texinode(TCP Function System)(Zftp Function System)(Calendar Function System)(Top)
chapter(TCP Function System)
cindex(TCP function system)
cindex(ztcp, function system based on)
sect(Description)

A module tt(zsh/net/tcp) is provided to provide network I/O over
TCP/IP from within the shell; see its description in
ifzman(\
zmanref(zshmodules)\
)\
ifnzman(\
noderef(Zsh Modules)\
).  This manual page describes a function suite based on the module.  
If the module is installed, the functions are usually installed at the
same time, in which case they will be available for
autoloading in the default function search path.  In addition to the
tt(zsh/net/tcp) module, the tt(zsh/zselect) module is used to implement
timeouts on read operations.  For troubleshooting tips, consult the
corresponding advice for the tt(zftp) functions described in
ifzman(\
zmanref(zshzftpsys)\
)\
ifnzman(\
noderef(Zftp Function System)\
).

There are functions corresponding to the basic I/O operations open, close,
read and send, named tt(tcp_open) etc., as well as a function
tt(tcp_expect) for pattern match analysis of data read as input.  The
system makes it easy to receive data from and send data to multiple named
sessions at once.  In addition, it can be linked with the shell's line
editor in such a way that input data is automatically shown at the
terminal.  Other facilities available including logging, filtering and
configurable output prompts.

To use the system where it is available, it should be enough to
`tt(autoload -U tcp_open)' and run tt(tcp_open) as documented below to
start a session.  The tt(tcp_open) function will autoload the remaining
functions.

startmenu()
menu(TCP Functions)
menu(TCP Parameters)
menu(TCP Examples)
menu(TCP Bugs)
endmenu()

texinode(TCP Functions)(TCP Parameters)()(TCP Function System)
sect(TCP User Functions)

subsect(Basic I/O)

startitem()
findex(tcp_open)
xitem(tt(tcp_open) [ tt(-qz) ] var(host port) [ var(sess) ])
xitem(tt(tcp_open) [ tt(-qz) ] [ tt(-s) var(sess) | tt(-l) var(sess)[tt(,)...] ] ... )
item(tt(tcp_open) [ tt(-qz) ] [ tt(-a) var(fd) | tt(-f) var(fd) ] [ var(sess) ])(
Open a new session.  In the first and simplest form, open a TCP connection
to host var(host) at port var(port); numeric and symbolic forms are
understood for both.

If var(sess) is given, this becomes the name of the session which can be
used to refer to multiple different TCP connections.  If var(sess) is
not given, the function will invent a numeric name value (note this is
em(not) the same as the file descriptor to which the session is attached).
It is recommended that session names not include `funny' characters, where
funny characters are not well-defined but certainly do not include
alphanumerics or underscores, and certainly do include whitespace.

In the second case, one or more sessions to be opened are given by name.
A single session name is given after tt(-s) and a comma-separated list
after tt(-l); both options may be repeated as many times as necessary.
A failure to open any session causes tt(tcp_open) to abort.
The host and port are read from the file tt(.ztcp_sessions) in the same
directory as the user's zsh initialisation files, i.e. usually the home
directory, but tt($ZDOTDIR) if that is set.  The file consists of lines
each giving a session name and the corresponding host and port, in that
order (note the session name comes first, not last), separated by
whitespace.

The third form allows passive and fake TCP connections.  If the option
tt(-a) is used, its argument is a file descriptor open for listening for
connections.  No function front-end is provided to open such a file
descriptor, but a call to `tt(ztcp -l) var(port)' will create one with the
file descriptor stored in the parameter tt($REPLY).  The listening port can
be closed with `tt(ztcp -c) var(fd)'.  A call to `tt(tcp_open -a) var(fd)'
will block until a remote TCP connection is made to var(port) on the local
machine.  At this point, a session is created in the usual way and is
largely indistinguishable from an active connection created with one of the
first two forms.

If the option tt(-f) is used, its argument is a file descriptor which is
used directly as if it were a TCP session.  How well the remainder of the
TCP function system copes with this depends on what actually underlies this
file descriptor.  A regular file is likely to be unusable; a FIFO (pipe) of
some sort will work better, but note that it is not a good idea for two
different sessions to attempt to read from the same FIFO at once.

If the option tt(-q) is given with any of the three forms, tt(tcp_open)
will not print informational messages, although it will in any case exit
with an appropriate status.

If the line editor (zle) is in use, which is typically the case if the
shell is interactive, tt(tcp_open) installs a handler inside zle which
will check for new data at the same time as it checks for keyboard input.
This is convenient as the shell consumes no CPU time while waiting; the
test is performed by the operating system.  Giving the option tt(-z) to
any of the forms of tt(tcp_open) prevents the handler from being
installed, so data must be read explicitly.  Note, however, this is not
necessary for executing complete sets of send and read commands from a
function, as zle is not active at this point.  Generally speaking, the
handler is only active when the shell is waiting for input at a command
prompt or in the tt(vared) builtin.  The option has no effect if zle is not
active; `tt([[ -o zle]])' will test for this.

The first session to be opened becomes the current session and subsequent
calls to tt(tcp_open) do not change it.  The current session is stored
in the parameter tt($TCP_SESS); see below for more detail about the
parameters used by the system.

The function tt(tcp_on_open), if defined, is called when a session
is opened.  See the description below.
)
findex(tcp_close)
item(tt(tcp_close) [ tt(-qn) ] [ tt(-a) | tt(-l) var(sess)[tt(,)...] | var(sess) ... ])(
Close the named sessions, or the current session if none is given,
or all open sessions if tt(-a) is given.  The options tt(-l) and tt(-s) are
both handled for consistency with tt(tcp_open), although the latter is
redundant.

If the session being closed is the current one, tt($TCP_SESS) is unset,
leaving no current session, even if there are other sessions still open.

If the session was opened with tt(tcp_open -f), the file descriptor is
closed so long as it is in the range 0 to 9 accessible directly from the
command line.  If the option tt(-n) is given, no attempt will be made to
close file descriptors in this case.  The tt(-n) option is not used for
genuine tt(ztcp) session; the file descriptors are always closed with the
session.

If the option tt(-q) is given, no informational messages will be printed.
)
findex(tcp_read)
redef(SPACES)(0)(tt(ifztexi(NOTRANS(@ @ @ @ @ @ @ @ @ ))ifnztexi(         )))
xitem(tt(tcp_read )[ tt(-bdq) ] [ tt(-t) var(TO) ] [ tt(-T) var(TO) ])
item(SPACES()[ tt(-a) | tt(-u) var(fd)[tt(,)...] | tt(-l) var(sess)[tt(,)...] | tt(-s) var(sess) ... ])(
Perform a read operation on the current session, or on a list of
sessions if any are given with tt(-u), tt(-l) or tt(-s), or all open
sessions if the option tt(-a) is given.  Any of the tt(-u), tt(-l) or
tt(-s) options may be repeated or mixed together.  The tt(-u) option
specifies a file descriptor directly (only those managed by this system
are useful), the other two specify sessions as described for
tt(tcp_open) above.

The function checks for new data available on all the sessions listed.
Unless the tt(-b) option is given, it will not block waiting for new data.
Any one line of data from any of the available sessions will be read,
stored in the parameter tt($TCP_LINE), and displayed to standard output
unless tt($TCP_SILENT) contains a non-empty string.  When printed to
standard output the string tt($TCP_PROMPT) will be shown at the start of
the line; the default form for this includes the name of the session being
read.  See below for more information on these parameters.  In this mode,
tt(tcp_read) can be called repeatedly until it returns status 2 which
indicates all pending input from all specified sessions has been handled.

With the option tt(-b), equivalent to an infinite timeout, the function
will block until a line is available to read from one of the specified
sessions.  However, only a single line is returned.

The option tt(-d) indicates that all pending input should be drained.  In
this case tt(tcp_read) may process multiple lines in the manner given
above; only the last is stored in tt($TCP_LINE), but the complete set is
stored in the array tt($tcp_lines).  This is cleared at the start of each
call to tt(tcp_read).

The options tt(-t) and tt(-T) specify a timeout in seconds, which may be a
floating point number for increased accuracy.  With tt(-t) the timeout is
applied before each line read.  With tt(-T), the timeout applies to the
overall operation, possibly including multiple read operations if the
option tt(-d) is present; without this option, there is no distinction
between tt(-t) and tt(-T).

The function does not print informational messages, but if the option
tt(-q) is given, no error message is printed for a non-existent session.

A return status of 2 indicates a timeout or no data to read.  Any other
non-zero return status indicates some error condition.

See tt(tcp_log) for how to control where data is sent by tt(tcp_read).
)
findex(tcp_send)
xitem(tt(tcp_send) [ tt(-cnq) ] [ tt(-s) var(sess) | tt(-l) var(sess)[tt(,)...] ] var(data) ...)
item(tt(tcp_send) [ tt(-cnq) ] tt(-a) var(data) ...)(
Send the supplied data strings to all the specified sessions in turn.  The
underlying operation differs little from a `tt(print -r)' to the session's
file descriptor, although it attempts to prevent the shell from dying owing
to a tt(SIGPIPE) caused by an attempt to write to a defunct session.

The option tt(-c) causes tt(tcp_send) to behave like tt(cat).  It reads
lines from standard input until end of input and sends them in turn to the
specified session+LPAR()s+RPAR() exactly as if they were given as var(data)
arguments to individual tt(tcp_send) commands.

The option tt(-n) prevents tt(tcp_send) from putting a newline at the end
of the data strings.

The remaining options all behave as for tt(tcp_read).

The data arguments are not further processed once they have been passed to
tt(tcp_send); they are simply passed down to tt(print -r).

If the parameter tt($TCP_OUTPUT) is a non-empty string and logging is
enabled then the data sent to each session will be echoed to the log
file+LPAR()s+RPAR() with tt($TCP_OUTPUT) in front where appropriate, much
in the manner of tt($TCP_PROMPT).
)
enditem()

subsect(Session Management)

startitem()
findex(tcp_alias)
xitem(tt(tcp_alias) [ tt(-q) ] var(alias)tt(=)var(sess) ...)
xitem(tt(tcp_alias) [ tt(-q) ] [ var(alias) ... ])
item(tt(tcp_alias) tt(-d) [ tt(-q) ] var(alias) ...)(
This function is not particularly well tested.

The first form creates an alias for a session name; var(alias) can then be
used to refer to the existing session var(sess).  As many aliases may be
listed as required.

The second form lists any aliases specified, or all aliases if none.

The third form deletes all the aliases listed.  The underlying sessions are
not affected.

The option tt(-q) suppresses an inconsistently chosen subset of error
messages.
)
findex(tcp_log)
item(tt(tcp_log) [ tt(-asc) ] [ tt(-n) | tt(-N) ] [ var(logfile) ])(
With an argument var(logfile), all future input from tt(tcp_read) will be
logged to the named file.  Unless tt(-a) (append) is given, this file will
first be truncated or created empty.  With no arguments, show the current
status of logging.

With the option tt(-s), per-session logging is enabled.  Input from
tt(tcp_read) is output to the file var(logfile)tt(.)var(sess).  As the
session is automatically discriminated by the filename, the contents are
raw (no tt($TCP_PROMPT)).  The option  tt(-a) applies as above.
Per-session logging and logging of all data in one file are not mutually
exclusive.

The option tt(-c) closes all logging, both complete and per-session logs.

The options tt(-n) and tt(-N) respectively turn off or restore output of
data read by tt(tcp_read) to standard output; hence `tt(tcp_log -cn)' turns
off all output by tt(tcp_read).

The function is purely a convenient front end to setting the parameters
tt($TCP_LOG), tt($TCP_LOG_SESS), tt($TCP_SILENT), which are described below.
)
findex(tcp_rename)
item(tt(tcp_rename) var(old) var(new))(
Rename session var(old) to session var(new).  The old name becomes invalid.
)
findex(tcp_sess)
item(tt(tcp_sess) [ var(sess) [ var(command) [ var(arg) ... ] ] ])(
With no arguments, list all the open sessions and associated file
descriptors.  The current session is marked with a star.  For use in
functions, direct access to the parameters tt($tcp_by_name), tt($tcp_by_fd)
and tt($TCP_SESS) is probably more convenient; see below.

With a var(sess) argument, set the current session to var(sess).
This is equivalent to changing tt($TCP_SESS) directly.

With additional arguments, temporarily set the current session while
executing `var(command) var(arg) ...'.  var(command) is re-evaluated
so as to expand aliases etc., but the remaining var(arg)s are passed
through as that appear to tt(tcp_sess).  The original session is restored
when tt(tcp_sess) exits.
)
enditem()

subsect(Advanced I/O)

startitem()
findex(tcp_command)
item(tt(tcp_command) var(send-option) ... var(send-argument) ...)(
This is a convenient front-end to tt(tcp_send).  All arguments are passed
to tt(tcp_send), then the function pauses waiting for data.  While data is
arriving at least every tt($TCP_TIMEOUT) (default 0.3) seconds, data is
handled and printed out according to the current settings.  Status 0 is
always returned.

This is generally only useful for interactive use, to prevent the display
becoming fragmented by output returned from the connection.  Within a
programme or function it is generally better to handle reading data by a
more explicit method.
)
findex(tcp_expect)
redef(SPACES)(0)(tt(ifztexi(NOTRANS(@ @ @ @ @ @ @ @ @ @ @ ))ifnztexi(           )))
xitem(tt(tcp_expect )[ tt(-q) ] [ tt(-p) var(var) | tt(-P) var(var) ] [ tt(-t) var(TO) | tt(-T) var(TO) ])
item(SPACES()[ tt(-a) | tt(-s) var(sess) | tt(-l) var(sess)[tt(,)...] ] var(pattern) ...)(
Wait for input matching any of the given var(pattern)s from any of the
specified sessions.  Input is ignored until an input line matches one of
the given patterns; at this point status zero is returned, the matching
line is stored in tt($TCP_LINE), and the full set of lines read during the
call to tt(tcp_expect) is stored in the array tt($tcp_expect_lines).

Sessions are specified in the same way as tt(tcp_read): the default is to
use the current session, otherwise the sessions specified by tt(-a),
tt(-s), or tt(-l) are used.

Each var(pattern) is a standard zsh extended-globbing pattern; note that it
needs to be quoted to avoid it being expanded immediately by filename
generation.  It must match the full line, so to match a substring there
must be a `tt(*)' at the start and end.  The line matched against includes
the tt($TCP_PROMPT) added by tt(tcp_read).  It is possible to include the
globbing flags `tt(#b)' or `tt(#m)' in the patterns to make backreferences
available in the parameters tt($MATCH), tt($match), etc., as described in
the base zsh documentation on pattern matching.

Unlike tt(tcp_read), the default behaviour of tt(tcp_expect) is to block
indefinitely until the required input is found.  This can be modified by
specifying a timeout with tt(-t) or tt(-T); these function as in
tt(tcp_read), specifying a per-read or overall timeout, respectively, in
seconds, as an integer or floating-point number.  As tt(tcp_read), the
function returns status 2 if a timeout occurs.

The function returns as soon as any one of the patterns given match.  If
the caller needs to know which of the patterns matched, the option tt(-p)
var(var) can be used; on return, tt($var) is set to the number of the
pattern using ordinary zsh indexing, i.e. the first is 1, and so on.  Note
the absence of a `tt($)' in front of var(var).  To avoid clashes, the
parameter cannot begin with `tt(_expect)'.  The index -1 is used if
there is a timeout and 0 if there is no match.

The option tt(-P) var(var) works similarly to tt(-p), but instead of
numerical indexes the regular arguments must begin with a prefix
followed by a colon: that prefix is then used as a tag to which var(var)
is set when the argument matches.  The tag tt(timeout) is used if there
is a timeout and the empty string if there is no match.  Note it is
acceptable for different arguments to start with the same prefix if the
matches do not need to be distinguished.

The option tt(-q) is passed directly down to tt(tcp_read).

As all input is done via tt(tcp_read), all the usual rules about output of
lines read apply.  One exception is that the parameter tt($tcp_lines) will
only reflect the line actually matched by tt(tcp_expect); use
tt($tcp_expect_lines) for the full set of lines read during the function
call.
)
findex(tcp_proxy)
item(tt(tcp_proxy))(
This is a simple-minded function to accept a TCP connection and execute a
command with I/O redirected to the connection.  Extreme caution should be
taken as there is no security whatsoever and this can leave your computer
open to the world.  Ideally, it should only be used behind a firewall.

The first argument is a TCP port on which the function will listen.

The remaining arguments give a command and its arguments to execute with
standard input, standard output and standard error redirected to the
file descriptor on which the TCP session has been accepted.
If no command is given, a new zsh is started.  This gives everyone on
your network direct access to your account, which in many cases will be a
bad thing.

The command is run in the background, so tt(tcp_proxy) can then accept new
connections.  It continues to accept new connections until interrupted.
)
findex(tcp_spam)
item(tt(tcp_spam) [ tt(-ertv) ] [ tt(-a) | tt(-s) var(sess) | tt(-l) var(sess)[tt(,)...] ] var(cmd) [ var(arg) ... ])(
Execute `var(cmd) [ var(arg) ... ]' for each session in turn.  Note this executes
the command and arguments; it does not send the command line as data
unless the tt(-t) (transmit) option is given.

The sessions may be selected explicitly with the standard tt(-a), tt(-s) or
tt(-l) options, or may be chosen implicitly.  If none of the three options
is given the rules are: first, if the array tt($tcp_spam_list) is set, this
is taken as the list of sessions, otherwise all sessions are taken.
Second, any sessions given in the array tt($tcp_no_spam_list) are removed
from the list of sessions.

Normally, any sessions added by the `tt(-a)' flag or when all sessions are
chosen implicitly are spammed in alphabetic order; sessions given by the
tt($tcp_spam_list) array or on the command line are spammed in the order
given.  The tt(-r) flag reverses the order however it was arrived it.

The tt(-v) flag specifies that a tt($TCP_PROMPT) will be output before each
session.  This is output after any modification to tt(TCP_SESS) by the
user-defined tt(tcp_on_spam) function described below.  (Obviously that
function is able to generate its own output.)

If the option tt(-e) is present, the line given as `var(cmd) [ var(arg) ... ]' is executed
using tt(eval), otherwise it is executed without any further processing.
)
findex(tcp_talk)
item(tt(tcp_talk))(
This is a fairly simple-minded attempt to force input to the line editor to
go straight to the default tt(TCP_SESS).

An escape string, tt($TCP_TALK_ESCAPE), default `tt(:)', is used to allow
access to normal shell operation.  If it is on its own at the start of the
line, or followed only by whitespace, the line editor returns to normal
operation.  Otherwise, the string and any following whitespace are skipped
and the remainder of the line executed as shell input without any change of
the line editor's operating mode.

The current implementation is somewhat deficient in terms of use of the
command history.  For this reason, many users will prefer to use some form
of alternative approach for sending data easily to the current session.
One simple approach is to alias some special character (such as `tt(%)') to
`tt(tcp_command -)tt(-)'.
)
findex(tcp_wait)
item(tt(tcp_wait))(
The sole argument is an integer or floating point number which gives the
seconds to delay.  The shell will do nothing for that period except wait
for input on all TCP sessions by calling tt(tcp_read -a).  This is similar
to the interactive behaviour at the command prompt when zle handlers are
installed.
)
enditem()

subsect(`One-shot' file transfer)
startitem()
xitem(tt(tcp_point) var(port))
item(tt(tcp_shoot) var(host) var(port))(
This pair of functions provide a simple way to transfer a file between
two hosts within the shell.  Note, however, that bulk data transfer is
currently done using tt(cat).  tt(tcp_point) reads any data arriving at
var(port) and sends it to standard output; tt(tcp_shoot) connects to
var(port) on var(host) and sends its standard input.  Any unused var(port)
may be used; the standard mechanism for picking a port is to think of a
random four-digit number above 1024 until one works.

To transfer a file from host tt(woodcock) to host tt(springes), on
tt(springes):

example(tcp_point 8091 >output_file)

and on tt(woodcock):

example(tcp_shoot springes 8091 <input_file)

As these two functions do not require tt(tcp_open) to set up a TCP
connection first, they may need to be autoloaded separately.
)
enditem()

sect(TCP User-defined Functions)

Certain functions, if defined by the user, will be called by the function
system in certain contexts.  This facility depends on the module
tt(zsh/parameter), which is usually available in interactive shells as the
completion system depends on it.  None of the functions need be defined;
they simply provide convenient hooks when necessary.

Typically, these are called after the requested action has been taken, so
that the various parameters will reflect the new state.

startitem()
findex(tcp_on_alias)
item(tt(tcp_on_alias) var(alias) var(fd))(
When an alias is defined, this function will be called with two arguments:
the name of the alias, and the file descriptor of the corresponding session.
)
findex(tcp_on_awol)
item(tt(tcp_on_awol) var(sess) var(fd))(
If the function tt(tcp_fd_handler) is handling input from the line
editor and detects that the file descriptor is no longer reusable,
by default it removes it from the list of file descriptors handled
by this method and prints a message.  If the function tt(tcp_on_awol)
is defined it is called immediately before this point.  It may
return status 100, which indicates that the normal handling should
still be performed; any other return status indicates that no further
action should be taken and the tt(tcp_fd_handler) should return
immediately with the given status.  Typically the action of tt(tcp_on_awol)
will be to close the session.

The variable tt(TCP_INVALIDATE_ZLE) will be a non-empty string if it is
necessary to invalidate the line editor display using `tt(zle -I)' before
printing output from the function.

(`AWOL' is military jargon for `absent without leave' or some
variation.  It has no pre-existing technical meaning known to the author.)
)
findex(tcp_on_close)
item(tt(tcp_on_close) var(sess) var(fd))(
This is called with the name of a session being closed and the file
descriptor which corresponded to that session.  Both will be invalid by
the time the function is called.
)
findex(tcp_on_open)
item(tt(tcp_on_open) var(sess) var(fd))(
This is called after a new session has been defined with the session name
and file descriptor as arguments.  If it returns a non-zero status,
opening the session is assumed to fail and the session is closed
again; however, tt(tcp_open) will continue to attempt to open any
remaining sessions given on the command line.
)
findex(tcp_on_rename)
item(tt(tcp_on_rename) var(oldsess) var(fd) var(newsess))(
This is called after a session has been renamed with the three arguments
old session name, file descriptor, new session name.
)
findex(tcp_on_spam)
item(tt(tcp_on_spam) var(sess) var(command ...))(
This is called once for each session spammed, just em(before) a command is
executed for a session by tt(tcp_spam).  The arguments are the session name
followed by the command list to be executed.  If tt(tcp_spam) was called
with the option tt(-t), the first command will be tt(tcp_send).

This function is called after tt($TCP_SESS) is set to reflect the session
to be spammed, but before any use of it is made.  Hence it is possible to
alter the value of tt($TCP_SESS) within this function.  For example, the
session arguments to tt(tcp_spam) could include extra information to be
stripped off and processed in tt(tcp_on_spam).

If the function sets the parameter tt($REPLY) to `tt(done)', the command
line is not executed; in addition, no prompt is printed for the tt(-v)
option to tt(tcp_spam).
)
findex(tcp_on_unalias)
item(tt(tcp_on_unalias) var(alias) var(fd))(
This is called with the name of an alias and the corresponding session's
file descriptor after an alias has been deleted.
)
enditem()

sect(TCP Utility Functions)

The following functions are used by the TCP function system but will rarely
if ever need to be called directly.

startitem()
findex(tcp_fd_handler)
item(tt(tcp_fd_handler))(
This is the function installed by tt(tcp_open) for handling input from
within the line editor, if that is required.  It is in the format
documented for the builtin `tt(zle -F)' in
ifzman(\
zmanref(zshzle)
)\
ifnzman(\
noderef(Zle Builtins)
).

While active, the function sets the parameter tt(TCP_HANDLER_ACTIVE) to 1.
This allows shell code called internally (for example, by setting
tt(tcp_on_read)) to tell if is being called when the shell is otherwise
idle at the editor prompt.
)
findex(tcp_output)
item(tt(tcp_output) [ tt(-q) ] tt(-P) var(prompt) tt(-F) var(fd) tt(-S) var(sess))(
This function is used for both logging and handling output to standard
output, from within tt(tcp_read) and (if tt($TCP_OUTPUT) is set)
tt(tcp_send).

The var(prompt) to use is specified by tt(-P); the default is the empty
string.  It can contain:
startitem()
item(tt(%c))(
Expands to 1 if the session is the current session, otherwise 0.  Used
with ternary expressions such as `tt(%LPAR()c.-.PLUS()RPAR())' to
output `tt(PLUS())' for the current session and `tt(-)' otherwise.
)
item(tt(%f))(
Replaced by the session's file descriptor.
)
item(tt(%s))(
Replaced by the session name.
)
item(tt(%%))(
Replaced by a single `tt(%)'.
)
enditem()

The option tt(-q) suppresses output to standard output, but not to any log
files which are configured.

The tt(-S) and tt(-F) options are used to pass in the session name and file
descriptor for possible replacement in the prompt.
)
enditem()

texinode(TCP Parameters)(TCP Examples)(TCP Functions)(TCP Function System)
sect(TCP User Parameters)

Parameters follow the usual convention that uppercase is used for scalars
and integers, while lowercase is used for normal and associative array.
It is always safe for user code to read these parameters.  Some parameters
may also be set; these are noted explicitly.  Others are included in this
group as they are set by the function system for the user's benefit,
i.e. setting them is typically not useful but is benign.

It is often also useful to make settable parameters local to a function.
For example, `tt(local TCP_SILENT=1)' specifies that data read during the
function call will not be printed to standard output, regardless of the
setting outside the function.  Likewise, `tt(local TCP_SESS=)var(sess)'
sets a session for the duration of a function, and `tt(local
TCP_PROMPT=)' specifies that no prompt is used for input during the
function.

startitem()
vindex(tcp_expect_lines)
item(tt(tcp_expect_lines))(
Array.  The set of lines read during the last call to tt(tcp_expect),
including the last (tt($TCP_LINE)).
)
vindex(tcp_filter)
item(tt(tcp_filter))(
Array. May be set directly.  A set of extended globbing patterns which,
if matched in tt(tcp_output), will cause the line not to be printed to
standard output.  The patterns should be defined as described for the
arguments to tt(tcp_expect).  Output of line to log files is not affected.
)
vindex(TCP_HANDLER_ACTIVE)
item(tt(TCP_HANDLER_ACTIVE))(
Scalar.  Set to 1 within tt(tcp_fd_handler) to indicate to functions
called recursively that they have been called during an editor session.
Otherwise unset.
)
vindex(TCP_LINE)
item(tt(TCP_LINE))(
The last line read by tt(tcp_read), and hence also tt(tcp_expect).
)
vindex(TCP_LINE_FD)
item(tt(TCP_LINE_FD))(
The file descriptor from which tt($TCP_LINE) was read.
tt(${tcp_by_fd[$TCP_LINE_FD]}) will give the corresponding session name.
)
vindex(tcp_lines)
item(tt(tcp_lines))(
Array. The set of lines read during the last call to tt(tcp_read),
including the last (tt($TCP_LINE)).
)
vindex(TCP_LOG)
item(tt(TCP_LOG))(
May be set directly, although it is also controlled by tt(tcp_log).
The name of a file to which output from all sessions will be sent.
The output is proceeded by the usual tt($TCP_PROMPT).  If it is not an
absolute path name, it will follow the user's current directory.
)
findex(TCP_LOG_SESS)
item(tt(TCP_LOG_SESS))(
May be set directly, although it is also controlled by tt(tcp_log).
The prefix for a set of files to which output from each session separately
will be sent; the full filename is tt(${TCP_LOG_SESS}.)var(sess).
Output to each file is raw; no prompt is added.  If it is not an absolute
path name, it will follow the user's current directory.
)
vindex(tcp_no_spam_list)
item(tt(tcp_no_spam_list))(
Array.  May be set directly.  See tt(tcp_spam) for how this is used.
)
vindex(TCP_OUTPUT)
item(tt(TCP_OUTPUT))(
May be set directly.  If a non-empty string, any data sent to a session by
tt(tcp_send) will be logged.  This parameter gives the prompt to be used
in a file specified by tt($TCP_LOG) but not in a file generated from
tt($TCP_LOG_SESS).  The prompt string has the same format as
tt(TCP_PROMPT) and the same rules for its use apply.
)
vindex(TCP_PROMPT)
item(tt(TCP_PROMPT))(
May be set directly.  Used as the prefix for data read by tt(tcp_read)
which is printed to standard output or to the log file given by
tt($TCP_LOG), if any.  Any `tt(%s)', `tt(%f)' or `tt(%%)' occurring in the
string will be replaced by the name of the session, the session's
underlying file descriptor, or a single `tt(%)', respectively.  The
expression `tt(%c)' expands to 1 if the session being read is the current
session, else 0; this is most useful in ternary expressions such as
`tt(%LPAR()c.-.PLUS()RPAR())' which outputs `tt(PLUS())' if the session is
the current one, else `tt(-)'.

If the prompt starts with tt(%P), this is stripped and the complete
result of the previous stage is passed through standard prompt tt(%)-style
formatting before being output.
)
vindex(TCP_READ_DEBUG)
item(tt(TCP_READ_DEBUG))(
May be set directly.  If this has non-zero length, tt(tcp_read) will give
some limited diagnostics about data being read.
)
vindex(TCP_SECONDS_START)
item(tt(TCP_SECONDS_START))(
This value is created and initialised to zero by tcp_open.

The functions tt(tcp_read) and tt(tcp_expect) use the shell's
tt(SECONDS) parameter for their own timing purposes.  If that parameter
is not of floating point type on entry to one of the functions, it will
create a local parameter tt(SECONDS) which is floating point and set the
parameter tt(TCP_SECONDS_START) to the previous value of tt($SECONDS).
If the parameter is already floating point, it is used without a local
copy being created and tt(TCP_SECONDS_START) is not set.  As the global
value is zero, the shell elapsed time is guaranteed to be the sum of
tt($SECONDS) and tt($TCP_SECONDS_START).

This can be avoided by setting tt(SECONDS) globally to a floating point
value using `tt(typeset -F SECONDS)'; then the TCP functions will never
make a local copy and never set tt(TCP_SECONDS_START) to a non-zero value.
)
vindex(TCP_SESS)
item(tt(TCP_SESS))(
May be set directly.  The current session; must refer to one of the
sessions established by tt(tcp_open).
)
vindex(TCP_SILENT)
item(tt(TCP_SILENT))(
May be set directly, although it is also controlled by tt(tcp_log).
If of non-zero length, data read by tt(tcp_read) will not be written to
standard output, though may still be written to a log file.
)
vindex(tcp_spam_list)
item(tt(tcp_spam_list))(
Array.  May be set directly.  See the description of the function
tt(tcp_spam) for how this is used.
)
vindex(TCP_TALK_ESCAPE)
item(tt(TCP_TALK_ESCAPE))(
May be set directly.  See the description of the function tt(tcp_talk) for
how this is used.
)
vindex(TCP_TIMEOUT)
item(tt(TCP_TIMEOUT))(
May be set directly.  Currently this is only used by the function
tt(tcp_command), see above.
)
enditem()

sect(TCP User-defined Parameters)

The following parameters are not set by the function system, but have
a special effect if set by the user.

startitem()
vindex(tcp_on_read)
item(tt(tcp_on_read))(
This should be an associative array; if it is not, the behaviour is
undefined.  Each key is the name of a shell function or other command,
and the corresponding value is a shell pattern (using tt(EXTENDED_GLOB)).
Every line read from a TCP session directly or indirectly using
tt(tcp_read) (which includes lines read by tt(tcp_expect)) is compared
against the pattern.  If the line matches, the command given in the key is
called with two arguments: the name of the session from which the line was
read, and the line itself.

If any function called to handle a line returns a non-zero status, the
line is not output.  Thus a tt(tcp_on_read) handler containing only
the instruction `tt(return 1)' can be used to suppress output of
particular lines (see, however, tt(tcp_filter) above).  However, the line
is still stored in tt(TCP_LINE) and tt(tcp_lines); this occurs after all
tt(tcp_on_read) processing.
)
enditem()

sect(TCP Utility Parameters)

These parameters are controlled by the function system; they may be read
directly, but should not usually be set by user code.

startitem()
findex(tcp_aliases)
item(tt(tcp_aliases))(
Associative array.  The keys are the names of sessions established with
tt(tcp_open); each value is a space-separated list of aliases which refer
to that session.
)
findex(tcp_by_fd)
item(tt(tcp_by_fd))(
Associative array.  The keys are session file descriptors; each
value is the name of that session.
)
findex(tcp_by_name)
item(tt(tcp_by_name))(
Associative array.  The keys are the names of sessions; each value is the
file descriptor associated with that session.
)
enditem()

texinode(TCP Examples)(TCP Bugs)(TCP Parameters)(TCP Function System)
sect(TCP Examples)

Here is a trivial example using a remote calculator.

To create a calculator server on port 7337 (see the tt(dc) manual page for
quite how infuriating the underlying command is):

example(tcp_proxy 7337 dc)

To connect to this from the same host with a session also named `tt(dc)':

example(tcp_open localhost 7337 dc)

To send a command to the remote session and wait a short while for output
(assuming tt(dc) is the current session):

example(tcp_command 2 4 + p)

To close the session:

example(tcp_close)

The tt(tcp_proxy) needs to be killed to be stopped.  Note this will not
usually kill any connections which have already been accepted, and also
that the port is not immediately available for reuse.

The following chunk of code puts a list of sessions into an xterm header,
with the current session followed by a star.

example(print -n "\033]2;TCP:" ${(k)tcp_by_name:/$TCP_SESS/$TCP_SESS\*} "\a")

texinode(TCP Bugs)()(TCP Examples)(TCP Function System)
sect(TCP Bugs)

The function tt(tcp_read) uses the shell's normal tt(read) builtin.  As
this reads a complete line at once, data arriving without a terminating
newline can cause the function to block indefinitely.

Though the function suite works well for interactive use and for data
arriving in small amounts, the performance when large amounts of data are
being exchanged is likely to be extremely poor.