tcl/doc/http.n

'\"
'\" Copyright (c) 1995-1997 Sun Microsystems, Inc.
'\" Copyright (c) 1998-2000 Ajuba Solutions.
'\" Copyright (c) 2004 ActiveState Corporation.
'\"
'\" See the file "license.terms" for information on usage and redistribution
'\" of this file, and for a DISCLAIMER OF ALL WARRANTIES.
'\"
.TH "http" n 2.10 http "Tcl Bundled Packages"
.so man.macros
.BS
'\" Note:  do not modify the .SH NAME line immediately below!
.SH NAME
http \- Client-side implementation of the HTTP/1.1 protocol
.SH SYNOPSIS
.nf
\fBpackage require http\fR ?\fB2.10\fR?
.\" See Also -useragent option documentation in body!

\fB::http::config\fR ?\fI\-option value ...\fR?
\fB::http::geturl \fIurl\fR ?\fI\-option value ...\fR?
\fB::http::formatQuery\fI key value\fR ?\fIkey value ...\fR?
\fB::http::quoteString\fI value\fR
\fB::http::reset\fI token\fR ?\fIwhy\fR?
\fB::http::wait \fItoken\fR
\fB::http::status \fItoken\fR
\fB::http::size \fItoken\fR
\fB::http::error \fItoken\fR
\fB::http::postError \fItoken\fR
\fB::http::cleanup \fItoken\fR
\fB::http::requestLine\fI token\fR
\fB::http::requestHeaders\fI token\fR ?\fIheaderName\fR?
\fB::http::requestHeaderValue\fI token headerName\fR
\fB::http::responseLine\fI token\fR
\fB::http::responseCode\fI token\fR
\fB::http::reasonPhrase\fI code\fR
\fB::http::responseHeaders\fI token\fR ?\fIheaderName\fR?
\fB::http::responseHeaderValue\fI token headerName\fR
\fB::http::responseInfo\fI token\fR
\fB::http::responseBody\fI token\fR
\fB::http::register \fIproto port command\fR ?\fIsocketCmdVarName\fR? ?\fIuseSockThread\fR? ?\fIendToEndProxy\fR?
\fB::http::registerError \fIsock\fR ?\fImessage\fR?
\fB::http::unregister \fIproto\fR
\fB::http::code \fItoken\fR
\fB::http::data \fItoken\fR
\fB::http::meta \fItoken\fR ?\fIheaderName\fR?
\fB::http::metaValue\fI token headerName\fR
\fB::http::ncode \fItoken\fR
.fi
.SH "EXPORTED COMMANDS"
.PP
Namespace \fBhttp\fR exports the commands \fBconfig\fR, \fBformatQuery\fR,
\fBgeturl\fR, \fBpostError\fR, \fBquoteString\fR, \fBreasonPhrase\fR,
\fBregister\fR,
\fBregisterError\fR, \fBrequestHeaders\fR, \fBrequestHeaderValue\fR,
\fBrequestLine\fR, \fBresponseBody\fR, \fBresponseCode\fR,
\fBresponseHeaders\fR, \fBresponseHeaderValue\fR, \fBresponseInfo\fR,
\fBresponseLine\fR,
\fBreset\fR, \fBunregister\fR, and \fBwait\fR.
.PP
It does not export the commands \fBcleanup\fR, \fBcode\fR, \fBdata\fR,
\fBerror\fR, \fBmeta\fR, \fBmetaValue\fR, \fBncode\fR,
\fBsize\fR, or \fBstatus\fR.
.BE
.SH DESCRIPTION
.PP
The \fBhttp\fR package provides the client side of the HTTP/1.1
protocol, as defined in RFC 9110 to 9112, which supersede RFC 7230
to RFC 7235, which in turn supersede RFC 2616.
The package implements the GET, POST, and HEAD operations
of HTTP/1.1.  It allows configuration of a proxy host to get through
firewalls.  The package is compatible with the \fBSafesock\fR security
policy, so it can be used by untrusted applets to do URL fetching from
a restricted set of hosts. This package can be extended to support
additional HTTP transport protocols, such as HTTPS, by providing
a custom \fBsocket\fR command, via \fB::http::register\fR.
.PP
The \fB::http::geturl\fR procedure does a HTTP transaction.
Its \fIoptions \fR determine whether a GET, POST, or HEAD transaction
is performed.
The return value of \fB::http::geturl\fR is a token for the transaction.
The token can be supplied as an argument to other commands, to manage the
transaction and examine its results.
.PP
If the \fB\-command\fR option is specified, then
the HTTP operation is done in the background.
\fB::http::geturl\fR returns immediately after generating the
HTTP request and the \fB\-command\fR callback is invoked
when the transaction completes.  For this to work, the Tcl event loop
must be active.  In Tk applications this is always true.  For pure-Tcl
applications, the caller can use \fB::http::wait\fR after calling
\fB::http::geturl\fR to start the event loop.
.PP
\fBNote:\fR The event queue is even used without the \fB\-command\fR option.
As a side effect, arbitrary commands may be processed while \fBhttp::geturl\fR
is running.
.PP
When the HTTP server has replied to the request, call the command
\fB::http::responseInfo\fR, which
returns a \fBdict\fR of metadata that is essential for identifying a
successful transaction and making use of the response.  See
section \fBMETADATA\fR for details of the information returned.
The response itself is returned by command \fB::http::responseBody\fR,
unless it has been redirected to a file by the \fI\-channel\fR option
of \fB::http::geturl\fR.
.SH COMMANDS
.\" COMMAND: config
.TP
\fB::http::config\fR ?\fIoptions\fR?
.
The \fB::http::config\fR command is used to set and query the name of the
proxy server and port, and the User-Agent name used in the HTTP
requests.  If no options are specified, then the current configuration
is returned.  If a single argument is specified, then it should be one
of the flags described below.  In this case the current value of
that setting is returned.  Otherwise, the options should be a set of
flags and values that define the configuration:
.RS
.\" OPTION: -accept
.TP
\fB\-accept\fI mimetypes\fR
.
The Accept header of the request.  The default is */*, which means that
all types of documents are accepted.  Otherwise you can supply a
comma-separated list of mime type patterns that you are
willing to receive.  For example,
.QW "image/gif, image/jpeg, text/*" .
.\" OPTION: -cookiejar
.TP
\fB\-cookiejar\fI command\fR
.VS TIP406
The cookie store for the package to use to manage HTTP cookies.
\fIcommand\fR is a command prefix list; if the empty list (the
default value) is used, no cookies will be sent by requests or stored
from responses. The command indicated by \fIcommand\fR, if supplied,
must obey the \fBCOOKIE JAR PROTOCOL\fR described below.
.VE TIP406
.\" OPTION: -pipeline
.TP
\fB\-pipeline\fI boolean\fR
.
Specifies whether HTTP/1.1 transactions on a persistent socket will be
pipelined.  See the \fBPERSISTENT SOCKETS\fR section for details. The default
is 1.
.\" OPTION: -postfresh
.TP
\fB\-postfresh\fI boolean\fR
.
Specifies whether requests that use the \fBPOST\fR method will always use a
fresh socket, overriding the \fB\-keepalive\fR option of
command \fBhttp::geturl\fR.  See the \fBPERSISTENT SOCKETS\fR section for
details. The default is 0.
.\" OPTION: -proxyauth
.TP
\fB\-proxyauth\fI string\fR
.
If non-empty, the string is supplied to the proxy server as the value of the
request header Proxy-Authorization.  This option can be used for HTTP Basic
Authentication.  If the proxy server requires authentication by another
technique, e.g. Digest Authentication, the \fB\-proxyauth\fR option is not
useful.  In that case the caller must expect a 407 response from the proxy,
compute the authentication value to be supplied, and use the \fB\-headers\fR
option to supply it as the value of the Proxy-Authorization header.
.\" OPTION: -proxyfilter
.TP
\fB\-proxyfilter\fI command\fR
.
The command is a callback that is made during
\fB::http::geturl\fR
to determine if a proxy is required for a given host.  One argument, a
host name, is added to \fIcommand\fR when it is invoked.  If a proxy
is required, the callback should return a two-element list containing
the proxy server and proxy port.  Otherwise the filter command should return
an empty list.
.RS
.PP
The default value of \fB\-proxyfilter\fR is \fBhttp::ProxyRequired\fR, and
this command returns the values of the \fB\-proxyhost\fR and
\fB\-proxyport\fR settings if they are non-empty.  The options
\fB\-proxyhost\fR, \fB\-proxyport\fR, and \fB\-proxynot\fR are used only
by \fBhttp::ProxyRequired\fR, and nowhere else in \fB::http::geturl\fR.
A user-supplied \fB\-proxyfilter\fR command may use these options, or
alternatively it may obtain values from elsewhere in the calling script.
In the latter case, any values provided for \fB\-proxyhost\fR,
\fB\-proxyport\fR, and \fB\-proxynot\fR are unused.
.PP
The \fB::http::geturl\fR command runs the \fB\-proxyfilter\fR callback inside
a \fBcatch\fR command.  Therefore an error in the callback command does
not call the \fBbgerror\fR handler.  See the \fBERRORS\fR section for
details.
.RE
.\" OPTION: -proxyhost
.TP
\fB\-proxyhost\fI hostname\fR
.
The host name or IP address of the proxy server, if any.  If this value is
the empty string, the URL host is contacted directly.  See
\fB\-proxyfilter\fR for how the value is used.
.\" OPTION: -proxynot
.TP
\fB\-proxynot\fI list\fR
.
A Tcl list of domain names and IP addresses that should be accessed directly,
not through the proxy server.  The target hostname is compared with each list
element using a case-insensitive \fBstring match\fR.  It is often convenient
to use the wildcard "*" at the start of a domain name (e.g. *.example.com) or
at the end of an IP address (e.g. 192.168.0.*).  See \fB\-proxyfilter\fR for
how the value is used.
.\" OPTION: -proxyport
.TP
\fB\-proxyport\fI number\fR
.
The port number of the proxy server.  See \fB\-proxyfilter\fR for how the
value is used.
.\" OPTION: -repost
.TP
\fB\-repost\fI boolean\fR
.
Specifies what to do if a POST request over a persistent connection fails
because the server has half-closed the connection.  If boolean \fBtrue\fR, the
request
will be automatically retried; if boolean \fBfalse\fR it will not, and the
application
that uses \fBhttp::geturl\fR is expected to seek user confirmation before
retrying the POST.  The value \fBtrue\fR should be used only under certain
conditions. See the \fBPERSISTENT SOCKETS\fR section for details. The
default is 0.
.\" OPTION: -threadlevel
.TP
\fB\-threadlevel\fI level\fR
.
Specifies whether and how to use the \fBThread\fR package.  Possible values
of \fIlevel\fR are 0, 1 or 2.
.RS
.IP \fB0\fR
(the default) do not use Thread
.IP \fB1\fR
use Thread if it is available, do not use it if it is unavailable
.IP \fB2\fR
use Thread if it is available, raise an error if it is unavailable
.PP
The Tcl \fBsocket -async\fR command can block in adverse cases (e.g. a slow
DNS lookup).  Using the Thread package works around this problem, for both
HTTP and HTTPS transactions.  Values of \fIlevel\fR other than 0 are
available only to the main interpreter in each thread.  See
section \fBTHREADS\fR for more information.
.RE
.\" OPTION: -urlencoding
.TP
\fB\-urlencoding\fI encoding\fR
.
The \fIencoding\fR used for creating the x-url-encoded URLs with
\fB::http::formatQuery\fR and \fB::http::quoteString\fR.
The default is \fButf-8\fR, as specified by RFC 2718.
.\" OPTION: -useragent
.TP
\fB\-useragent\fI string\fR
.
The value of the User-Agent header in the HTTP request.  In an unsafe
interpreter, the default value depends upon the operating system, and
the version numbers of \fBhttp\fR and \fBTcl\fR, and is (for example)
.QW "\fBMozilla/5.0 (Windows; U; Windows NT 10.0) http/2.10.2 Tcl/9.0.0\fR" .
A safe interpreter cannot determine its operating system, and so the default
in a safe interpreter is to use a Windows 10 value with the current version
numbers of \fBhttp\fR and \fBTcl\fR.
.\" OPTION: -zip
.TP
\fB\-zip\fI boolean\fR
.
If the value is boolean \fBtrue\fR, then by default requests will send a header
.QW "\fBAccept-Encoding: gzip,deflate\fR" .
If the value is boolean \fBfalse\fR, then by default requests will send a header
.QW "\fBAccept-Encoding: identity\fR" .
In either case the default can be overridden for an individual request by
supplying a custom \fBAccept-Encoding\fR header in the \fB\-headers\fR option
of \fBhttp::geturl\fR. The default value is 1.
.RE
.\" COMMAND: geturl
.TP
\fB::http::geturl\fI url\fR ?\fIoptions\fR?
.
The \fB::http::geturl\fR command is the main procedure in the package.
The \fB\-query\fR or \fB\-querychannel\fR option causes a POST operation and
the \fB\-validate\fR option causes a HEAD operation;
otherwise, a GET operation is performed.  The \fB::http::geturl\fR command
returns a \fItoken\fR value that can be passed as an argument to other commands
to get information about the transaction.  See the \fBMETADATA\fR and
\fBERRORS\fR section for
details.  The \fB::http::geturl\fR command blocks until the operation
completes, unless the \fB\-command\fR option specifies a callback
that is invoked when the HTTP transaction completes.
\fB::http::geturl\fR takes several options:
.RS
.\" OPTION: -binary
.TP
\fB\-binary\fI boolean\fR
.
Specifies whether to force interpreting the URL data as binary.  Normally
this is auto-detected (anything not beginning with a \fBtext\fR content
type or whose content encoding is \fBgzip\fR or \fBdeflate\fR is
considered binary data).
.\" OPTION: -blocksize
.TP
\fB\-blocksize\fI size\fR
.
The block size used when reading the URL.
At most \fIsize\fR bytes are read at once.  After each block, a call to the
\fB\-progress\fR callback is made (if that option is specified).
.\" OPTION: -channel
.TP
\fB\-channel\fI name\fR
.
Copy the URL contents to channel \fIname\fR instead of saving it in
a Tcl variable for retrieval by \fB::http::responseBody\fR.
.\" OPTION: -command
.TP
\fB\-command\fI callback\fR
.
The presence of this option causes \fB::http::geturl\fR to return immediately.
After the HTTP transaction completes, the value of \fIcallback\fR is expanded,
an additional argument is added, and the resulting command is evaluated.
The additional argument is the \fItoken\fR returned
from \fB::http::geturl\fR. This token is the name of an array that is
described in the \fBSTATE ARRAY\fR section.  Here is a template for the
callback:
.RS
.PP
.CS
proc httpCallback {token} {
    upvar 0 $token state
    # Access state as a Tcl array defined in this proc
    ...
    return
}
.CE
.PP
The \fB::http::geturl\fR command runs the \fB\-command\fR callback inside
a \fBcatch\fR command.  Therefore an error in the callback command does
not call the \fBbgerror\fR handler.  See the \fBERRORS\fR section for
details.
.RE
.\" OPTION: -guesstype
.TP
\fB\-guesstype\fI boolean\fR
.
Attempt to guess the \fBContent-Type\fR and character set when a misconfigured
server provides no information.  The default value is \fIfalse\fR (do
nothing).  If boolean \fItrue\fR then, if the server does not send a
\fBContent-Type\fR header, or if it sends the value "application/octet-stream",
\fBhttp::geturl\fR will attempt to guess appropriate values.  This is not
intended to become a general-purpose tool, and currently it is limited to
detecting XML documents that begin with an XML declaration.  In this case
the \fBContent-Type\fR is changed to "application/xml", the binary flag
state(binary) is changed to 0, and the character set is changed to
the one specified by the "encoding" tag of the XML line, or to utf-8 if no
encoding is specified.  Not used if a \fB\-channel\fR is specified.
.\" OPTION: -handler
.TP
\fB\-handler\fI callback\fR
.
If this option is absent, \fBhttp::geturl\fR processes incoming data itself,
either appending it to the state(body) variable or writing it to the -channel.
But if the \fB\-handler\fR option is present, \fBhttp::geturl\fR does not do
this processing and instead calls \fIcallback\fR.
Whenever HTTP data is available, the value of \fIcallback\fR is expanded, an
additional two arguments are added, and the resulting command is evaluated.
The two additional
arguments are: the socket for the HTTP data and the \fItoken\fR returned from
\fB::http::geturl\fR.  The token is the name of a global array that is
described in the \fBSTATE ARRAY\fR section.  The procedure is expected
to return the number of bytes read from the socket.  Here is a
template for the callback:
.RS
.PP
.CS
proc httpHandlerCallback {socket token} {
    upvar 0 $token state
    # Access socket, and state as a Tcl array defined in this proc
    # For example...
    ...
    set data [read $socket 1000]
    set nbytes [string length $data]
    ...
    return $nbytes
}
.CE
.PP
The \fBhttp::geturl\fR code for the \fB\-handler\fR option is not compatible
with either compression or chunked transfer-encoding.  If \fB\-handler\fR is
specified, then to work around these issues \fBhttp::geturl\fR will reduce the
HTTP protocol to 1.0, and override the \fB\-zip\fR option (i.e. it will
send the header \fBAccept-Encoding: identity\fR instead
of \fBAccept-Encoding: gzip,deflate\fR).
.PP
If options \fB\-handler\fR and \fB\-channel\fR are used together, the handler
is responsible for copying the data from the HTTP socket to the specified
channel.  The name of the channel is available to the handler as element
\fB\-channel\fR of the token array.
.PP
The \fB::http::geturl\fR command runs the \fB\-handler\fR callback inside
a \fBcatch\fR command.  Therefore an error in the callback command does
not call the \fBbgerror\fR handler.  See the \fBERRORS\fR section for
details.
.RE
.\" OPTION: -headers
.TP
\fB\-headers\fI keyvaluelist\fR
.
This option is used to add headers not already specified
by \fB::http::config\fR to the HTTP request.  The
\fIkeyvaluelist\fR argument must be a list with an even number of
elements that alternate between keys and values.  The keys become
header field names.  Newlines are stripped from the values so the
header cannot be corrupted.  For example, if \fIkeyvaluelist\fR is
\fBPragma no-cache\fR then the following header is included in the
HTTP request:
.RS
.PP
.CS
Pragma: no-cache
.CE
.RE
.\" OPTION: -keepalive
.TP
\fB\-keepalive\fI boolean\fR
.
If boolean \fBtrue\fR, attempt to keep the connection open for servicing
multiple requests.  Default is 0.
.\" OPTION: -method
.TP
\fB\-method\fI type\fR
.
Force the HTTP request method to \fItype\fR. \fB::http::geturl\fR will
auto-select GET, POST or HEAD based on other options, but this option overrides
that selection and enables choices like PUT and DELETE for WebDAV support.
.RS
.PP
It is the caller's responsibility to ensure that the headers and request body
(if any) conform to the requirements of the request method.  For example, if
using \fB\-method\fI POST\fR to send a POST with an empty request body, the
caller must also supply the option
.PP
.CS
\-headers {Content-Length 0}
.CE
.RE
.\" OPTION: -myaddr
.TP
\fB\-myaddr\fI address\fR
.
Pass an specific local address to the underlying \fBsocket\fR call in case
multiple interfaces are available.
.\" OPTION: -progress
.TP
\fB\-progress\fI callback\fR
.
If the \fB\-progress\fR option is present,
then the \fIcallback\fR is made after each transfer of data from the URL.
The value of \fIcallback\fR is expanded, an additional three arguments are
added, and the resulting command is evaluated.
The three additional arguments are: the \fItoken\fR returned from
\fB::http::geturl\fR, the expected total size of the contents from the
\fBContent-Length\fR response header, and the current number of bytes
transferred so far.  The token is the name of a global array that is
described in the \fBSTATE ARRAY\fR section.  The expected total size may
be unknown, in which
case zero is passed to the callback.  Here is a template for the
progress callback:
.RS
.PP
.CS
proc httpProgress {token total current} {
    upvar 0 $token state
    # Access state as a Tcl array defined in this proc
    ...
    return
}
.CE
.RE
.\" OPTION: -protocol
.TP
\fB\-protocol\fI version\fR
.
Select the HTTP protocol version to use. This should be 1.0 or 1.1 (the
default). Should only be necessary for servers that do not understand or
otherwise complain about HTTP/1.1.
.\" OPTION: -query
.TP
\fB\-query\fI query\fR
.
This flag (if the value is non-empty) causes \fB::http::geturl\fR to do a
POST request that passes the string
\fIquery\fR verbatim to the server as the request payload.
The content format (and encoding) of \fIquery\fR is announced by the request
header \fBContent-Type\fR which is set by the option \fB\-type\fR.  Any value
of \fB\-type\fR is permitted, and it is the responsibility of the caller to
supply \fIquery\fR in the correct format.
.RS
.PP
If \fB\-type\fR is not specified, it defaults to
\fIapplication/x-www-form-urlencoded\fR, which requires \fIquery\fR to be an
x-url-encoding formatted query-string (this \fB\-type\fR and query format are
used in a POST submitted from an html form).  The \fB::http::formatQuery\fR
procedure can be used to do the formatting.
.RE
.\" OPTION: -queryblocksize
.TP
\fB\-queryblocksize\fI size\fR
.
The block size used when posting query data to the URL.
At most
\fIsize\fR
bytes are written at once.  After each block, a call to the
\fB\-queryprogress\fR
callback is made (if that option is specified).
.\" OPTION: -querychannel
.TP
\fB\-querychannel\fI channelID\fR
.
This flag causes \fB::http::geturl\fR to do a POST request that passes the
data contained in \fIchannelID\fR to the server. The data contained in
\fIchannelID\fR must be an x-url-encoding
formatted query unless the \fB\-type\fR option below is used.
If a \fBContent-Length\fR header is not specified via the \fB\-headers\fR
options, \fB::http::geturl\fR attempts to determine the size of the post data
in order to create that header.  If it is
unable to determine the size, it returns an error.
.\" OPTION: -queryprogress
.TP
\fB\-queryprogress\fI callback\fR
.
If the \fB\-queryprogress\fR option is present,
then the \fIcallback\fR is made after each transfer of data to the URL
in a POST request (i.e. a call to \fB::http::geturl\fR with
option \fB\-query\fR or \fB\-querychannel\fR) and acts exactly like
the \fB\-progress\fR option (the callback format is the same).
.\" OPTION: -strict
.TP
\fB\-strict\fI boolean\fR
.
If true then the command will test that the URL complies with RFC 3986, i.e.
that it has no characters that should be "x-url-encoded" (e.g. a space should
be encoded to "%20").  Default value is 1.
.\" OPTION: -timeout
.TP
\fB\-timeout\fI milliseconds\fR
.
If \fImilliseconds\fR is non-zero, then \fB::http::geturl\fR sets up a timeout
to occur after the specified number of milliseconds.
A timeout results in a call to \fB::http::reset\fR and to
the \fB\-command\fR callback, if specified.
The return value of \fB::http::status\fR (and the value of the \fIstatus\fR key
in the dictionary returned by \fB::http::responseInfo\fR) is \fBtimeout\fR
after a timeout has occurred.
.\" OPTION: -type
.TP
\fB\-type\fI mime-type\fR
.
Use \fImime-type\fR as the \fBContent-Type\fR value, instead of the
default value (\fBapplication/x-www-form-urlencoded\fR) during a
POST operation.
.\" OPTION: -validate
.TP
\fB\-validate\fI boolean\fR
.
If \fIboolean\fR is non-zero, then \fB::http::geturl\fR does an HTTP HEAD
request.  This server returns the same status line and response headers as it
would for a HTTP GET request, but omits the response entity
(the URL "contents").  The response headers are available after the
transaction using command \fB::http::responseHeaders\fR or, for selected
information, \fB::http::responseInfo\fR.
.RE
.\" COMMAND: formatQuery
.TP
\fB::http::formatQuery\fI key value\fR ?\fIkey value ...\fR?
.
This procedure does x-url-encoding of query data.  It takes an even
number of arguments that are the keys and values of the query.  It
encodes the keys and values, and generates one string that has the
proper & and = separators.  The result is suitable for the
\fB\-query\fR value passed to \fB::http::geturl\fR.
.\" COMMAND: quoteString
.TP
\fB::http::quoteString\fI value\fR
.
This procedure does x-url-encoding of string.  It takes a single argument and
encodes it.
.\" COMMAND: reset
.TP
\fB::http::reset\fI token\fR ?\fIwhy\fR?
.
This command resets the HTTP transaction identified by \fItoken\fR, if any.
This sets the \fBstate(status)\fR value to \fIwhy\fR, which defaults to
\fBreset\fR, and then calls the registered \fB\-command\fR callback.
.\" COMMAND: wait
.TP
\fB::http::wait\fI token\fR
.
This command blocks and waits for the
transaction to complete.  This only works in trusted code because it
uses \fBvwait\fR.  Also, it is not useful for the case where
\fB::http::geturl\fR is called \fIwithout\fR the \fB\-command\fR option
because in this case the \fB::http::geturl\fR call does not return
until the HTTP transaction is complete, and thus there is nothing to
wait for.
.\" COMMAND: status
.TP
\fB::http::status\fI token\fR
.
This command returns a description of the status of the HTTP transaction.
The return value is the empty string until the HTTP transaction is
completed; after completion it has one of the values ok, eof, error,
timeout, and reset.  The meaning of these values is described in the
section \fBERRORS\fR (below).
.PP
.RS
The name "status" is not related to the terms "status line" and
"status code" that are defined for a HTTP response.
.RE
.\" COMMAND: size
.TP
\fB::http::size\fI token\fR
.
This command returns the number of bytes
received so far from the URL in the \fB::http::geturl\fR call.
.\" COMMAND: error
.TP
\fB::http::error\fI token\fR
.
This command returns the error information if the HTTP transaction failed,
or the empty string if there was no error.  The information is a Tcl list of
the error message, stack trace, and error code.
.\" COMMAND: postError
.TP
\fB::http::postError\fI token\fR
.
A POST request is a call to \fB::http::geturl\fR with either
the \fB\-query\fR or \fB\-querychannel\fR option.
The \fB::http::postError\fR command returns the error information generated
when a HTTP POST request sends its request-body to the server; or the empty
string if there was no error.  The information is a Tcl list of the error
message, stack trace, and error code.  When this type of error occurs,
the \fB::http::geturl\fR command continues the transaction and attempts to
receive a response from the server.
.\" COMMAND: cleanup
.TP
\fB::http::cleanup\fI token\fR
.
This procedure cleans up the state associated with the connection
identified by \fItoken\fR.  After this call, the procedures
like \fB::http::responseBody\fR cannot be used to get information
about the operation.  It is \fIstrongly\fR recommended that you call
this function after you are done with a given HTTP request.  Not doing
so will result in memory not being freed, and if your app calls
\fB::http::geturl\fR enough times, the memory leak could cause a
performance hit...or worse.
.\" COMMAND: requestLine
.TP
\fB::http::requestLine\fI token\fR
.
This command returns the "request line" sent to the server.
The "request line" is the first line of a HTTP client request, and has three
elements separated by spaces: the HTTP method, the URL relative to the server,
and the HTTP version. Examples:
.PP
.RS
GET / HTTP/1.1
GET /introduction.html?subject=plumbing HTTP/1.1
POST /forms/order.html HTTP/1.1
.RE
.\" COMMAND: requestHeaders
.TP
\fB::http::requestHeaders\fI token\fR ?\fIheaderName\fR?
.
This command returns the HTTP request header names and values, in the
order that they were sent to the server, as a Tcl list of the form
?name value ...?  Header names are case-insensitive and are converted to lower
case.  The return value is not a \fBdict\fR because some header names may occur
more than once.  If one argument is supplied, all request headers
are returned.  If two arguments are supplied, the
second provides the value of a header name.  Only headers with the requested
name (converted to lower case) are returned.  If no such headers are found,
an empty list is returned.
.\" COMMAND: requestHeaderValue
.TP
\fB::http::requestHeaderValue\fI token headerName\fR
.
This command returns the value of the HTTP request header named
\fIheaderName\fR.  Header names are case-insensitive and are converted to
lower case.  If no such header exists, the return value is the empty string.
If there are multiple headers named \fIheaderName\fR, the result is obtained
by joining the individual values with the string ", " (comma and space),
preserving their order.
.\" COMMAND: responseLine
.TP
\fB::http::responseLine\fI token\fR
.
This command returns the first line of the server response: the
HTTP "status line".  The "status line" has three
elements separated by spaces: the HTTP version, a three-digit numerical
"status code", and a "reason phrase".  Only the reason phrase may contain
spaces.  Examples:
.PP
.RS
HTTP/1.1 200 OK
HTTP/1.0 404 Not Found
.RE
.RS
The "status code" is a three-digit number in the range 100 to 599.
A value of 200 is the normal return from a GET request, and its matching
"reason phrase" is "OK".  Codes beginning with 4 or 5 indicate errors.
Codes beginning with 3 are redirection errors.  In this case the
\fBLocation\fR response header specifies a new URL that contains the
requested information.
.PP
The "reason phrase" is a textual description of the "status code": it may
vary from server to server,
and can be changed without affecting the HTTP protocol.  The recommended
values (RFC 7231 and IANA assignments) for each code are provided by the
command \fB::http::reasonPhrase\fR.
.RE
.\" COMMAND: responseCode
.TP
\fB::http::responseCode\fI token\fR
.
This command returns the "status code" (200, 404, etc.) of the server
"status line".  If a three-digit code cannot be found, the full status
line is returned.  See command \fB::http::responseLine\fR for more information
on the "status line".
.\" COMMAND: reasonPhrase
.TP
\fB::http::reasonPhrase\fI code\fR
.
This command returns the IANA recommended "reason phrase" for a particular
"status code" returned by a HTTP server.  The argument \fIcode\fR is a valid
status code, and therefore is an integer in the range 100 to 599 inclusive.
For numbers in this range with no assigned meaning, the command returns the
value "Unassigned".  Several status codes are used only in response to the
methods defined by HTTP extensions such as WebDAV, and not in response to a
HEAD, GET, or POST request method.
.PP
.RS
The "reason phrase" returned by a HTTP server may differ from the recommended
value, without affecting the HTTP protocol.  The value returned by
\fB::http::geturl\fR can be obtained by calling either command
\fB::http::responseLine\fR (which returns the full status line) or command
\fB::http::responseInfo\fR (which returns a dictionary, with
the "reason phrase" stored in key \fIreasonPhrase\fR).
.PP
A registry of valid status codes is maintained at
https://www.iana.org/assignments/http-status-codes/http-status-codes.xhtml
.RE
.\" COMMAND: responseHeaders
.TP
\fB::http::responseHeaders\fI token\fR ?\fIheaderName\fR?
.
The response from a HTTP server includes metadata headers that describe the
response body and the transaction itself.
This command returns the HTTP response header names and values, in the
order that they were received from the server, as a Tcl list of the form
?name value ...?  Header names are case-insensitive and are converted to lower
case.  The return value is not a \fBdict\fR because some header names may occur
more than once, notably \fBSet-Cookie\fR.  If the second argument is not
supplied, all response headers are returned.  If the second argument is
supplied, it provides the value of a header name.  Only headers with the
requested name (converted to lower case) are returned.  If no such headers
are found, an empty list is returned.  See section \fBMETADATA\fR for more
information.
.\" COMMAND: responseHeaderValue
.TP
\fB::http::responseHeaderValue\fI token headerName\fR
.
This command returns the value of the HTTP response header named
\fIheaderName\fR.  Header names are case-insensitive and are converted to
lower case.  If no such header exists, the return value is the empty string.
If there are multiple headers named \fIheaderName\fR, the result is obtained
by joining the individual values with the string ", " (comma and space),
preserving their order.  Multiple headers with the same name may be processed
in this manner, except \fBSet-Cookie\fR which does not conform to the
comma-separated-list syntax and cannot be combined into a single value.
Each \fBSet-Cookie\fR header must be treated individually, e.g. by processing
the return value of \fB::http::responseHeaders\fI token\fR \fBSet-Cookie\fR.
.\" COMMAND: responseInfo
.TP
\fB::http::responseInfo\fI token\fR
.
This command returns a \fBdict\fR of selected response metadata that are
essential for identifying a successful transaction and making use of the
response, along with other metadata that are informational.  The keys of
the \fBdict\fR are \fIstage\fR, \fIstatus\fR, \fIresponseCode\fR,
\fIreasonPhrase\fR, \fIcontentType\fR, \fIbinary\fR, \fIredirection\fR,
\fIupgrade\fR, \fIerror\fR, \fIpostError\fR, \fImethod\fR, \fIcharset\fR,
\fIcompression\fR, \fIhttpRequest\fR, \fIhttpResponse\fR, \fIurl\fR,
\fIconnectionRequest\fR, \fIconnectionResponse\fR, \fIconnectionActual\fR,
\fItransferEncoding\fR, \fItotalPost\fR, \fIcurrentPost\fR, \fItotalSize\fR,
and \fIcurrentSize\fR.  The meaning of these keys is described in the
section \fBMETADATA\fR below.
.RS
.PP
It is always worth checking the value of \fIbinary\fR after a HTTP transaction,
to determine whether a misconfigured server has caused http to interpret a
text resource as a binary, or vice versa.
.PP
After a POST transaction, check the value of \fIpostError\fR to verify that
the request body was uploaded without error.
.RE
.\" COMMAND: responseBody
.TP
\fB::http::responseBody\fI token\fR
.
This command returns the entity sent by the HTTP server (unless
\fI-channel\fR was used, in which case the entity was delivered to the
channel, and the command returns the empty string).
.RS
.PP
Other terms for
"entity", with varying precision, include "representation of resource",
"resource", "response body after decoding", "payload",
"message body after decoding", "content(s)", and "file".
.RE
.\" COMMAND: register
.TP
\fB::http::register\fI proto port command\fR ?\fIsocketCmdVarName\fR? ?\fIuseSockThread\fR? ?\fIendToEndProxy\fR?
.
This procedure allows one to provide custom HTTP transport types
such as HTTPS, by registering a prefix, the default port, and the
command to execute to create the Tcl \fBchannel\fR. The optional
arguments configure how \fBhttp\fR uses the custom transport, and have
default values that are compatible with older versions of \fBhttp\fR
in which \fB::http::register\fR has no optional arguments.
.RS
.PP
Argument \fIsocketCmdVarName\fR is the name of a variable provided by
the transport, whose value is the command used by the transport to open
a socket.  Its default value is set by the transport and is "::socket",
but if the name of the variable is supplied to \fB::http::register\fR,
then \fBhttp\fR will set a new value in order to make optional
facilities available.  These facilities are enabled by the optional
arguments \fIuseSockThread\fR, \fIendToEndProxy\fR, which take boolean
values with default value \fIfalse\fR.
.PP
Iff argument \fIuseSockThread\fR is supplied and is boolean \fItrue\fR,
then iff permitted by the value [\fBhttp::config\fI \-threadlevel\fR]
and by the availability of package \fBThread\fR, sockets created for
the transport will be opened in a different thread so that a slow DNS
lookup will not cause the script to block.
.PP
Iff argument \fIendToEndProxy\fR is supplied and is boolean \fItrue\fR,
then when \fBhttp::geturl\fR accesses a server via a proxy, it will
open a channel by sending a CONNECT request to the proxy, and it will
then make its request over this channel.  This allows end-to-end
encryption for HTTPS requests made through a proxy.
.PP
For example,
.RS
.PP
.CS
\fBpackage require http\fR
\fBpackage require tls\fR

::http::register https 443 ::tls::socket ::tls::socketCmd 1 1

set token [::http::geturl https://my.secure.site/]
.CE
.RE
.RE
.\" COMMAND: registerError
.TP
\fB::http::registerError\fI sock\fR ?\fImessage\fR?
.
This procedure allows a registered protocol handler to deliver an error
message for use by \fBhttp\fR.  Calling this command does not raise an
error. The command is useful when a registered protocol detects an problem
(for example, an invalid TLS certificate) that will cause an error to
propagate to \fBhttp\fR.  The command allows \fBhttp\fR to provide a
precise error message rather than a general one.  The command returns the
value provided by the last call with argument \fImessage\fR, or the empty
string if no such call has been made.
.\" COMMAND: unregister
.TP
\fB::http::unregister\fI proto\fR
.
This procedure unregisters a protocol handler that was previously
registered via \fB::http::register\fR, returning a six-item list of
the values that were previously supplied to \fB::http::register\fR
if there was such a handler, and an error if there was no such handler.
.\" COMMAND: code
.TP
\fB::http::code\fI token\fR
.
An alternative name for the command \fB::http::responseLine\fR
.\" COMMAND: data
.TP
\fB::http::data\fI token\fR
.
An alternative name for the command \fB::http::responseBody\fR.
.\" COMMAND: meta
.TP
\fB::http::meta\fI token\fR ?\fIheaderName\fR?
.
An alternative name for the command \fB::http::responseHeaders\fR
.\" COMMAND: ncode
.TP
\fB::http::ncode\fI token\fR
.
An alternative name for the command \fB::http::responseCode\fR
.SH ERRORS
The \fB::http::geturl\fR procedure will raise errors in the following cases:
invalid command line options,
or an invalid URL.
These errors mean that it
cannot even start the network transaction.
For synchronous \fB::http::geturl\fR calls (where \fB\-command\fR is
not specified), it will raise an error if
the URL is on a non-existent host
or at a bad port on an existing host.
It will also raise an error for any I/O errors while
writing out the HTTP request line and headers, or
reading the HTTP reply headers or data.  Because \fB::http::geturl\fR
does not return a token in these cases, it does all the required
cleanup and there is no issue of your app having to call
\fB::http::cleanup\fR.
.PP
For asynchronous \fB::http::geturl\fR calls, all of the above error
situations apply, except that if there is any error while reading the
HTTP reply headers or data, no exception is thrown.  This is because
after writing the HTTP headers, \fB::http::geturl\fR returns, and the
rest of the HTTP transaction occurs in the background.  The command
callback can check if any error occurred during the read by calling
\fB::http::responseInfo\fR to check the transaction status.
.PP
Alternatively, if the main program flow reaches a point where it needs
to know the result of the asynchronous HTTP request, it can call
\fB::http::wait\fR and then check status and error, just as the
synchronous call does.
.PP
The \fB::http::geturl\fR command runs the \fB\-command\fR, \fB\-handler\fR,
and \fB\-proxyfilter\fR callbacks inside a \fBcatch\fR command.  Therefore
an error in the callback command does not call the \fBbgerror\fR handler.
When debugging one of these
callbacks, it may be convenient to report errors by using a
\fBcatch\fR command within the callback command itself, e.g. to write
an error message to stdout.
.PP
In any case, you must still call
\fB::http::cleanup\fR to delete the state array when you are done.
.PP
There are other possible results of the HTTP transaction
determined by examining the status from \fB::http::status\fR (or the value
of the \fIstatus\fR key in the dictionary returned
by \fB::http::responseInfo\fR).
These are described below.
.IP \fBok\fR
If the HTTP transaction completes entirely, then status will be \fBok\fR.
However, you should still check the \fB::http::responseLine\fR value to get
the HTTP status.  The \fB::http::responseCode\fR procedure provides just
the numeric error (e.g., 200, 404 or 500) while the \fB::http::responseLine\fR
procedure returns a value like
.QW "HTTP 404 File not found" .
.IP \fBeof\fR
If the server closes the socket without replying, then no error
is raised, but the status of the transaction will be \fBeof\fR.
.IP \fBerror\fR
The error message, stack trace, and error code are accessible
via \fB::http::error\fR.  The error message is also provided by the value of
the \fIerror\fR key in the dictionary returned by \fB::http::responseInfo\fR.
.IP \fBtimeout\fR
A timeout occurred before the transaction could complete.
.IP \fBreset\fR
The user has called \fB::http::reset\fR.
.IP \fB""\fR
(empty string) The transaction has not yet finished.
.PP
Another error possibility is that \fB::http::geturl\fR failed to
write the whole of the POST request body (\fB\-query\fR or \fB\-querychannel\fR
data) to the server.  \fB::http::geturl\fR stores the error message for later
retrieval by the \fB::http::postError\fR or \fB::http::responseInfo\fR
commands, and then attempts to complete the transaction.
If it can read the server's response the status will be \fBok\fR, but it is
important to call \fB::http::postError\fR or \fB::http::responseInfo\fR after
every POST to check that the data was sent in full.
If the server has closed the connection the status will be \fBeof\fR.
.SH "METADATA"
.PP
.SS "MOST USEFUL METADATA"
When a HTTP server responds to a request, it supplies not only the entity
requested, but also metadata.  This is provided by the first line (the
"status line") of the response, and by a number of HTTP headers.  Further
metadata relates to how \fB::http::geturl\fR has processed the response
from the server.
.PP
The most important metadata can be accessed with the command
\fB::http::responseInfo\fR.
This command returns a \fBdict\fR of metadata that are essential for
identifying a successful transaction and making use of the response,
along with other metadata that are informational.  The keys of
the \fBdict\fR are:
.PP
.RS
.RS
.\" TODO: Find a better way to mark this up!
\fB===== Essential Values =====\fR
.RE
.RE
.IP \fBstage\fR
This value, set by \fB::http::geturl\fR, describes the stage that the
transaction has reached. Values, in order of the transaction lifecycle,
are: "created", "connecting", "header", "body", and "complete".  The
other \fBdict\fR keys will not be available until the value of \fBstage\fR
is "body" or "complete".  The key \fBcurrentSize\fR has its final value only
when \fBstage\fR is "complete".
.IP \fBstatus\fR
This value, set by \fB::http::geturl\fR, is "ok" for a successful transaction;
"eof", "error", "timeout", or "reset" for an unsuccessful transaction; or ""
if the transaction is still in progress.  The value is the same as that
returned by command \fB::http::status\fR. The meaning of these values is
described in the section \fBERRORS\fR (above).
.IP \fBresponseCode\fR
The "HTTP status code" sent by the server in the first line (the "status line")
of the response.  If the value cannot be extracted from the status line, the
full status line is returned.
.IP \fBreasonPhrase\fR
The "reason phrase" sent by the server as a description of the HTTP status code.
If the value cannot be extracted from the status line, the full status
line is returned.
.IP \fBcontentType\fR
The value of the \fBContent-Type\fR response header or, if the header was not
supplied, the default value "application/octet-stream".
.IP \fBbinary\fR
This boolean value, set by \fB::http::geturl\fR, describes how the command
has interpreted the entity returned by the server (after decoding any
compression specified by the \fBContent-Encoding\fR response header).
This decoded entity is accessible as the return value of the
command \fB::http::responseBody\fR.
.PP
.RS
The value is \fBtrue\fR if http has interpreted the decoded entity as binary.
The value returned by \fB::http::responseBody\fR is a Tcl binary string.
This is a suitable format for image data, zip files, etc.
\fB::http::geturl\fR chooses this value if the user has requested a binary
interpretation by passing the option \fB\-binary\fR to the command, or if the
server has supplied a binary content type in a \fBContent-Type\fR response
header, or if the server has not supplied any \fBContent-Type\fR header.
.PP
The value is \fBfalse\fR in other cases, and this means that http has
interpreted the decoded entity as text. The text has been converted, from the
character set notified by the server, into Tcl's internal Unicode format;
the value returned by \fB::http::responseBody\fR is an ordinary Tcl string.
.PP
It is always worth checking the value of "binary" after a HTTP transaction,
to determine whether a misconfigured server has caused http to interpret a
text resource as a binary, or vice versa.
.RE
.IP \fBredirection\fR
The URL that is the redirection target. The value is that of the \fBLocation\fR
response header.  This header is sent when a response has status code
3XX (redirection).
.IP \fBupgrade\fR
If not empty, the value indicates the protocol(s) to which the server will
switch after completion of this transaction, while continuing to use the
same connection.  When the server intends to switch protocols, it will also
send the value "101" as the status code (the \fBresponseCode\fR key), and the
word "upgrade" as an element of the \fBConnection\fR response header (the
\fBconnectionResponse\fR key), and it will not send a response body.
See the section \fBPROTOCOL UPGRADES\fR for more information.
.IP \fBerror\fR
The error message, if there is one.  Further information, including a stack
trace and error code, are available from command \fB::http::error\fR.
.IP \fBpostError\fR
The error message (if any) generated when a HTTP POST request sends its
request-body to the server.  Further information, including a stack trace
and error code, are available from command \fB::http::postError\fR.  A POST
transaction may appear complete, according to the
keys \fBstage\fR, \fBstatus\fR, and \fBresponseCode\fR, but it is important
to check this \fBpostError\fR key in case an error occurred when uploading
the request-body.
.PP
.RS
.RS
\fB===== Informational Values =====\fR
.RE
.RE
.IP \fBmethod\fR
The HTTP method used in the request.
.IP \fBcharset\fR
The value of the charset attribute of the \fBContent-Type\fR response header.
The charset value is used only for a text resource.  If the server did not
specify a charset, the value defaults to that of the
variable \fB::http::defaultCharset\fR, which unless it has been deliberately
modified by the caller is \fBiso8859-1\fR.  Incoming text data is automatically
converted from the character set defined by \fBcharset\fR to Tcl's internal
Unicode representation, i.e. to a Tcl string.
.IP \fBcompression\fR
A copy of the \fBContent-Encoding\fR response-header value.
.IP \fBhttpRequest\fR
The version of HTTP specified in the request (i.e. sent in the request line).
The value is that of the option \fB\-protocol\fR supplied
to \fB::http::geturl\fR (default value "1.1"), unless the command reduced the
value to "1.0" because it was passed the \fB\-handler\fR option.
.IP \fBhttpResponse\fR
The version of HTTP used by the server (obtained from the response
"status line").  The server uses this version of HTTP in its response, but
ensures that this response is compatible with the HTTP version specified in the
client's request.  If the value cannot be extracted from the status line, the
full status line is returned.
.IP \fBurl\fR
The requested URL, typically the URL supplied as an argument
to \fB::http::geturl\fR but without its "fragment" (the final part of the URL
beginning with "#").
.IP \fBconnectionRequest\fR
The value, if any, sent to the server in \fBConnection\fR request header(s).
.IP \fBconnectionResponse\fR
The value, if any, received from the server in \fBConnection\fR response
header(s).
.IP \fBconnectionActual\fR
This value, set by \fB::http::geturl\fR, reports whether the connection was
closed after the transaction (value "close"), or left open (value "keep-alive").
.IP \fBtransferEncoding\fR
The value of the Transfer-Encoding response header, if it is present.
The value is either "chunked" (indicating HTTP/1.1 "chunked encoding") or
the empty string.
.IP \fBtotalPost\fR
The total length of the request body in a POST request.
.IP \fBcurrentPost\fR
The number of bytes of the POST request body sent to the server so far.
The value is the same as that returned by command \fB::http::size\fR.
.IP \fBtotalSize\fR
A copy of the \fBContent-Length\fR response-header value.
The number of bytes specified in a \fBContent-Length\fR header, if one
was sent.  If none was sent, the value is 0.  A correctly configured server
omits this header if the transfer-encoding is "chunked", or (for older
servers) if the server closes the connection when it reaches the end of
the resource.
.IP \fBcurrentSize\fR
The number of bytes fetched from the server so far.
.PP
.SS "MORE METADATA"
The dictionary returned by \fB::http::responseInfo\fR is the most useful
subset of the available metadata.  Other metadata include:
.PP
1. The full "status line" of the response, available as the return value
of command \fB::http::responseLine\fR.
.PP
2. The full response headers, available as the return value of
command \fB::http::responseHeaders\fR.  This return value is a list of the
response-header names and values, in the order that they were received from
the server.
.PP
The return value is not a \fBdict\fR because some header names may
occur more than once, notably \fBSet-Cookie\fR. If the value is read
into a \fBdict\fR or into an array (using array set), only the last header
with each name will be preserved.
.PP
.RS
Some of the header names (metadata keys) are listed below, but the HTTP
standard defines several more, and servers are free to add their own.
When a dictionary key is mentioned below, this refers to the \fBdict\fR
value returned by command \fB::http::responseInfo\fR.
.IP \fBContent-Type\fR
The content type of the URL contents.  Examples include \fBtext/html\fR,
\fBimage/gif,\fR \fBapplication/postscript\fR and
\fBapplication/x-tcl\fR.  Text values typically specify a character set, e.g.
\fBtext/html; charset=UTF-8\fR.  Dictionary key \fIcontentType\fR.
.IP \fBContent-Length\fR
The advertised size in bytes of the contents, available as dictionary
key \fItotalSize\fR.  The actual number of bytes read by \fB::http::geturl\fR
so far is available as dictionary key \fBcurrentSize\fR.
.IP \fBContent-Encoding\fR
The compression algorithm used for the contents.
Examples include \fBgzip\fR, \fBdeflate\fR.
Dictionary key \fIcontent\fR.
.IP \fBLocation\fR
This header is sent when a response has status code 3XX (redirection).
It provides the URL that is the redirection target.
Dictionary key \fIredirection\fR.
.IP \fBSet-Cookie\fR
This header is sent to offer a cookie to the client.  Cookie management is
done by the \fB::http::config\fR option \fB\-cookiejar\fR, and so
the \fBSet-Cookie\fR headers need not be parsed by user scripts.
See section \fBCOOKIE JAR PROTOCOL\fR.
.IP \fBConnection\fR
The value can be supplied as a comma-separated list, or by multiple headers.
The list often has only one element, either "close" or "keep-alive".
The value "upgrade" indicates a successful upgrade request and is typically
combined with the status code 101, an \fBUpgrade\fR response header, and no
response body.  Dictionary key \fIconnectionResponse\fR.
.IP \fBUpgrade\fR
The value indicates the protocol(s) to which the server will switch
immediately after the empty line that terminates the 101 response headers.
Dictionary key \fIupgrade\fR.
.RE
.PP
.SS "EVEN MORE METADATA"
.IP 1.
Details of the HTTP request.  The request is determined by the options
supplied to \fB::http::geturl\fR and \fB::http::config\fR.  However, it is
sometimes helpful to examine what \fB::http::geturl\fR actually sent to the
server, and this information is available through
commands \fB::http::requestHeaders\fR and \fB::http::requestLine\fR.
.IP 2.
The state array: the internal variables of \fB::http::geturl\fR.
It may sometimes be helpful to examine this array.
Details are given in the next section.
.SH "STATE ARRAY"
The \fB::http::geturl\fR procedure returns a \fItoken\fR that can be used
as an argument to other \fB::http::*\fR commands, which examine and manage
the state of the HTTP transaction.  For most purposes these commands are
sufficient.  The \fItoken\fR can also be used to access
the internal state of the transaction, which is stored in a Tcl array.
This facility is most useful when writing callback commands for the
options \fB\-command\fR, \fB\-handler\fR, \fB\-progress\fR,
or \fB\-queryprogress\fR.
Use the following command inside the proc to define an easy-to-use
array \fIstate\fR as a local variable within the proc
.PP
.CS
upvar 0 $token state
.CE
.PP
Once the data associated with the URL is no longer needed, the state
array should be unset to free up storage.
The \fB::http::cleanup\fR procedure is provided for that purpose.
.PP
The following elements of the array are supported, and are the origin of the
values returned by commands as described below.  When a dictionary key is
mentioned below, this refers to the \fBdict\fR value returned by
command \fB::http::responseInfo\fR.
.RS
.IP \fBbinary\fR
For dictionary key \fIbinary\fR.
.IP \fBbody\fR
For command \fB::http::responseBody\fR.
.IP \fBcharset\fR
For dictionary key \fIcharset\fR.
.IP \fBcoding\fR
For dictionary key \fIcompression\fR.
.IP \fBconnection\fR
For dictionary key \fIconnectionActual\fR.
.IP \fBcurrentsize\fR
For command \fB::http::size\fR; and for dictionary key \fIcurrentSize\fR.
.IP \fBerror\fR
For command \fB::http::error\fR; part is used in dictionary key \fIerror\fR.
.IP \fBhttp\fR
For command \fB::http::responseLine\fR.
.IP \fBhttpResponse\fR
For dictionary key \fIhttpResponse\fR.
.IP \fBmeta\fR
For command \fB::http::responseHeaders\fR. Further discussion above in the
section \fBMORE METADATA\fR.
.IP \fBmethod\fR
For dictionary key \fImethod\fR.
.IP \fBposterror\fR
For dictionary key \fIpostError\fR.
.IP \fBpostErrorFull\fR
For command \fB::http::postError\fR.
.IP \fB\-protocol\fR
For dictionary key \fIhttpRequest\fR.
.IP \fBquerylength\fR
For dictionary key \fItotalPost\fR.
.IP \fBqueryoffset\fR
For dictionary key \fIcurrentPost\fR.
.IP \fBreasonPhrase\fR
For dictionary key \fIreasonPhrase\fR.
.IP \fBrequestHeaders\fR
For command \fB::http::requestHeaders\fR.
.IP \fBrequestLine\fR
For command \fB::http::requestLine\fR.
.IP \fBresponseCode\fR
For dictionary key \fIresponseCode\fR.
.IP \fBstate\fR
For dictionary key \fIstage\fR.
.IP \fBstatus\fR
For command \fB::http::status\fR; and for dictionary key \fIstatus\fR.
.IP \fBtotalsize\fR
For dictionary key \fItotalSize\fR.
.IP \fBtransfer\fR
For dictionary key \fItransferEncoding\fR.
.IP \fBtype\fR
For dictionary key \fIcontentType\fR.
.IP \fBupgrade\fR
For dictionary key \fIupgrade\fR.
.IP \fBurl\fR
For dictionary key \fIurl\fR.
.RE
.SH "PERSISTENT CONNECTIONS"
.PP
.SS "BASICS"
.PP
See RFC 7230 Sec 6, which supersedes RFC 2616 Sec 8.1.
.PP
A persistent connection allows multiple HTTP/1.1 transactions to be
carried over the same TCP connection.  Pipelining allows a
client to make multiple requests over a persistent connection without
waiting for each response.  The server sends responses in the same order
that the requests were received.
.PP
If a POST request fails to complete, typically user confirmation is
needed before sending the request again.  The user may wish to verify
whether the server was modified by the failed POST request, before
sending the same request again.
.PP
A HTTP request will use a persistent socket if the call to
\fBhttp::geturl\fR has the option \fB\-keepalive true\fR. It will use
pipelining where permitted if the \fBhttp::config\fR option
\fB\-pipeline\fR is boolean \fBtrue\fR (its default value).
.PP
The http package maintains no more than one persistent connection to each
server (i.e. each value of
.QW "domain:port" ).
If \fBhttp::geturl\fR is called to make a request over a persistent
connection while the connection is busy with another request, the new
request will be held in a queue until the connection is free.
.PP
The http package does not support HTTP/1.0 persistent connections
controlled by the \fBKeep-Alive\fR header.
.SS "SPECIAL CASES"
.PP
This subsection discusses issues related to closure of the
persistent connection by the server, automatic retry of failed requests,
the special treatment necessary for POST requests, and the options for
dealing with these cases.
.PP
In accordance with RFC 7230, \fBhttp::geturl\fR does not pipeline
requests that use the POST method.  If a POST uses a persistent
connection and is not the first request on that connection,
\fBhttp::geturl\fR waits until it has received the response for the previous
request; or (if \fBhttp::config\fR option \fB\-postfresh\fR is boolean
\fBtrue\fR) it uses a new connection for each POST.
.PP
If the server is processing a number of pipelined requests, and sends a
response header
.QW "\fBConnection: close\fR"
with one of the responses (other than the last), then subsequent responses
are unfulfilled. \fBhttp::geturl\fR will send the unfulfilled requests again
over a new connection.
.PP
A difficulty arises when a HTTP client sends a request over a persistent
connection that has been idle for a while.  The HTTP server may
half-close an apparently idle connection while the client is sending a
request, but before the request arrives at the server: in this case (an
.QW "asynchronous close event" )
the request will fail.  The difficulty arises because the client cannot
be certain whether the POST modified the state of the server.  For HEAD or
GET requests, \fBhttp::geturl\fR opens another connection and retransmits
the failed request. However, if the request was a POST, RFC 7230 forbids
automatic retry by default, suggesting either user confirmation, or
confirmation by user-agent software that has semantic understanding of
the application.  The \fBhttp::config\fR option \fB\-repost\fR allows for
either possibility.
.PP
Asynchronous close events can occur only in a short interval of time.  The
\fBhttp\fR package monitors each persistent connection for closure by the
server.  Upon detection, the connection is also closed at the client end,
and subsequent requests will use a fresh connection.
.PP
If the \fBhttp::geturl\fR command is called with option \fB\-keepalive true\fR,
then it will both try to use an existing persistent connection
(if one is available), and it will send the server a
.QW "\fBConnection: keep-alive\fR"
request header asking to keep the connection open for future requests.
.PP
The \fBhttp::config\fR options \fB\-pipeline\fR, \fB\-postfresh\fR, and
\fB\-repost\fR relate to persistent connections.
.PP
Option \fB\-pipeline\fR, if boolean \fBtrue\fR, will pipeline GET and HEAD
requests made over a persistent connection.  POST requests will not be
pipelined - if the
POST is not the first transaction on the connection, its request will not
be sent until the previous response has finished.  GET and HEAD requests
made after a POST will not be sent until the POST response has been
delivered, and will not be sent if the POST fails.
.PP
Option \fB\-postfresh\fR, if boolean \fBtrue\fR, will override the
\fBhttp::geturl\fR option \fB\-keepalive\fR, and always open a fresh connection
for a POST request.
.PP
Option \fB\-repost\fR, if \fBtrue\fR, permits automatic retry of a POST request
that fails because it uses a persistent connection that the server has
half-closed (an
.QW "asynchronous close event" ).
Subsequent GET and HEAD requests in a failed pipeline will also be retried.
\fIThe \fB\-repost\fI option should be used only if the application understands
that the retry is appropriate\fR - specifically, the application must know
that if the failed POST successfully modified the state of the server, a repeat
POST would have no adverse effect.
.VS TIP406
.SH "COOKIE JAR PROTOCOL"
.PP
Cookies are short key-value pairs used to implement sessions within the
otherwise-stateless HTTP protocol. (See RFC 6265 for details; Tcl does not
implement the Cookie2 protocol as that is rarely seen in the wild.)
.PP
Cookie storage management commands \(em
.QW "cookie jars"
\(em must support these subcommands which form the HTTP cookie storage
management protocol. Note that \fIcookieJar\fR below does not have to be a
command name; it is properly a command prefix (a Tcl list of words that will
be expanded in place) and admits many possible implementations.
.PP
Though not formally part of the protocol, it is expected that particular
values of \fIcookieJar\fR will correspond to sessions; it is up to the caller
of \fB::http::config\fR to decide what session applies and to manage the
deletion of said sessions when they are no longer desired (which should be
when they not configured as the current cookie jar).
.\" METHOD: getCookies
.TP
\fIcookieJar \fBgetCookies \fIprotocol host requestPath\fR
.
This command asks the cookie jar what cookies should be supplied for a
particular request. It should take the \fIprotocol\fR (typically \fBhttp\fR or
\fBhttps\fR), \fIhost\fR name and \fIrequestPath\fR (parsed from the \fIurl\fR
argument to \fB::http::geturl\fR) and return a list of cookie keys and values
that describe the cookies to supply to the remote host. The list must have an
even number of elements.
.RS
.PP
There should only ever be at most one cookie with a particular key for any
request (typically the one with the most specific \fIhost\fR/domain match and
most specific \fIrequestPath\fR/path match), but there may be many cookies
with different names in any request.
.RE
.\" METHOD: storeCookie
.TP
\fIcookieJar \fBstoreCookie \fIcookieDictionary\fR
.
This command asks the cookie jar to store a particular cookie that was
returned by a request; the result of this command is ignored. The cookie
(which will have been parsed by the http package) is described by a
dictionary, \fIcookieDictionary\fR, that may have the following keys:
.RS
.IP \fBdomain\fR
This is always present. Its value describes the domain hostname \fIor
prefix\fR that the cookie should be returned for.  The checking of the domain
against the origin (below) should be careful since sites that issue cookies
should only do so for domains related to themselves. Cookies that do not obey
a relevant origin matching rule should be ignored.
.IP \fBexpires\fR
This is optional. If present, the cookie is intended to be a persistent cookie
and the value of the option is the Tcl timestamp (in seconds from the same
base as \fBclock seconds\fR) of when the cookie expires (which may be in the
past, which should result in the cookie being deleted immediately). If absent,
the cookie is intended to be a session cookie that should be not persisted
beyond the lifetime of the cookie jar.
.IP \fBhostonly\fR
This is always present. Its value is a boolean that describes whether the
cookie is a single host cookie (true) or a domain-level cookie (false).
.IP \fBhttponly\fR
This is always present. Its value is a boolean that is true when the site
wishes the cookie to only ever be used with HTTP (or HTTPS) traffic.
.IP \fBkey\fR
This is always present. Its value is the \fIkey\fR of the cookie, which is
part of the information that must be return when sending this cookie back in a
future request.
.IP \fBorigin\fR
This is always present. Its value describes where the http package believes it
received the cookie from, which may be useful for checking whether the
cookie's domain is valid.
.IP \fBpath\fR
This is always present. Its value describes the path prefix of requests to the
cookie domain where the cookie should be returned.
.IP \fBsecure\fR
This is always present. Its value is a boolean that is true when the cookie
should only used on requests sent over secure channels (typically HTTPS).
.IP \fBvalue\fR
This is always present. Its value is the value of the cookie, which is part of
the information that must be return when sending this cookie back in a future
request.
.PP
Other keys may always be ignored; they have no meaning in this protocol.
.RE
.VE TIP406
.SH "PROTOCOL UPGRADES"
.PP
The HTTP/1.1 \fBConnection\fR and \fBUpgrade\fR request headers inform the
server that the client wishes to change the protocol used over the existing
connection (RFC 7230).
This mechanism can be used to request a WebSocket (RFC 6455), a
higher version of the HTTP protocol (HTTP 2), or TLS encryption.  If the
server accepts the upgrade request, its response code will be 101.
.PP
To request a protocol upgrade when calling \fBhttp::geturl\fR,
the \fB\-headers\fR option must supply appropriate values for \fBConnection\fR
and \fBUpgrade\fR, and
the \fB\-command\fR option must supply a command that implements the requested
protocol and can also handle the server response if the server refuses the
protocol upgrade.  For upgrade requests \fBhttp::geturl\fR ignores the value of
option \fB\-keepalive\fR, and always uses the value \fB0\fR so that the upgrade
request is not made over a connection that is intended for multiple HTTP
requests.
.PP
The Tcllib library \fBwebsocket\fR implements WebSockets, and makes the
necessary calls to commands in the \fBhttp\fR package.
.PP
There is currently no native Tcl client library for HTTP/2.
.PP
The \fBUpgrade\fR mechanism is not used to request TLS in web browsers, because
\fBhttp\fR and \fBhttps\fR are served over different ports.  It is used by
protocols such as Internet Printing Protocol (IPP) that are built on top of
\fBhttp(s)\fR and use the same TCP port number for both secure and insecure
traffic.
.PP
In browsers, opportunistic encryption is instead implemented by the
\fBUpgrade-Insecure-Requests\fR client header.  If a secure service is
available, the server response code is a 307 redirect, and the response header
\fBLocation\fR specifies the target URL.  The browser must
call \fBhttp::geturl\fR again in order to fetch this URL.
See https://w3c.github.io/webappsec-upgrade-insecure-requests/
.PP
.SH THREADS
.PP
.SS "PURPOSE"
.PP
Command \fB::http::geturl\fR uses the Tcl \fB::socket\fR command with
the \fB\-async\fR option to connect to a remote server, but the return from
this command can be delayed in adverse cases (e.g. a slow DNS lookup),
preventing the event loop from processing other events.
This delay is avoided if the \fB::socket\fR command is evaluated in another
thread.  The Thread package is not part of Tcl but is provided in
"Batteries Included" distributions.  Instead of the \fB::socket\fR command,
the http package uses \fB::http::socket\fR which makes connections in the
manner specified by the value of \fB\-threadlevel\fR and the availability
of package Thread.
.PP
.SS "WITH TLS (HTTPS)"
.PP
The same \fB\-threadlevel\fR configuration applies to both HTTP and HTTPS
connections.
HTTPS is enabled by using the \fBhttp::register\fR command, typically by
specifying the \fB::tls::socket\fR command of the tls package to handle TLS
cryptography.  The \fB::tls::socket\fR command connects to the remote server by
using the command specified by the value of variable \fB::tls::socketCmd\fR, and
this value defaults to "::socket".  If http::geturl finds
that \fB::tls::socketCmd\fR has this value, it replaces it with the value
"::http::socket".  If \fB::tls::socketCmd\fR has a value other than "::socket",
i.e. if the script or the Tcl installation has replaced the value "::socket"
with the name of a different command, then http does not change the value.
The script or installation that modified \fB::tls::socketCmd\fR is responsible
for integrating \fB::http::socket\fR into its own replacement command.
.PP
.SS "WITH A CHILD INTERPRETER"
.PP
The peer thread can transfer the socket only to the main interpreter of the
script's thread.  Therefore the thread-based \fB::http::socket\fR works with
non-zero \fB\-threadlevel\fR values only if the script runs in the main
interpreter.  A child interpreter must use \fB\-threadlevel 0\fR unless the
parent interpreter has provided alternative facilities.  The main parent
interpreter may grant full \fB\-threadlevel\fR facilities to a child
interpreter, for example by aliasing, to \fB::http::socket\fR in the child,
a command that runs \fBhttp::socket\fR in the parent, and then transfers
the socket to the child.
.PP
.SH EXAMPLE
.PP
This example creates a procedure to copy a URL to a file while printing a
progress meter, and prints the response headers associated with the URL.
.PP
.CS
proc httpcopy { url file {chunk 4096} } {
    set out [open $file w]
    set token [\fB::http::geturl\fR $url -channel $out \e
            -progress httpCopyProgress -blocksize $chunk]
    close $out

    # This ends the line started by httpCopyProgress
    puts stderr ""

    upvar 0 $token state
    set max 0
    foreach {name value} $state(meta) {
        if {[string length $name] > $max} {
            set max [string length $name]
        }
        if {[regexp -nocase ^location$ $name]} {
            # Handle URL redirects
            puts stderr "Location:$value"
            return [httpcopy [string trim $value] $file $chunk]
        }
    }
    incr max
    foreach {name value} $state(meta) {
        puts [format "%-*s %s" $max $name: $value]
    }

    return $token
}
proc httpCopyProgress {args} {
    puts -nonewline stderr .
    flush stderr
}
.CE
.SH "SEE ALSO"
safe(n), socket(n), safesock(n)
.SH KEYWORDS
internet, security policy, socket, www
'\" Local Variables:
'\" mode: nroff
'\" End: