition, params, id,
   description, other, fields, parts, body, close.

> (entity-type entity-struct)

   In rfc822 BNF extension:

   type := discrete-type | composite-type

   discrete-type := "text" / "image" / "audio" / "video" /
                        "application" / extension-token

   composite-type := "message" / "multipart" / extension-token

   extension-token := ietf-token / x-token

   ietf-token := <An extension token defined by a standards-track
                  RFC and registered with IANA.>

   x-token := <The two characters "X-" or "x-" followed, with no
               intervening white space, by any token>

   In English, it returns one of the symbols: text, image, audio,
   video, application, an ietf-extension, and x-token, or a local
   extension (not recommended by the way.)

> (entity-subtype entity-struct)

   subtype := extension-token / iana-token

   Also a symbol, it depends on the value of type.  Quoting RFC 1700:

 Type            Subtype         Description                 Reference
 ----            -------         -----------                 ---------
 text            plain                                   [RFC1521,NSB]
                 richtext                                [RFC1521,NSB]
                 tab-separated-values                   [Paul Lindner]

 multipart       mixed                                   [RFC1521,NSB]
                 alternative                             [RFC1521,NSB]
                 digest                                  [RFC1521,NSB]
                 parallel                                [RFC1521,NSB]
                 appledouble                [MacMime,Patrik Faltstrom]
                 header-set                             [Dave Crocker]

 message         rfc822                                  [RFC1521,NSB]
                 partial                                 [RFC1521,NSB]
                 external-body                           [RFC1521,NSB]
                 news                        [RFC 1036, Henry Spencer]

 application     octet-stream                            [RFC1521,NSB]
                 postscript                              [RFC1521,NSB]
                 oda                                     [RFC1521,NSB]
                 atomicmail                           [atomicmail,NSB]
                 andrew-inset                       [andrew-inset,NSB]
                 slate                           [slate,terry crowley]
                 wita              [Wang Info Transfer,Larry Campbell]
                 dec-dx            [Digital Doc Trans, Larry Campbell]
                 dca-rft        [IBM Doc Content Arch, Larry Campbell]
                 activemessage                          [Ehud Shapiro]
                 rtf                                    [Paul Lindner]
                 applefile                  [MacMime,Patrik Faltstrom]
                 mac-binhex40               [MacMime,Patrik Faltstrom]
                 news-message-id              [RFC1036, Henry Spencer]
                 news-transmission            [RFC1036, Henry Spencer]
                 wordperfect5.1                         [Paul Lindner]
                 pdf                                    [Paul Lindner]
                 zip                                    [Paul Lindner]
                 macwriteii                             [Paul Lindner]
                 msword                                 [Paul Lindner]
                 remote-printing                         [RFC1486,MTR]

 image           jpeg                                    [RFC1521,NSB]
                 gif                                     [RFC1521,NSB]
                 ief             Image Exchange Format       [RFC1314]
                 tiff            Tag Image File Format           [MTR]

 audio           basic                                   [RFC1521,NSB]

 video           mpeg                                    [RFC1521,NSB]
                 quicktime                              [Paul Lindner]


> (entity-charset entity-struct)

   The value of the `charset' parameter of the Content-Type header.
   By default, the symbol: us-ascii.

> (entity-encoding entity-struct)

   The value of the Content-Transfer-Encoding header, one of the
   symbols: 7bit, 8bit, binary, quoted-printable, base64.  If no
   Content-Transfer-Encoding header is given, message should be 7bit,
   thus, 7bit is the default value of this field.

> (entity-disposition entity-struct)

   This one is a litter bit complex, so it is yet another structure.
   These `disposition' structures represent the value of the
   Content-Disposition header, as defined in rfc2183:

  disposition := "Content-Disposition" ":"
                  disposition-type
                  *(";" disposition-parm)

   disposition structs have the following fields: type, filename,
   creation, modification, read, size, and params.  Only type is
   required.

> (disposition-type disposition-struct)

   disposition-type := 'inline / 'attachment  / extension-token

> (disposition-filename disposition-struct)

   string representing the `filename' parameter of the
   Content-Disposition header.

> (disposition-creation disposition-struct)

   string representing the `creation-date' parameter, which must
   follow this BNF grammar:

 date-time   =  [ day "," ] date time        ; dd mm yy
                                   ;  hh:mm:ss zzz

 day         =  "Mon"  / "Tue" /  "Wed"  / "Thu"
            /  "Fri"  / "Sat" /  "Sun"

 date        =  1*2DIGIT month 2DIGIT        ; day month year
                                   ;  e.g. 20 Jun 82

 month       =  "Jan"  /  "Feb" /  "Mar"  /  "Apr"
            /  "May"  /  "Jun" /  "Jul"  /  "Aug"
            /  "Sep"  /  "Oct" /  "Nov"  /  "Dec"

 time        =  hour zone                    ; ANSI and Military

 hour        =  2DIGIT ":" 2DIGIT [":" 2DIGIT]
                                   ; 00:00:00 - 23:59:59

 zone        =  "UT"  / "GMT"                ; Universal Time
                                   ; North American : UT
            /  "EST" / "EDT"                ;  Eastern:  - 5/ - 4
            /  "CST" / "CDT"                ;  Central:  - 6/ - 5
            /  "MST" / "MDT"                ;  Mountain: - 7/ - 6
            /  "PST" / "PDT"                ;  Pacific:  - 8/ - 7
            /  1ALPHA                       ; Military: Z = UT;
                                   ;  A:-1; (J not used)
                                   ;  M:-12; N:+1; Y:+12
            / ( ("+" / "-") 4DIGIT )        ; Local differential

> (disposition-modification disposition-struct)

   String representing the `modification-date' parameter, it has the
   same syntax as `disposition-creation'.

> (disposition-read disposition-struct)

   String representing the `read-date' parameter, it has the same
   syntax as `disposition-creation'.

> (disposition-size disposition-struct)

   Number representing the `size' parameter (the size of the file).

> (disposition-params disposition-struct)

   List of extra parameters (non-standard) following the
   Content-Disposition header.

> (entity-params entity-struct)

   List of `parameters' given with all the MIME headers above.  Each
   of these parameters is a pair of the form: (parameter-name . value)

> (entity-id entity-struct)

   The value of the Content-Id header.

> (entity-description entity-struct)

   The value of the Content-description header.

> (entity-other entity-struct)

   MIME extension fields (non-standard) they all start with
   "Content-".

> (entity-fields entity-struct)

   A list of all non-MIME headers.

> (entity-parts entity-struct)

   A list of `message'-structs see above.  Each entry represents a
   part of the current MIME entity (only multipart, or message
   entities can have parts).

> (entity-body entity-struct)

   A procedure that consumes an output port and writes the message
   body into the port, properly decoded (e.g. if body was
   quoted-printable encoded, then the body is written already
   quoted-printable decoded.).  If type is multipart, or message,
   clearly no bytes are written to port. The procedure only works
   once (since the encoded body is pulled from a stream).

==========================================================================
_Base 64 Encoding_, _base64_
==========================================================================

To load directly: (require (lib "base64.ss" "net"))
Module files: _base64.ss_ - provides the procedures documented below
              _base64-unit.ss_ - provides unit net:base64@
              _base64-sig.ss_ - provides signature net:base64^

ABSTRACT -------------------------------------------------------------

Implements Base 64 (mime-standard) encoder and decoder.

PROCEDURES -----------------------------------------------------------

> (base64-encode bytes) -> bytes

  Consumes a byte string and returns its Base 64 encoding as a new
  byte string.  The returned string is broken into 72-byte lines
  separated by CRLF combinations, and always ends with a CRLF
  combination unless the input is empty.

> (base64-decode bytes) -> bytes

  Consumes a byte string and returns its Base 64 decoding as a new
  byte string.

> (base64-encode-stream input-port output-port [newline-bytes]) -> void

  Reads bytes from `input-port' and writes the encoded result to
  `output-port', breaking the output into 72-character lines separated
  by `newline-bytes, and ending with `newline-bytes unless the
  input stream is empty. The default `newline-bytes contains just a
  newline (not CRLF). The procedure returns when it encounters an
  end-of-file from `input-port'.

> (base64-decode-stream input-port output-port) -> void

  Reads a Base 64 encoding from `input-port' and writes the decoded
  result to `output-port'. The procedure returns when it encounters an
  end-of-file or Base 64 terminator (=) from `input-port'.

==========================================================================
_Quoted Printable Encoding_, _qp__
==========================================================================

To load directly: (require (lib "qp.ss" "net"))
Module files: _qp.ss_ - provides the procedures documented below
              _qp-unit.ss_ - provides unit net:qp@
              _qp-sig.ss_ - provides signature net:qp^

ABSTRACT -------------------------------------------------------------

Author: Francisco Solsona <solsona@acm.org>

Implements Quoted Printable (mime-standard) encoder and decoder. As
proposed by the rfc2045 section 6.7.

EXCEPTIONS -----------------------------------------------------------


  qp-wrong-input

  This exception is raised when input to any of the qp-procedures (see
  below) is neither a string, nor an input port.

  qp-wrong-line-size

  If the length of a supposedly quoted-printable line is longer than
  76 this exception is raised.

  qp-error

  is a catch all exception.

PROCEDURES -----------------------------------------------------------


> (qp-encode string) -> string

  Consumes a string and returns its quoted printable representation as
  a new string. The encoded string uses "\r\n" where necessary to
  create shorter lines.

> (qp-decode string) -> string

  Consumes a string and returns its un-quoted printable representation
  as a new string. Non-soft linebreaks are preserved in whatever form
  they exist (CR, LR, or CRLF) in the input string.

> (qp-encode-stream input output [newline-string]) -> void

  Reads characters from `input' and writes the quoted printable
  encoded result to `output'.

  The `newline-string' argument is used for soft line-breaks (after
  "="). The default `newline-string' is #"\n", though it should
  probably be #"\r\n" instead.

  Other linebreaks are preserved in whatever form they exist (CR, LR,
  or CRLF) in the input stream.

> (qp-decode-stream input output) -> void

   Reads characters from `input' and writes de-quoted-printable result
   to `output'.  Non-soft linebreaks are preserved in whatever form
   they exist (CR, LR, or CRLF) in the input stream.


==========================================================================
_FTP client_
==========================================================================

To load directly: (require (lib "ftp.ss" "net"))
Module files: _ftp.ss_ - provides the procedures documented below
              _ftp-unit.ss_ - provides unit net:ftp@
              _ftp-sig.ss_ - provides signature net:ftp^

ABSTRACT -------------------------------------------------------------

Author: Micah Flatt <mflatt@elephant-ear.com>

Implements FTP client actions.

PROCEDURES -----------------------------------------------------------

> (ftp-establish-connection server-host-string server-port-k user-string passwd-string)

  Establishes an FTP connection with the given server using the
  supplied username and password. The result is an opaque `ftp-conn'
  structure representing the connection.

  The username and password strings are encoded to bytes using the
  current locale's encoding.

> (ftp-close-connection ftp-conn)

  Closes an FTP connection.

> (ftp-cd ftp-conn new-dir-string)

  Changes the current directory on the FTP server to `new-dir-string'.
  The `new-dir-string' is not interpreted at all, but simply passed on
  to the server (encoded using the current locale's encoding). It must
  not contain a newline.

> (ftp-directory-list ftp-conn)

  Returns a list of files and directories in the current directory of
  the server, assuming that the server provides directory information
  in the quasi-standard Unix format.

  Each file or directory is represented by a list of three
  strings. The first string is either "-", "d", or "l", depending on
  whether the items is a file, directory, or link, respectively.  The
  second item is the file's date; to convert this value to seconds
  consistent with `file-seconds', pass the date string to
  `ftp-make-file-seconds', below.  The third string is the name of the
  file or directory.

  All strings are decoded from bytes using the current locale's
  encoding.

> (ftp-make-file-seconds ftp-date-string)

  Takes a date string produced by `ftp-directory-list' and converts it
  to seconds.

> (ftp-download-file ftp-conn local-directory-path filename-string)

  Downloads `filename-string' from the server's current directory and
  puts it in `local-directory-path' using the same name. If the file
  already exists in the local directory, it is replaced, but only
  after the transfer succeeds (i.e., the file is first downloaded to a
  temporary file, then moved into place on success).

==========================================================================
_TCP redirect_
==========================================================================

The "tcp-redirect.ss" library provides an function to generate a unit
with the signature net:tcp^ that redirects certain port numbers to
intra-process listeners that do not actually use TCP. The net:tcp^
signature is imported, for example, by the url@ unit of "url-unit.ss".

Module file: _tcp-redirect.ss_ - provides the `tcp-redirect-function
             _tcp-sig.ss_ - defines _net:tcp^_, which contains the
                            following:
                                 tcp-abandon-port
                                 tcp-accept
                                 tcp-accept-ready?
                                 tcp-addresses
                                 tcp-close
                                 tcp-connect
                                 tcp-connect/enable-break
                                 tcp-listen
                                 tcp-listener?
             _tcp-unit.ss_ - defines _tcp@_ which implements
                             net:tcp^ using the MzScheme functions of
                             the same name
             ssl-tcp-unit.ss - see below

PROCEDURES -----------------------------------------------------------

> (tcp-redirect port-number-list)

  Returns a unit that implements net:tcp^ (from "tcp-sig.ss"). For
  port numbers not listed in `port-number-list', the unit's
  implementations are the MzScheme implementations.

  For the port numbers listed in `port-number-list', and for
  connections to "127.0.0.1", the unit's implementation does not use
  TCP connections, but instead uses internal buffered channels. Such
  channels behave exactly as TCP listeners and ports.

==========================================================================
_SSL redirect__
==========================================================================

The _ssl-tcp-unit.ss_ library provides `make-ssl-tcp@', which returns
a unit that implements net:tcp^ from "tcp-sig.ss"

PROCEDURES -----------------------------------------------------------

> (make-ssl-tcp@
   server-cert-file server-key-file server-root-cert-files
   server-suggest-auth-file
   client-cert-file client-key-file client-root-cert-files)

  Returns a unit that implements net:tcp^ using the SSL functions from
  (lib "mzssl.ss" "openssl"). The arguments to `make-ssl-tcp@' control
  the certificates and keys uses by server and client connections:

    `server-cert-file' - a PEM file for a server's certificate; #f means
     no certificate (which is unlikely to work with any SSL client)
    `server-key-file' - a private key PEM to go with `server-cert-file';
     #f means no key (which is likely renders a certificate useless)
    `server-root-cert-files' - a list of PEM files for trusted root
     certificates; #f disables verification of peer client certificates

    `server-suggest-auth-file' - PEM file for root certificates to be
     suggested to peer clients that must supply certificates

    `client-cert-file' - a PEM file for a client's certificate; #f means
     no certificate (usually fine)
    `client-key-file' - a private key PEM to go with `client-cert-file';
     #f means no key (which is likely renders a certificate useless)
    `client-root-cert-files' - a list of PEM files for trusted root
     certificates; #f disables verification of peer server certificates

==========================================================================
_cookies_, HTTP _State Management_
==========================================================================

To load directly: (require (lib "cookie.ss" "net"))
Module files: _cookie.ss_ - provides the procedures documented below
              _cookie-unit.ss_ - provides unit net:cookie@
              _cookie-sig.ss_ - provides signature net:cookie^


ABSTRACT -------------------------------------------------------------

The cookie package helps programmers use cookies as specified in RFC
2109 HTTP State Management Mechanism, aka HTTP cookies:
http://www.w3.org/Protocols/rfc2109/rfc2109

EXCEPTIONS -----------------------------------------------------------

> cookie-error
  struct (cookie-error exn) ()

  All errors are signaled by raising a cookie-error

TYPES ----------------------------------------------------------------

> cookie

  The opaque structure that represents a cookie

PROCEDURES -----------------------------------------------------------

> (set-cookie name value)
  : String String -> cookie

  Creates a new cookie, with default values for required fields:

> (cookie:add-comment cookie comment)
  : cookie String -> cookie

> (cookie:add-domain cookie domain)
  : cookie String -> cookie

> (cookie:add-max-age cookie seconds)
  : cookie Nat -> cookie

> (cookie:add-path cookie path)
  : cookie String -> cookie

> (cookie:secure cookie secure?)
  : cookie Boolean -> cookie

> (cookie:version cookie version)
  : cookie Nat -> cookie

  These procedures all modify the fields of an existing cookie.  The
  modification is destructive, and the modified cookie is returned.

  `domain' must match a prefix of the request-URI.

  `seconds' should be a non-negative natural number representing the
  number of seconds you want the client to retain the cookie.

  `version' is a required field; the library defaults to the only
  known incarnation of HTTP cookies: 1.

> (print-cookie cookie)
  : cookie-structure -> String

  Printing a cookie obtains the string that represents the cookie you
  formed with the procedures above.

  Empty fields will not appear in the output except when there is a
  required default.  Currently only the Version is always added.

> (get-cookie name cookies)
  : String String -> (listof String)

  Returns a list with all the values (strings) associated with name.

> (get-cookie/single name cookies)
  : String String -> (U String #f)

  Returns the `first' value string associated to name, or #f if no
  association is found.

  The method used to obtain the `Cookie' header depends on the web
  server.  It may be an environment variable (CGI), or you may have to
  read it from the `input port' (FastCGI), or maybe it comes in an
  `initial-request' structure, etc.  The above two procedures can be
  used to extract fields from a `Cookie' header.


EXAMPLES -------------------------------------------------------------

> Creating a cookie

(let ((c (cookie:add-max-age
          (cookie:add-path
           (set-cookie "foo" "bar")
           "/servlets")
          3600)))
  (print-cookie c))
=>
"foo=bar; Max-Age=3600; Path=/servlets; Version=1"

To use this output in a "regular" CGI, instead of the last line you
would use:

  (display (format "Set-Cookie: ~a" (print-cookie c)))

and to use with the PLT Web Server, you would use:

  (make-response/full code message (current-seconds) mime
                      `((Set-Cookie . ,(print-cookie c)))
                      body)

> Parsing a cookie

Imagine your Cookie header looks like this:

(cookies . "test2=2; test3=3; xfcTheme=theme6; xfcTheme=theme2")

Then, to get the values of the xfcTheme cookie, you would use:

  (get-cookie "xfcTheme" cookies)
=> ("theme6" "theme2")
  (get-cookie/single "xfcTheme" cookies)
=> "theme6"

If you try to get a cookie that simply is not there, you will get:

 (get-cookie/single "foo" cookies)
=> #f
 (get-cookie "foo" cookies)
=> ()

Note that not having a cookie is normally not an error.  Most clients
won't have a cookie set then first arrive at your site.

==========================================================================
_URL encoding_, _URL decoding_, _application/x-www-form-urlencoded_
==========================================================================

To load directly: (require (lib "uri-codec.ss" "net"))
Module files: _uri-codec.ss_ - provides the procedures documented below
              _uri-codec-unit.ss_ - provides unit net:uri-codec@
              _uri-codec-sig.ss_ - provides signature net:uri-codec^


ABSTRACT -------------------------------------------------------------

The module provides functions to encode and decode strings using
the URI encoding rules given in RFC 2396, and to encode and decode
name/value pairs using the application/x-www-form-urlencoded
mimetype given the in HTML 4.0 specification.  There are minor
differences between the two encodings.

The URI encoding uses allows a few characters to be represented `as
is': a-Z, A-Z, 0-9, -, _, ., !, ~, *, ', ( and ).  The remaining
characters are encoded as %xx, where xx is the hex representation
of the integer value of the character (where the mapping
character<->integer is determined by US-ASCII if the integer is
<128).

The encoding, inline with RFC 2396's recommendation, represents a
character as is, if possible.  The decoding allows any characters
to be represented by their hex values, and allows characters to be
incorrectly represented `as is'.

The rules for the application/x-www-form-urlencoded mimetype given
in the HTML 4.0 spec are:

  1. Control names and values are escaped. Space characters are
  replaced by `+', and then reserved characters are escaped as
  described in [RFC1738], section 2.2: Non-alphanumeric characters
  are replaced by `%HH', a percent sign and two hexadecimal digits
  representing the ASCII code of the character. Line breaks are
  represented as "CR LF" pairs (i.e., `%0D%0A').

  2. The control names/values are listed in the order they appear in
  the document. The name is separated from the value by `=' and
  name/value pairs are separated from each other by either `;' or `&'.
  When encoding, `;' is used as the separator by default. When
  decoding, both `;' and `&' are parsed as separators by default.

NB: RFC 2396 supersedes RFC 1738.

This differs slightly from the straight encoding in RFC 2396 in
that `+' is allowed, and represents a space.  We follow this
convention, encoding a space as `+' and decoding `+' as a space.
There appear to be some brain-dead decoders on the web, so we also
encode `!', `~', `'', `(' and ')' using their hex representation.
This is the same choice as made by the Java URLEncoder.

TYPES ----------------------------------------------------------------

> alist

  (list-of (cons symbol string))  Represents decoded name/value pairs

PROCEDURES -----------------------------------------------------------

> (uri-encode string)
  : String -> String

  Encode a string using the URI encoding rules

> (uri-decode string)
  : String -> String

  Decode a string encoded using the URI encoding rules

> (form-urlencoded-encode string)
  : String -> String

  Encode a string using the application/x-www-form-urlencoded encoding
  rules. (This string contains no non-ASCII characters, of course.)

> (form-urlencoded-decode string)
  : String -> String

  Decode a string encoded using the application/x-www-form-urlencoded
  encoding rules.

> (alist->form-urlencoded alist)
  : alist -> String

  Encode an alist of key/value pairs using the
  application/x-www-form-urlencoded encoding rules.

  The `separator-mode-sym' argument must be either 'amp or 'semi to
  select the separator. The default is 'semi.
 

> (form-urlencoded->alist string [separator-mode-sym])
  : String -> alist

  Decode a string encoded using the application/x-www-form-urlencoded
  encoding rules into an alist of key/value pairs. All keys are
  downcased for conversion to symbols.

  The `separator-mode-sym' argument must be either 'amp, 'semi, or
  'both to select the separator. The default is 'both.

> (current-alist-separator-mode)
> (current-alist-separator-mode mode-sym)

  A parameter that determines the separator used/recognized between
  alist pairs in `form-urlencoded->alist', `alist->form-urlencoded',
  `url->string', and `string->url'. The possible mode symbols are
  'amp, 'semi, and 'amp-or-semi.

  The default value is 'amp-or-semi, which means that both "&" and ";"
  are treated as separators when parsing, and ";" is used as a
  separator when encoding. The other modes use/recognize only of the
  separators.

  Examples for '((name . "shriram") (host "nw")):

    Mode                     Parse                    Generate
   ------             --------------------      --------------------

   'amp               name=shriram&host=nw      name=shriram&host=nw

   'semi              name=shriram;host=nw      name=shriram;host=nw

   'amp-or-semi       name=shriram&host=nw      name=shriram;host=nw
                               or
                      name=shriram;host=nw
