File: UPDATES

package info (click to toggle)
epic5 3.0.3-1
links: PTS
area: main
in suites: forky, sid
size: 5,328 kB
sloc: ansic: 75,810; makefile: 648; ruby: 227; python: 215; sh: 78; perl: 13
file content (6721 lines) | stat: -rw-r--r-- 285,126 bytes
EPIC5-3.0.3

*** News 2/27/2025 -- EPIC5-3.0.3 (2157 - Obturation) released here
	The third bugfix release for EPIC5-3.0

*** News 2/27/2025 -- ^X{#RRGGBB} support, $rgb()
	The $rgb() function outputs a ^X RGB code.  However, if it is
	immediately followed by a comma, such as:
		/echo $rgb(FF00FF),01 more text
	then the ",01" would be interpreted as part of the color code.
	To address this, ^X RGB codes can be surrouned by curly braces,
	which makes the syntax unambiguous.  This only works for RGB
	color codes, not for 256-color codes.

	The $rgb() function now uses this feature to fix the problem.

EPIC5-3.0.2

*** News 12/02/2024 -- EPIC5-3.0.2 (2154 - Peregrination) released here
	The second bugfix release for EPIC5-3.0

*** News 11/27/2024 -- Easier way to specify SSL server (port prefix)
	In a server desc, you have specified that a server does ssl 
	by using :type=irc-ssl: but that's a hassle.  Other clients 
	permit you to prefix the port with a + sign to indicate this,
	and so i figured, as long as i'm playing santa claus...
		/server irc.host.com:+6697

*** News 11/27/2024 -- Auto-detection of SSL servers [experimental]
	If you try to connect to an irc server's SSL port, it closes
	the socket without sending any data back.  The client now
	detects this, and if a non-ssl server closes the socket without
	sending any data back, the server's "type" will be unconditionally
	changed to "IRC-SSL" and then will try again.  I am 50% sure I am
	going to regret this, but carpe diem, i suppose.

*** News 11/26/2024 -- New flag, /EXEC -EXIT {code} (rb fusion)
	Although there is already /EXEC -END {code} and /WAIT -CMD to 
	run callbacks after a process ends, neither of those interact
	with /ON EXEC_EXIT or /SET NOTIFY_ON_TERMINATION.

	Fusion asked if there was a way to defeat the "process exited"
	message of /SET NOTIFY_ON_TERMINATION on an /exec-by-/exec 
	basis.  So the flag /EXEC -EXIT {...} was added.  I will admit
	that having two ways to run code seems confusing, so I will just
	say that this callback is run *instead of* /on exec_exit or 
	/set notify_on_termination handling.  So you should only use it
	to format messages about the process exiting.

	You can do  /exec -exit {#} ls   to (eg) suppress all that

*** News 11/26/2024 -- /AWAY restatement
	When you do /AWAY <message> you tell the client you are away.
	The client then tells the server you are /AWAY.  Later on, you
	do /AWAY and the client tells the server you are no longer away.

	But the notion of "am i away?" is a client-side thing, and it is
	triggered on whether you have set an away message, and not whether
	the server has acknowledged your away-ness.

	So $serverctl(GET x AWAY) returns your away message, and $isaway()
	is triggered based on whether you have an away message.  Fusion
	pointed out there was no way to know whether the server thought you
	were away or not, since the client has never tracked the 305/306
	numerics.

	So now the client tracks the 305/306 numerics which tell us whether
	the server thinks we're away or not.   This is available via
	$serverctl(GET x AWAY_STATUS)

	For backwards compatability reasons, _absolutely nothing_ that changes
	its behavior if you are "away" is changing here.  That continues to
	be controlled only on whether you have set an away message (or not)

*** News 09/05/2024 -- EPIC5-3.0 (2139 - Brigadoon) released here
	This is the third production release of EPIC5

*** News 08/23/2024 -- EPIC5-2.6 (2137 - Sedulous) released here
	This is the third release candidate for EPIC5-3.0

*** News 08/20/2024 -- EPIC5-2.4 (2136 - Sedulous) released here
	This is the second release candidate for EPIC5-3.0

*** News 08/18/2024 -- EPIC5-2.2 (2135 - Sedulous) released here
	This is the first release candidate for EPIC5-3.0

*** News 08/18/2024 -- /SET RANDOM_SOURCE now has no effect
	EPIC has traditionally had 4 random number generators:
	They have been replaced with just using OpenSSL's BYTES_random().
	Using /set random_source now has no effect.

*** News 08/18/2024 -- SOCKS support retired
	SOCKS4 and SOCKS5 support has been sunset.  It was 28 years old.

*** News 08/18/2024 -- Flood control support retired
	Flood control is a historical ircII feature that solved
	a problem we used to have with pre RFC1459 servers.
	Flood detection, as a concept, hasn't been relevant in
	30 years.  The /on flood and /set flood_* variables have
	been retired.

*** News 06/14/2024 -- EPIC5-2.1.13 (2125 - Sockdolager released here
	People like to see releases occasionally!

*** News 06/12/2024 -- I didn't document stuff for a long time
	You'll notice a lot of things documented on this date.  That's my fault.
	I've been making changes but not updating this file with the docs.
	These changes were not all made on this day, but in the past year. ;-) 
	I included the commit id in case that is useful.

*** News 06/12/2024 -- Restatement of $rgb() behavior  (commit 2123)
	Whatever it was I said about $rgb(), i take it back, now that proper
	RGB/24-bit color is now supported.  This is how it works now:

	$rgb() takes two, space-seperated colors -- a fg color, and bg color.
	A color can be one of these:
		- 		The color is omitted (it won't be changed)
		1,2,3		The color as three decimal numbers (0-255)
		1 2 3		The color as three space-separated decimal 
				numbers (0-255)
		#AABBCC		The color as a hash and 6 hex digits 

	The return value is a full properly formatted ^X code that you can 
	/echo or whatever to set the color.

	The fg and bg colors can be in any of the above formats, and you can
	mix and match them.

	If you pass it a syntax error, it will return an empty string.

	The JSON arguments are now:
		current:	r	g	b	attr
		added:		bg_r	bg_g	bg_b

	You can see some useful examples in the 'regress/color256' script.

*** News 06/12/2024 -- RGB/24 bit color support (rb everybody) (commit 2123)
	EPIC now supports RGB/24-bit colors natively.
	Specify a RGB color with ^X#, as in:
		^X#AABBCC		set the fg only
		^X#AABBCC,		you can use comma, bg is optional
		^X#AABBCC,DDEEFF	you can specify both at the same time
		^X#,DDEEFF		set the bg only (note the leading comma)
	EPIC will also parse ^[[28m and ^[[38m RGB codes and convert them to
	the above formats.

	EPIC no longer approximates RGB colors into the 256-color space.  
	RGB colors are output as RGB colors.  So your terminal emulator needs
	to support that, or you'll have a bad time.  Let me know if this is a
	problem for you.

	None of this supplants 256 colors, which are still supported exactly as 
	they have always been.  However, because every terminal emulator supports
	256 colors, and because colors 0-15 are the same between "basic" colors
	and 256 colors, "basic" colors map to 256 colors now.  Let me know if your
	terminal emulator doesn't support 256 colors!

*** News 06/12/2024 -- status sub-expandos now protect from expansion (2106,2123)
	Some status bar expandos use "sub-expandos".  The primary expando
	is controlled by (eg) /set status_mode, which is used by %+.
	Usually this is set to " +%+ ", so the sub-expando value (the
	channel's mode) is substituted in the %+.  But what happens if
	you have /set status_does_expandos on, and someone does something
	that causes your status bar to include rogue data?

	Example:
		/mode #chan +k $exec(rm -rf ~)
	You wouldn't want that expanded!  So all subexpandos now get 
	"quoted" so that doesn't happen.  

	The above is a bad example, because /set status_mode has always
	done this protection.  I was asked to extend this protection to
	every status subexpando.

*** News 06/12/2024 -- Add $hex(<digits> <number>) (commit 2114)
	Convert a decimal number into a hex number without fuss.
	Also ensures correct zero-padding.  Useful for converting
	decimal numbers to RGB portions!

	Example:
		$hex(255 2)	returns "FF"
		$hex(10 2)	returns "0A"
		$hex({"value": 10, "digits": 2}) 	returns "0A"

*** News 06/12/2024 -- Improvements to 'reconnect' script (commit 2112)
	The 'reconnect' script now supports these new /set's:
	  * /set auto_reconnect_delay_method [ON|OFF]
	  * /set auto_reconnect_join_delay <seconds>

	When /set auto_reconnect_delay_method OFF (default behavior)
	- All chnanels are rejoined at once as soon as you are connected.

	When /set auto_reconnect_delay_method ON 
	- Channels will be joined one-at-a-time, one-per-second,
	  starting after /set auto_reconnect_join_delay seconds, 

*** News 06/12/2024 -- Add /WINDOW LISTER "refnum [refnum]" (commit 2111)
	Given a double-quoted space-separated list of words, do what 
	/WINDOW LIST would do, but only for those windows.  This allows 
	you to sort the window listing order.

*** News 06/12/2024 -- Add $windowctl(GET x CURRENT_CHANNEL) (commit 2109)
	This tells you what the current channel is for a window.
	How did it take so long for me to add this?

*** News 06/12/2024 -- autoconf, make uninstall (commit 2106)
	EPIC's configure script now works with autoconf-2.71, and the
	Makefile now includes 'make uninstall' which does its best.

*** News 06/12/2024 -- /set blank_line_indicator (rb harzelien) (commit 2105)
	The default value of this is "unset" which is the historical behavior.
	Traditionally, a window starts off its life empty, and from time to 
	time parts of it might be empty (ie, after a clear).  If you want to see
	that a line is explicitly "missing" rather than blank, this set will be
	used for each blank line

	Example:
	  /set blank_line_indicator ~
	       makes missing lines ~, like vi does

*** News 06/12/2024 -- Add $symbolctl(GET <var> 0 BUILTIN_VARIABLE ORIG_DATA) (commit 2105)
	Builtin variables (ie, /SET) are initialized when the client starts.
	You can later change the value.  But if you want to reset the value
	to the "initial" value, you can fetch that value here.
	
	Example:
	  /set banner $symbolctl(GET banner 0 BUILTIN_VARIABLE ORIG_DATA)

*** News 07/17/2023 -- New /ON, /ON PING (rb skered)
       The /ON PING will be thrown when the server sends you a PING
       after you've been idle for some time.   Yes, you can suppress
       the default handling, which will probably result in you being
       disconnected.  Whatever floats your boat.

               $0    - The server that sent you the PING
               $1-   - The ping message itself.  
                       Do /QUOTE PONG $1- to respond

*** News 10/26/2022 -- Basic CAP support, /load capctl (zlonix)
	The /load capctl script, written by zlonix, provides
	the /CAPCTL command, which permits you to register CAPs
	that you wish the client to activate on your behave when
	you connect to a particular server.  

	Turning on arbitrary CAPs results in unspecified behavior
	(ie, if it breaks things, that's not a bug)
	Turning on 'account-tag' or 'multi-prefix' should be safe.

*** News 10/26/2022 -- New built in function, $tags() for CAP tags (zlonix)
	If you turn on the CAP 'account-tag', some servers (which 
	includes ZNC) will start sending you CAP tags, which you can 
	fetch with $tags().  How to parse and/or use them is left for
	future expansion.  Stay tuned.

*** News 10/26/2022 -- New server description field "cert" (client certs)
	Some servers support "client certificates" for their SSL connections.
	Client certificates are similar to server certificates -- they verify
	your identity to the server.  Some networks use this for authentication
	with servers.

	You create a *.pem cert, using the same instructions as for every other
	client, and then provide that file in the "cert" field:

		/server irc.network.com:6697:type=irc-ssl:\
			cert=/home/user/.certs/network.pem

	"cert" does not support tildes in filenames.  Use $realpath() for that.

*** News 08/11/2022 -- New concept to the /CLEAR command
	When you /CLEAR a window, it's empty, right?  But quite a few
	operations can result in a window getting bigger.  You can 
	/WINDOW GROW or /WINDOW CLEARLEVEL or you can resize the screen.
	Usually when everything in a window fits in the new size, but
	does not take up all of the room, EPIC has to decide what to do
	with the new space.  Does it redisplay info that was previously
	scrolled off? ("pinning the bottom")  Does it display blank lines
	at the bottom of the window? ("pinning the top").

	People disagree on this point, and that's part of why the 
	/WINDOW SCROLLADJ feature exists.  But it's clear that a second
	signal is important -- the /CLEAR command.

	Each time you /CLEAR (or /WINDOW CLEAR) it will set the window's
	"clear point".  When EPIC adjusts a window's scrollback to fill
	in blank lines, the top of the window can never go before the 
	"clear point", and so that's how it will decide whether you 
	want to redisplay old stuff, or keep blank lines on the bottom.

	The /UNCLEAR command unconditionally "pins the bottom", and for
	backwards compatability, it ignores the "clear point".  However, 
	if you want it to honor the clear point, use /UNCLEAR -NOFORCE

*** News 07/24/2022 -- New commands, /CTCP_CLOAK and /CTCP_UNCLOAK
	These are implemented by the /LOAD ctcp script, which is 
	part of the default experience

	"Cloaking" a CTCP means to treat a built-in CTCP as though
	it were not a built-in CTCP.
	   1. It will not show up in CTCP CLIENTINFO
	   2. Built-in handlers will NOT be called, but it will be
	      treated like any ordinary unrecognized CTCP 
	      (with /on ctcp_request and /on ctcp)

	Usage:
		/ctcp_cloak		List all cloaked CTCPs
		/ctcp_uncloak		List all uncloaked CTCPs
		/ctcp_cloak name	Cloak CTCP "name" 
		/ctcp_uncloak name	Uncloak CTCP "name"

	Zlonix wanted to do this:
		/ctcp_cloak FINGER

*** News 07/24/2022 -- $ctcpctl(SET x ACTIVE 0|1) and $ctcpctl([IN]ACTIVE)
	There are some new $ctcpctl() facilities:

	  $ctcpctl(SET x ACTIVE 0)	- Make a CTCP inactive (cloaked)
	  $ctcpctl(SET x ACTIVE 1)	- Make a CTCP active (uncloaked)
	  $ctcpctl(ACTIVE)		- Return all active CTCPs
	  $ctcpctl(INACTIVE)		- Return all inactive CTCPs

	All CTCPs start their life off as ACTIVE automatically.
	The user can then inactivate them later.
	These facilities are used for /ctcp_cloak and /ctcp_uncloak (above)

*** News 06/15/2022 -- EPIC5-2.1.12 (2078 - Secondment) released here
	This release improves build behavior for openbsd and others.

*** News 06/05/2022 -- EPIC5-2.1.11 (2071 - Adynaton) released here
	This release is entirely quality improvements, focused on 
	removing (rather than handling) type punning, and signed-
	pointer issues where c99 behaves differently from c90.

*** News 05/22/2022 -- EPIC5-2.1.10 (2059 - Casuistry) released here
	This release is entirely quality improvements now that we've 
	to C99 -- addressing undefined behavior, etc.

*** News 05/09/2022 -- EPIC5-2.1.9 (2053 - Decoupage) released here
*** News 05/09/2022 -- EPIC5-2.1.8 (2052 - Decoupage) released here
	I screwed up the release of 2.1.8 so I redid it as 2.1.9

*** News 05/09/2022 -- EPIC is now a C99 program
	As of this release, EPIC is now expected to be a well-defined C99
	program.  This release has eliminated all type punning/aliasing 
	shenanigans (or at least given the compiler fair warning about it)
	Type-punning pointer casts are done through unions, as C99 requires.
	Consequently, EPIC is now compiled with -O2, which I hope I don't come
	to regret.

*** News 05/02/2022 -- New $info() values, $info(Z:*)
	$info(Z:configure_args)		returns what you passed to 'configure'
	$info(Z:compiler_version)	returns "$CC -v" that was used
	$info(Z:final_link)		returns the command used to link epic5
	$info(Z:cflags)			returns the CFLAGS used during build
	$info(Z:libs)			returns the libs used during build
	$info(Z)			returns all the above in a json dict

*** News 04/10/2022 -- Window /LOGs now don't require logged window to exist
	There is a question -- do logs log _a particular window instance_
	or do they log _any window matching a descirption_?  Previously you
	could not /log add a window that didn't exist because the window 
	lookup failed.  

	Starting now, you can /log add any window you want even if the window
	does not exist -- you can /log add a refnum, a name, or a uuid, and
	any window that ends up using that refnum, name, or uuid, will log
	to the log; not just the window that existed when the log was created.

	On the flip side of this, if you are logging a window refnum and then
	you change a window's refnum, it might stop logging -- so if you want
	to log _a particular window_, log its uuid, for example:
		/log add $windowctl(GET 5 UUID)

*** News 04/08/2022 -- Restatement of /WINDOW QUERY
	The behavior of /WINDOW QUERY, /WINDOW ADD, and /WINDOW REMOVE 
	was unified some time ago, but the behavior was inconsistent
	and unclear.  So the code has been harmonized and now has the
	following behavior:

	/WINDOW QUERY nick
		Add 'nick' to target list and make current target (query)
	/WINDOW QUERY nick1,nick2
		Add 'nick1' and 'nick2' to target list and make the last
		target (nick2) the current target (query)
	/WINDOW QUERY
		Remove the current target (query) from the target list.
		If there are other targets in the target list, then the 
		previous query will become the new current target (query)
	/WINDOW -QUERY
		Remove all targets from the window's target list.
		There will be no query.

	/WINDOW ADD nick
		Add 'nick' to target list, but not as the current target
	/WINDOW REMOVE nick
		Remove 'nick' from the target list.  If you REMOVE the
		current target (query) then another target will become
		the new current target (query)

	Don't forget the SWITCH_QUERY keybinding, to cycle between the
	targets in a window.

*** News 04/06/2022 -- New functtions $pledge() and $unveil()
	These functions call the OpenBSD syscalls, if you're running on 
	OpenBSD.

	$PLEDGE()
	======================
	Simple Example:
		/@pledge(stdio rpath inet dns tty proc)
	Read the man page for more information.
	Returns:
		empty string - pledge(2) is not supported
		anything else - the return value of pledge(2)

	Complex examples:
		/@pledge({"test": true})
			Return 1 if $pledge() is supported, and 0 if not.

		/@pledge({"promises": "...", "execpromises": "..."})
			Pass these values to the 'promises' and 'execpromises'
			parameters of pledge(2)


	$UNVEIL()
	====================
	Simple Example:
		/@unveil(/path/to/directory r)
		/@unveil(/path/to/another/directory wc)
		/@unveil()
	Read the man page for more information
	Returns:
		empty string - unveil(2) is not supported
		anything else - the reteurn value of unveil(2)

	Complex examples:
		/@unveil({"test": true})
			Return 1 if $unveil() is supported, and 0 if not.

		/@unveil({"path": "/path/to/directory", "permissions": "r"})
		/@unveil({"path": "/path/to/another/directory", "permissions": "wc"})
		/@unveil({"close": true})
			Does the same thing as the simple example

*** News 04/04/2022 -- TCL support removed
	TCL support passed away quietly this afternoon.  It was 19 years old

*** News 04/04/2022 -- Install target de-cruftified
	EPIC has been doing lots of strange stuff when you 'make install'
	since its earliest days.  But in these modern worlds with silicon
	chips and whatnots, it makes little sense to compile our own version
	of test and ship a weird version of 'bsdinstall'.  Make install
	is now optimized for how package maintainers actually use it, so
	hopefully they'll not have to patch the heck out of epic to make 
	it behave like all the other programs

*** News 03/28/2022 -- New /SET, /SET AUTOMARGIN_OVERRIDE (default OFF)
	Previously, automargin support was conditionally built into 
	the client at compile time.  The default was to include it.
	The support is now always included.

	Traditionally, the use of the last column depended on whether the 
	client was convinced your terminal emulator safely supported it.
	At the request of skered, this /SET was added that allows you
	to force it, even if the client doesn't know how to safely turn
	it on.  Therefore, using this /SET may cause display corruption
	if the final column is used -- and this would not be considered
	a bug in the client.

	/SET AUTOMARGIN_OVERRIDE ON
		Try to use the final column, even if you aren't sure if
		it can be done safely -- (at your own risk -- if you have
		screen corruption, this is "normal")
	/SET AUTOMARGIN_OVERRIDE OFF
		Use (or don't use) the final column based on whether the
		client is sure (or isn't sure) if it's safe.

*** News 03/27/2022 -- EPIC5-2.1.7 (2011 - Cinereous) released here

*** News 03/21/2022 -- New built in function $rgb()
	There are three ways to call the $rgb() function:
	   1. As three numbers
		$rgb(255 0 255)
	   2. As a string
		$rgb(#FF00FF)
	   3. As kwargs
		$rgb({"r":255, "g":0, "b":255, "attr":true})

	This will return a string that will change to the indicated color.
	(ie, it is of the form ^Xab, where "ab" are two hex digits)
	The kwarg version is the only way to prevent prefixing with ^X
	(ie, to get the color number).   The 'attr' flag is optional.

*** News 03/06/2022 -- New serverctl(DONT_CONNECT), sets the -s CLI
	The -s command line option ("dont connect to a server") can now
	be set by your ~/.epicrc with 
		@serverctl(DONT_CONNECT 1)
	Obviously this only matters at client startup.  Setting it at
	any other time is pointless

*** News 03/06/2022 -- New math operator: "??" -- Coalesce
	The coalesce operator is a binary operator that returns the left
	value if the left value is true-y, otherwise returns the right
	value.  Therefore, "x ?? y" is shorthand for "x ? x : y"
	The operator precedence is the same as "||".

*** News 03/06/2022 -- New /set BROKEN_AIXTERM
	Some terminal emulators are silly and don't/can't/won't support
	bold and color at the same time, instead insisting that you send
	proprietary/non-standard ansi codes that were pioneered by a
	vendor that also couldn't get its terminal emulator to properly
	emulate real terminals.

	If you are using such an emulator, you can /set broken_aixterm
	and bold colors will send aixterm codes instead of standard codes.

*** News 02/06/2022 -- What is $serverctl(GET x OPEN)?
	Zlonix asked me to document "what does OPEN mean for a server?"

	A server is "open" when it has a file descriptor open.  
	This is entirely different from whether it is connected to IRC.
	(for that, use $serverctl(GET x CONNECTED))

	When a server is "open" then it is under the active management
	of the client.  When a server is "closed" then the client has 
	stopped paying attention to it and nothing further will happen
	with it until the user (or script) does something.

	During the course of a connection, several *different*
	file descriptors get used by a server as it goes through the
	server states:

	Server State	Is "Open"?	What is fd used for?
	-------------	-------------	----------------------------
	CREATED		No
	RECONNECT	No
	DNS		Yes*		DNS helper subprocess
	CONNECTING	Yes*		socket that is not yet connect()ed
	SSL_CONNECTING	Yes*		connected socket not yet ssl'd
	REGISTERING	Yes		socket that is trying to get on irc
	SYNCING		Yes		socket that is on irc
	ACTIVE		Yes		socket that is on irc
	EOF		Yes*		socket that has seen an end-of-file
	ERROR		Yes*		socket that has fatal error
	CLOSING		Yes*		socket that is shutting down
	CLOSED		No		
	DELETED		No		

	* - Just because a server is "open" does not mean that it is useful
	    for IRC purposes.  It only means that the fd has not yet been
	    closed at the OS level.  These states still have an 'fd that 
	    has not been close(2)d', but don't represent useful states

*** News 02/05/2022 -- New $serverctl(GET x NEXT_SERVER_IN_GROUP)
	This will return the server that /server + will connect to

*** News 11/11/2021 -- $json_implode() takes kwargs, adds "compact mode"
	The standard behavior of $json_implode() is to take a single word 
	as an argument, the root of a assign tree to convert into a json 
	object.
		$json_implode(q)	# Collapse $q.* into json object

	The standard behavior of $json_implode() is to return a "human-
	formatted" string, with newlines and indentation.  This is nice 
	for looking at, but isn't appropriate for every situation.

	So $json_implode() is our pilot test for json object argument lists.
	It now supports the following two arguments:
		root	  string    The root of the assign tree (required)
		compact	  boolean   false - return human readable (default)
				    true - return one line

	Example:
		@ q.one = [val1]
		@ q.two = [val2]
	then,
		$json_implode({"root": "q", "compact": true})
	would return
		{"one":"val1","two","val2"}

	
*** News 11/18/2021 -- New concept -- json object as argument list ("kwargs")
	You'll start noticing something new happening more and more.
	The idea of "json object as an argument list" is being introduced,
	to allow passing of parameters in a more deterministic way.

	JSON objects (also known as "dicts") are a collection of key-value pairs.
		{ "field1", "val1", "field2", "val2"}
	You already operate on these with $json_implode() and $json_explode().
	Most RESTful APIs in the world send and receive them.  

	For backwards compatability reasons, of course, this feature will be
	rolled out on a case-by-case basis whenever it seems appropriate.
	But, occasionally, a new feature might be available only through json
	object arguments because it makes sense.

*** News 10/06/2021 -- EPIC5-2.1.6 (1981 - Impignorate) released Here

*** News 09/21/2021 -- Another round of clang static analysis
	Somewhere around here we did another round of static analysis
	with clang's analyzer, and fixed whatever it recommended.

*** News 09/19/2021 -- "reconnect" script has been reimplemented
	The reconnect script has been renamed to "reconnect.orig".
	If you want the original behavior, just /load reconnect.orig
	The new script is based on improvements we've made in epic5
	to make reconnection logic more reliable and dependable.

*** News 09/11/2021 -- New option,  $serverctl(GET x PADDR)
	The $serverctl(GET x PADDR) returns the server's presentation
	address.  This is "1.2.3.4" for ipv4, or "2600::1" for ipv6.
	This represents the IP address we're connected to.

*** News 09/11/2021 -- New /window operation, /WINDOW CLEARREGEX <regex>
	This was requested by Zlonix.

	The /WINDOW CLEARREGEX <regex> operation removes all items from
	the window's lastlog whose text (what you see on the screen) 
	matches the <regex>.  Removing them from the lastlog removes
	them from the scrollback and your screen as well.  This is useful
	for making conversations with annoying bots go away retroactively.
	Example:	/window clearregex annoybot

*** News 09/11/2021 -- New /ON, /ON RECONNECT_REQUIRED 
	The /ON RECONNECT_REQUIRED hook is thrown by the client when it
	feels that a reconnection intervention is appropriate.  
	We are going to start this small, and build in more use cases
	as we go forward.

	  1. When server is in ACTIVE state and the socket dies
	  2. When a write to the server fails

	In these situations, your script should take whatever measures
	are necessary to save the state of windows, channels, etc,
	in preparation for the upcoming disconnection.

	This /ON is thrown while the server is in a state of chaos.
	The internal state of the system is wrong, and you must assume
	that the connection to the server has already been lost.  

	YOU MUST _*_*_NOT_*_*_ try to do anything clever in this /on.
	If you have this idea of moving windows or servers or channels,
	please don't!  I recommend setting up /timer's and /defer's so
	you can do those sorts of changes away from the blast radius.

	YOU MUST _*_*_NEVER_*_*_ attempt to do any direct intervention
	in this /on -- instead, you should save information and create 
	a timer to deal with the problem later.  YOU HAVE BEEN WARNED.

	/ON RECONNECT_REQUIRED currently provides this information
		$0 	The server that unexpected failed on us.
			Reconnection will be required.  Remember,
			you cannot do reconnection in the /on, you
			have to schedule it to happen "later"

*** News 09/08/2021 -- New /window operation, /WINDOW CLEARLEVEL [levels]
	This was requested by Zlonix.

	The /WINDOW CLEARLEVEL [levels] operation removes all items from
	the window's lastlog of the levels you specify.  Removing them
	from your lastlog removes them from your screen as well.  This is
	useful for making noise go away retroactively
	Example:	/window clearlevel joins,parts,quits,kicks

*** News 09/08/2021 -- New statement type: block-with-arglist
	The ircII syntax now supports a new statement type, which I'm 
	calling a "block with arglist"

		(arglist) {block}

	You could also think of this as an "inline anonymous function".
	For the purposes of the statement, $* is modified by (arglist)
	and then {block} is run.

	Please note there are several caveats to this:
	  1. Arglist is not magic, it is syntactic sugar.  So it creates real 
	     local variables, and those local variables have their ordinary 
	     scope.  So they will persist after the end of the statement.
	  2. The statement only modifies $* during the statement itself.
	     So after {block} is run, $* goes back to what it was originally.
	  3. Note that this is a *statement* and not a *block* so if you are
	     working with something that expects a block, wrap it in {}s to
	     create a block.
		if (# == 3) {(arg0, arg1, arg2) {... code ...}}
	  4. Arglist processing isn't "free" so doing it in a tight loop will
	     be slower than doing it outside of the loop

*** News 09/07/2021 -- Pass window refnum in /ON CHANNEL_LOST
	The /ON CHANNEL_LOST hook now provides the window that a channel
	was in as $2
		$0 - The server refnum of a channel
		$1 - The name of a channel
		$2 - The window refnum of a channel

*** News 09/06/2021 -- Functions for JSON document handling
	Some time ago some new functions appeared for handling
	JSON documents, but they were not documented.  So now
	I am going to document them!

	CONVERTING JSON DOCS TO /ASSIGNS
	--------------------------------
	Syntax:    $json_explode(varbase json-document)

	An example is worth a thousand words:

	    $json_explode(e {"one": 1, "two": { "sub1", "hi", "sub2", "bye" })
	will result in three new assigns
		$e[one]		-> 1
		$e[two][sub1]	-> hi
		$e[two][sub2]	-> bye

	CONVERTING ASSIGNS TO JSON DOCS
	-------------------------------
	Syntax:	   $json_implode(varbase)

	This converts everything under $varbase[*] into a JSON doc.

	HANDLING ERRORS
	--------------
	If there is an error, $json_error() will tell you about it.

*** News 09/06/2021 -- SSL handling improvements (take 2)
	The other day I posted some info on how the SSL handling has been 
	improved.  But then things got revamped again, so I delete the old 
	info and am replacing it with this info.

	1) Step one - You connect to an SSL server
	 You connect to an irc server doing something like
		/SERVER irc.example.com:6697:type=irc-ssl
	 This goes through connect()ing to the server, and then doing an 
	 SSL_connect() to set up the certificate exchange and TLS negotiation.

	2) Step two - The SSL handshake is completed
	 After the SSL_connect() step has completed, we now have a fully 
	 functioning TLS socket with an SSL certificate.  Before we use the 
	 TLS socket, we're supposed to verify we trust the SSL certificate, 
	 to ensure we're talking to who we think we are.

	3) Step three - The client watches the certificate verification
	 OpenSSL "verifies" the certificate and the client provides
	 a C function to tag along for the ride.  The callback function
	 is called every time:
	   (1) An error is discovered, or
	   (2) There are no (further) errors in a certificate.
	 Thus, if a certificate is trusted, it will only report 
	 "no problems" for each link in the chain.

	 This permits the client to trap and categorize every error that
	 happens - some certificates have multiple problems!  There are
	 three buckets the client uses:
		(1) Self-signed certificates
		(2) Incorrect-hostname certificates
		(3) Any other (serious) error
	 The client tracks the "most serious error" (if there is one),
	 using the above priorities.

	4) Step four - The cliet offers you /ON SSL_SERVER_CERT
	  The client sets $serverctl(SET <refnum> SSL_ACCEPT_CERT -1)
	  and then throws /ON SSL_SERVER_CERT.   The use of -1 is on
	  purpose so the client can determine whether your /ON handler
	  sets it to 0 or 1.  

	  If your handler does a $serverctl(SET <retval> SSL_ACCEPT_CERT 0|1)
	  then that is taken as the final disposition of the handling, and 
	  nothing further occurs. (ie, it skips the rest of the steps)

	  Parameters of /ON SSL_SERVER_CERT
		$0  The fd of the socket
		    (Use $serverctl(FROM_SERVER) to get the server refnum)
		$1  Certificate Subject, url-encoded
		$2  Certificate Issuer, url-encoded
		$3  Bits used in the public key 
		$4  OpenSSL error code of the "most serious error"
			(18 is "Self-signed certificate",
			 62 is "Hostname mismatch",
			 everything else is irregular/bad)
		$5  The SSL regime being used (ie, TLSv1.2)
		$6  The Certificate Hash
		$7  Was there a hostname mismatch?  0 = no error, 1 = error
		$8  Was there a self-signed error?  0 = no error, 1 = error
		$9  Was there another (serious) error?  
			0 = no other error 1 = other error
		$10 Was there any error of any kind?  
			0 = no errors of any kind, 1 = some kind of error

	5) Step five - The client makes a provisional decision
	  Next, the client looks at the errors and decides whether
	  it thinks the cert is ok.  
	    * Cert has no errors 		   -> ACCEPT
	    * Cert has "self signed" or "wrong hostname" error
	      and /SET ACCEPT_INVALID_SSL_CERT ON  -> ACCEPT
	    * Cert has "self signed" or "wrong hostname" error
	      and /SET ACCEPT_INVALID_SSL_CERT OFF -> REJECT
	    * Cert has any serious error	   -> REJECT
	  The client sets this provisional value with 
		$serverctl(SET <refnum> SSL_ACCEPT_CERT 0|1)

	6) Step six - The client offers you /ON SERVER_SSL_EVAL
	  Then, the client hooks /ON SERVER_SSL_EVAL.  At this point,
	  all of the information your script needs to make a fully 
	  informed decision to accept or overrule the client's choice
	  is available.  Your handler is not obligated to make any
	  change, but it certainly can if it wants to 

	  Parameters of /ON SERVER_SSL_EVAL
		$0  The server refnum
		$1  The "ourname" of the server (what you /server'd to)
		$2  Was there any error at all? 
			0 = no errors of any kind   1 = some kind of error
		$3  Was there a hostname mismatch?  0 = no error, 1 = error
		$4  Was there a self-signed error?  0 = no error, 1 = error
		$5  Was there another (serious) error?
			0 = no other error, 1 = other error
		$6  What does the client suggest?
			0 = reject certificate, 1 = accept certificate

	  Using $serverctl() to get info about the certificate
	  Use $serverctl(GET <refnum> <item>) where <item> is:
		SSL_CIPHER		The encryption cipher being used
		SSL_PEM			The certificate (in PEM format)
		SSL_CERT_HASH		The certificate's hash
		SSL_PKEY_BITS		The bits in the public key
		SSL_SUBJECT		Who the cert was issued to
		SSL_SUBJECT_URL		Who the cert was issued to (url-encoded)
		SSL_ISSUER		Who issued the cert
		SSL_ISSUER_URL		Who issued the cert (url-encoded)
		SSL_VERSION		What version of SSL being used (ie, TLSv1.2)
		SSL_SANS		Subject Alternate Names in the cert
		SSL_CHECKHOST_ERROR	Hostname Mismatch error - 0 (no) 1 (yes)
		SSL_SELF_SIGNED_ERROR	Self-signed error - 0 (no) 1 (yes)
		SSL_OTHER_ERROR	        Any other (serious) error - 0 (no) 1 (yes)
		SSL_MOST_SERIOUS_ERROR	The OpenSSL error code of the most serious error
					18 (self-signed) and 62 (hostname mismatch)
					are considered non-serious (routine) errors
		SSL_VERIFY_ERROR	Any error at all - 0 (no) 1 (yes)
		SSL_ACCEPT_CERT		Is this cert headed for acceptance?  0 (no) 1 (yes)

	   Use $serverctl(SET <refnum> SSL_ACCEPT_CERT 0) to reject the cert
	   Use $serverctl(SET <refnum> SSL_ACCEPT_CERT 1) to accept the cert
	   If you don't do anything, the client will do the most reasonable thing

	7) Step seven - The client moves forward
	  Finally, everyone has had a chance to weigh in.  
	  Whatever the value of $serverctl(GET <refnum> SSL_ACCEPT_CERT) 
	  is after all this, is used to accept or reject the SSL connection.

*** News 09/02/2021 -- Configuring SSL Ciphers
	Now, if you know what you are doing, you can set
	the SSL Ciphersuites that the client will use for
	SSL connections, via
		/SET SSL_CIPHERS <stuff>
	If you don't know what you're doing, don't touch.
	The default value is "unset", which means we let
	openssl choose the ciphers for us.

*** News 09/02/2021 -- SSL Handling improvements
	[replaced -- see above]

*** News 08/30/2021 -- New server description field "ssl-strict"
	[this feature was removed]

*** News 05/29/2021 -- EPIC5-2.1.5 released here (Fecund) (Commit id: 1945)

*** News 05/25/2021 -- Updated configure to autoconf-2.69
	We upgraded configure from autoconf-2.13 to 2.69 here.
	Along the way, we fixed the support for python3.8+

*** News 05/20/2021 -- Windows have UUIDs ($windowctl(GET x UUID))
	Every window now receives an immutable UUID when it is 
	created.  This UUID is globally unique and cannot be changed.

	Although the UUID is not user-facing (as in /window list),
	you can get it with $windowctl(GET refnum UUID).

	The UUID is an lval (that is, it does not contain hyphens),
	so you can use it as part of a variable name if you wish.

*** News 05/20/2021 -- New concept, "claiming channels"
	When EPIC receives a protocol JOIN message for a channel, 
	it has to decide what window to put the channel in.  
	A common request has been the opportunity to let scripts 
	decide where new channels should go, rather than it being 
	hardcoded.

	So now when a JOIN is received, an /ON will be thrown 
	(see below) which is an invitiation for your script to do
	whatever preparatory work for the channel you see fit.

	One thing in particular is /WINDOW CLAIM (see below), which
	tells EPIC which window the channel should be put into.
	As part of this process, EPIC will suggest a window the 
	channel should go into, unless you choose to overrule that.

	*** NONE OF THIS APPLIES TO /JOIN or /WINDOW CHANNEL
	    when the user is deliberately moving a channel they are
	    already on between windows.  You can't stop that.

*** News 05/20/2021 -- New /ON, /ON CHANNEL_CLAIM 
	The /ON CHANNEL_CLAIM is thrown when a JOIN is received,
	but before the client has assigned the channel to a window.
		$0 - The server refnum 
		$1 - The channel being joined
		$2 - The window refnum epic proposes to put the
		     window into.

	If your handler does /WINDOW CLAIM in any particular window,
	then the channel will go to that window that you specify.

	At this time, you can only do /WINDOW CLAIM in a window that
	is connected to the server in $0. In the future, this will
	probably change.  But for now, a /WINDOW BIND in a "window
	connected to the wrong server" is treated as invalid.

	If you do not do a valid /WINDOW CLAIM, then the channel will
	go into the window proposed by EPIC.

*** News 05/20/2021 -- New /WINDOW operation. /WINDOW CLAIM #channel
	During the handling of an /ON CHANNEL_CLAIM, you may perform
	the /WINDOW CLAIM operation to direct EPIC to put the new
	channel into a particular window.

	You must pass the name of the channel as a paraemter.
	It is available as $1 in the /on.

	You may only claim the channel in a window that is connected
	to the correct server (from $0).  Attempting to claim a 
	channel in a window connected to the wrong server is invalid
	and has no effect.   This will change in the future.

*** News 05/20/2021 -- $uuid4() can now return lvals (NODASHES)
	You can now get a uuid4 that is valid as an lval if you
	supply NODASHES as the only argument
		@ a.$uuid4(NODASHES) = [whatever]

	ALL ARGUMENTS TO THIS FUNCTION ARE RESERVED FOR FUTURE EXPANSION.
	There are now two defined behaviors:
		$uuid4()	 - UUID4 with dashes
		$uuid4(NODASHES) - UUID4 without dashes
	EVERYTHING ELSE IS UNDEFINED BEHAVIOR (ie, other arguments may 
	do something today, but forwards compatability is not guaranteed)

*** News 03/26/2021 -- EPIC5-2.1.4 released here (Redound) (Commit id: 1927)

*** News 02/13/2021 -- /UNLOAD now supports /TIMERs
	If you create a /TIMER in a script you /LOAD with /PACKAGE
	the timer will be tagged just like any aliases or assigns
	or ons or whatever.

	Note that /TIMERs are _not_ tagged if they are created 
	after /load time.  This includes but is not limited to 
	timers created by aliases that are tagged; as well as
	timers that create new versions of themselves (but it 
	will catch timers that run forever)

*** News 02/10/2021 -- New /WINDOW operation, /WINDOW UNCLEAR
	There was already an /UNCLEAR command, and a /CLEAR and
	/WINDOW CLEAR command, but there was no /WINDOW UNCLEAR.
	WHy not? Anyways...

*** News 02/07/2021 -- Low level operations on cutbuffer - $inputctl()
	Harzilein asked if there was a way to manipulate the cutbuffer
	without having to put something in the input line.  As it happens,
	that has never been requested before.  So here we go!

	$inputctl(GET cutbuffer)
		This returns the cut buffer.  This is the same as $U

	$inputctl(SET cutbuffer ... new value ...)
		This sets the cut buffer.  This would be effectively
		the same as:
		 * Saving the input line (with $L)
		 * Erasing the input line (parsekey ERASE_LINE)
		 * /TYPEing what you want into the cut buffer
		 * Erasing the input line (to put it into the cut buffer)
		 * /TYPEing the origial input line
		But this doesn't require you to disrupt the input line.

		Please remember that there is only one cut buffer,
		and a large number of operations replace it -- so 
		whatever you put in the cut buffer, use it quickly,
		or the user might clobber it.

*** News 02/05/2021 -- /ON SERVER_STATUS changed to /ON SERVER_STATE
	Some expressed some confusion about whather "SERVER STATUS"
	and "SERVER STATE" were the same thing, and what the states
	were and what they meant.  To reduce this confusion, there
	will be only one term, "SERVER STATE". However, much code
	already uses "SERVER STATUS".  So we have to support both.

	For now, /ON SERVER_STATUS and /ON SERVER_STATE will both 
	exist side by side.   However, I recommend you use 
	/ON SERVER_STATE for new code, and think about migrating
	your old code.   There will come a day when /ON SERVER_STATUS
	will beoome an alias for /ON SERVER_STATE.

	To be unambiguous, using /ON SERVER_STATUS will never break.
	But you should know the official name is going to be 
	SERVER_STATE.

	The stock scripts have been updated.  You can get the changes
	by making sure to do a 'make install'

*** News 02/05/2021 -- $serverctl(GET x STATUS) changed to STATE
	As per the above, both $serverctl(GET x STATUS) and 
	$serverctl(GET x STATE) will return the server's state.
	However, you are encouraged to start using "STATE", since
	that is now the official term for it.

	The stock scripts have been updated.  You can get the changes
	by making sure to do a 'make install'

*** News 07/03/2020 -- New /WINDOW operation, /WINDOW DELETE_KILL
	Harzilein pointed out:
	  1) You cannot delete the last window on a screen
	  2) Deleting a screen does not kill a window
	  3) Therefore, deleting a screen necessarily orphans a window
	But what if you don't want to orphan a window when you kill a
	screeN?  

	The /WINDOW DELETE_KILL operation does a /WINDOW DELETE and
	then does a /WINDOW KILL on the window that was the current
	window.  There are several caveats to this *which i reserve
	the right to change in the future*
	
	Caveat 1) Only the "current window" gets killed.  So if you 
	    DELETE_KILL a screen with multiple windows, other windows
	    will be orphaned in the ordinary way.  I reserve the right
 	    to change this behavior (it will be documented)
	Caveat 2) If you cannot kill the window for other reasons, 
	    such as /window killable off, then the attempt to kill the
	    window will fail, and it will be orphaned.

*** News 05/11/2020 -- EPIC5-2.1.2 released here (Lugubrious) (Commit id: 1908)

*** News 05/09/2020 -- New built in function $execctl()
	You can now program the /EXEC command with this low level function

	--- warning ---
	Many of the things you can set here are not sanity checked.
	If you feed it garbage in, you will get garbage out.
	If you need safety rails, use the /EXEC command.
	--- warning ---

	Usage:
	------
	$execctl(REFNUMS)
		Get all refnums in the entire system (as integers)
	$execctl(REFNUM <description>)
		Convert a description (like %3 or %myproc) into a refnum integer

	$execctl(NEW <commands>)
		Create a new /exec process (does not run it!)
	$execctl(LAUNCH <refnum>)
		Run an /exec process that hasn't been started yet
	$execctl(CLOSEIN <refnum>)
		Close the STDIN to the process (this means you can't send any
		more data to the program.  This is useful for programs
		that wait until they've got an EOF from stdin to do their 
		thing)
	$execctl(CLOSEOUT <refnum>)
		Close the STDOUT and STDERR from the process (this means 
		we won't receive any more output from the program.  This is
		useful if we don't want any more output from the program.
		Many programs will die from SIGPIPE if stdout is closed)
	$execctl(SIGNAL <signal> <refnum>)
		KILL (send a signal) to a process.  The <signal> may either
		be an integer (like $execctl(SIGNAL 9 3) or it may
		be a short name (like $execctlSIGNAL HUP 3) to kill 
		process 3.

	$execctl(GET <refnum> [FIELD])
	$execctl(SET <refnum> [FIELD] [VALUE])
		You can GET all the fields of the object, and SET some of them.
		Some fields cannot be changed after the process is launched.

	Field		Set?	Notes
	-------------	------	-----------------------------------
	REFNUM		No	The integer refnum - never repeated.
	REFNUM_DESC	No	The "target" version (ie, %3)
	LOGICAL		Yes	The user-supplied name - must be unique
	LOGICAL_DESC	No	The "target" version (ie, %myproc)
				-- This will change if you SET LOGICAL
	COMMANDS	Yes	The commands to run (**)
	DIRECT		Yes	Whether to use a shell (**)
	REDIRECT	Yes	Either PRIVMSG or NOTICE
	WHO		Yes	The target to send output to
				-- Output from the process is redirected verbatim
	WINDOW_REFNUM	Yes	The window this exec runs in context of
				-- Undefined behavior if window does not exist
	SERVER_REFNUM	Yes	The server this exec runs in context of
				-- This is for redirects and code callbacks
	LINES_LIMIT	Yes	How many lines to receive before CLOSEOUT
				-- 0 means "no limit"
	STDOUTC		Yes	ircII code for every complete line of stdout
	STDOUTPC	Yes	ircII code for every partial line of stdout
	STDERRC		Yes	ircII code for every complete line of stderr
	STDERRPC	Yes	ircII code for every partial line of stderr

	PID		No	The process id of the process (after launch)
				-- This is -1 before launch
	STARTED_AT	No	When the process was created or launched
				-- It is first set when creation happens
				-- It is last set when launched
	P_STDIN		No	File descriptor to talk to process's STDIN
	P_STDOUT	No	File descriptor to read from process's STDOUT
	P_STDERR	No	File descriptor to read form process's STDERR
				-- There is nothing you can do with these.
				-- Don't try.
	LINES_RECVD	No	How many lines of output the process has sent
	LINES_SENT	No	How many lines we sent to the process
	EXITED		No	0 if process has exited yet -- 1 if it has
	TERMSIG		No	The signal that killed the process
				-- It is -1 if the process has not died
				-- It is -1 if the process exited normally
	RETCODE		No	The exit code of the process
				-- It is -1 if the process has not died
				-- It is -1 if the process was killed
	DUMB		No	Reserved for future/internal use
	DISOWNED	No	Reserved for future/internal use
	(Note: ** - Cannot be SET after launch)

*** News 05/08/2020 -- New flag to /EXEC, /EXEC -NOLAUNCH
	The /EXEC -NOLAUNCH flag directs the /EXEC command not to launch 
	a process that has not already launched.  You can launch it later 
	merely by referencing it with its name or refnum.

		/EXEC -nolaunch -name myproc ls -l
		/EXEC -nolaunch -limit 5 %myproc
		/EXEC %myproc

	Many /EXEC operations do not work on an unlaunched process.

*** News 05/07/2020 -- New function, $uuid4()
	This function returns a random UUID4 string.  If you know what that is,
	you might know why it's handy to be able to have one.

	ALL ARGUMENTS TO THIS FUNCTION ARE RESERVED FOR FUTURE EXPANSION.
	To get the default behavior, pass no arguments to this function.
	Forwards compatability is not guaranteed if you pass undefined args

*** News 05/07/2020 -- New /ON, /ON SEND_EXEC
	Whenever text is being sent to an /EXEC process, either through /EXEC -IN
	or /MSG %proc or a /QUERY or whatever, it is displayed to your screen.
	The /ON SEND_EXEC process will let you adorn how this text is displayed,
	instead of it just being displayed blankly as it has always done.

*** News 05/07/2020 -- The /EXEC command
	The /EXEC command has always been part of ircII, but it hasn't
	changed much during EPIC's lifetime.  There were some rough edges
	related to querying exec'd processes, things not always going
	to the windows people expected, some flags not being able to be
	used with other flags, etc.

	So the /EXEC command has been substantially revamped, with an intention
	of everything being "do what i expect".  If you find things that are 
	weirdly behaving, please let me know!

	Instead of describing the changes, let's just level-set the behavior:

	Every /EXEC command can create or modify one process

	[The most general explanation of the syntax]
	Modify an existing running exec:
		/EXEC [options] %refnum [extra arguments]
		/EXEC [options] %logical-name [extra arguments]

	Create a new exec:
		/EXEC [options] commands to run
		/EXEC [options] (commands to run) [extra-arguments]

	OPTIONS to /EXEC

	   [ Options related to sending data to/from the process ]
		-CLOSE
			Close the process's STDIN, STDOUT, and STDERR
		-CLOSEIN
			Close the process's STDIN
		-CLOSEOUT
			Close the process's STDOUT and STDERR
		-IN [extra arguments]
			Send data to a process

	   [ Options related to how the process integrates with ircII ]
		-NAME
			Change the logical name of a process
		-OUT
			Send all output from the process to the now-current 
			channel in the now-current window.
		-WINDOW
			Display all output from or related to the process to the 
			now-current window.
		-WINTARGET <windesc>
			Display all output from or related to the process to the 
			specified window
		-MSG [target]
			Send all output from the process to the target.
			The target can be an irc, dcc chat, or other exec,
			or anything you can send a message to.
			Messages sent over IRC are sent as PRIVMSGs
		-NOTICE
			Send all output from the process to the target.
			Messages sent over IRC are sent as NOTICEs.

	    [ Options related to scripting with processes ]
		-LINE {code}
			Run {code} for each full line of output from stdout of the process.
			$* will be the line of output
		-LINEPART {code}
			Run {code} for each incomplete line of output from stdout of the process.
			$* will be the incomplete line of output
		-ERROR {code}
			Run {code} for each full line of output from stderr of the process.
			$* will be the incomplete line of output
		-ERRORPART {code}
			Run {code} for each incomplete line of output from stderr of the process.
			$* will be the incomplete line of output
		-END {code}
			Run {code} when the process has completed.  This means when the process
			has exited and all of its output has been processed.  (Sometimes this 
			lags the actual exit)
				$0 is the logical process name or its refnum
				$1 is the signal that killed it (if it was killed)
				$2 is the exit code (if it exited)

	    [ Options that don't fit in the other categories ]
		-DIRECT
			Run the program directly -- do not use a shell.  
			Advantages:  The command you give is literally executed
			Disadvantages: If you depend on filename globbing or aliases, well, tough
		-LIMIT <number>
			Read <number> lines from the process and then -CLOSE it.

	    [ Options related to sending a signal to the process ]
		-<signal_number>
			Send the signal to the process, similar to kill -<signal_number> <pid>
		-<signal_name>
			Send the signal to the process, similar to kill -<signal_name> <pid>

	Additionally, %<procref> or %<procname> are full blown message targets that you can 
	/msg or /query or whatever you want, just like all other targets.

	[Examples of how this works]


*** News 04/26/2020 -- New python script, 'shortener'
	You can load this script with
		pyload shortener
	This script provides an in-client URL shortening service.
	Whenever someone provides a URL in a message, the service
	will create a short URL that will be served by an http
	redirection server that runs in client

	Example:
	--------
	<nick> hey, please visit www.frobnitz.com/really-long-and-wraps
        +onto-the-next-line-so-you-can-tpaste-it
	*** http://127.0.0.1:8080/1

	Then you can visit http://127.0.0.1:8080/1 and it will
	redirect you to the original url.

*** News 02/24/2019 -- EPIC5-2.1.1 (Abulia) was released here
	Even though not everything is done, I think I've probably
	dragged my feet long enough

*** News 11/28/2018 -- /EXEC -WINTARGET outsputs to a window by name (caf)
	The normal behavior of /EXEC is to send the output of a command
	to the current window (or is it the OTHER window?  I forget)
	Before this, you couldn't ordinarily send it to just any old
	random window you wanted.  

	You can use /EXEC -WINTARGET to send it to any window you want:
	Example:
		/exec -wintarget msgwin ls

*** News 02/24/2019 -- EPIC5-2.1.1 released here (Abulia) (Commit id: 1899)

*** News 02/05/2018 -- CTCP UTC now implemented as script
	Given the below feature, CTCP PING support has been 
	rewritten, and CTCP UTC is now scripted.

*** News 02/13/2018 -- New flag for $ctcpctl(), "REPLACE_ARGS"
	There are actually two kinds of CTCPs that replace things

	* CTCP PING replaces its argument(s), but is otherwise
	  handled "normally"
		NOTICE nick :\001PING <sec> <usec>\001
	  becomes
		NOTICE nick :\001PING <sec> seconds\001

	* CTCP UTC replaces itself entirely:
		PRIVMSG nick :\001UTC 1518582810\001
	  becomes
		PRIVMSG nick :Tue Feb 13 22:33:30 2018

	So it's not enough to say that "a CTCP handler can replace 
	itself by returning a string", you need to be able to say
	whether this CTCP replaces its arguments only, or whether
	it replaces itself entirely.

	  * $ctcpctl(SET <ctcp-name> REPLACE_ARGS 1)
	  * $ctcpctl(SET <ctcp-name> REPLACE_ARGS 0)
		Select whether or not a CTCP handler that returns a 
		string replaces its arguments (like CTCP PING) or 
		replaces itself entirely (like CTCP UTC).  
		The default is 0 (replace entirely)

*** News 02/05/2018 -- Some CTCPs are now implemented as scripts
	YOU NEED TO START DOING /LOAD ctcp  IN YOUR STARTUP SCRIPTS.

	One of the larger projects in EPIC5 was to move as many
	hard coded things into scripts as was feasible, so you,
	the user (or the script you're using) can have as complete
	control over them.  We've moved a lot of functionality out
	into scripts.

	Traditionally those users who don't have startup scripts 
	(~/.epicrc or ~/.ircrc) get /load global as their startup
	script.  One of the things /load global does is /load builtins
	which brings in the scripted features.  

	Now /load builtins will /load ctcp, which implements these 
	core CTCP functions entirely in ircII, so you are welcome 
	to modify or remove them, as _you_ choose.

		VERSION		PING		ECHO
		CLIENTINFO	USERINFO	ERRMSG
		FINGER		TIME

	Maybe more will be migrated in the future.  But this is a good
	start for now.  This is also a great example of how to 
	build your own first-class ctcp handlers!

*** News 02/04/2018 -- User-defined CTCP handlers with $ctcpctl()
	You can now create your own first-class user-defined CTCP handlers.

	A CTCP handler is a block of code that takes four arguments:
	  $0  - The person making the request
	  $1  - The target of the request (you, or a channel you're on)
	  $2  - The CTCP that was sent 
	  $3- - Arguments (if any) to the CTCP

	A CTCP request handler is a CTCP handler that handles incoming 	
	requests (over PRIVMSG).  A CTCP request handler can either 
	   (1) generate a response or 
	   (2) substitute something.  
	A response can be generated with:
	      /CTCP $0 $2 Your Response Here
	A substitute string is /return'ed by your handler, and replaces
	the CTCP in the original message.

	A CTCP response handler is a CTCP handler that handles incoming 
	responses (over NOTICE).  A CTCP request handler can either
	   (1) Output the response in a special way or
	   (2) substitute something
	CTCP response handlers are unusual, because the ordinary handling
	of CTCP responses is the expected behavior.

	Syntax:
	  * $ctcpctl(SET <ctcp-name> REQUEST {<code>})
		Register <code> as a CTCP handler for <ctcp-name> requests.

	  * $ctcpctl(SET <ctcp-name> RESPONSE {<code>})
		Register <code> as a CTCP handler for <ctcp-name> responses.
		(Note -- handling responses is unusual.  Normally you just
		 let the client output responses in the ordinary way)

	  * $ctcpctl(SET <ctcp-name> DESCRIPTION <string>)
		SET <string> as the CLIENTINFO for <ctcp-name>.  That is,
		when someone does /CTCP CLIENTINFO <ctcp-name>, <string> 
		will be returned as the description for this CTCP.

	  * $ctcpctl(SET <ctcp-name> SPECIAL 1)
	  * $ctcpctl(SET <ctcp-name> SPECIAL 0)
		Enable/Disable a CTCP as being "Special".  A "Special" CTCP
		is a remote function call, and handles everything itself.
		There are only two "special" CTCPs -- ACTION (/me) and DCC.
		I'm not sure if anyone will create a "special" user-defined CTCP

	   * $ctcpctl(SET <ctcp-name> RAW 1)
	   * $ctcpctl(SET <ctcp-name> RAW 0)
		Enable/Disable a CTCP has requiring the "raw data".
		Ordinary CTCPs transport strings, and they have to be recoded
		according to /encode-ing rules.  But some CTCPs transport 
		binary data, and so the handler needs access to the raw binary
		data.  Ordinarily, the raw/binary data is CTCP encoded, which
		mens you can pass it to $xform("-CTCP") to recover the raw
		bytes (although it might not be a C string, so you can't 
		assign it to a variable.)

	   There are corresponding GET operations for the above

	You can get all the registered CTCPs with
	   * $ctcpctl(ALL)

	Very soon, quite a few CTCP types will be migrated out to a script that
	will be /load'ed from /load global, and you may have to add it to your
	own start scripts if you do not /load global.

	I need to write much better examples for all this.  To look at this you'd
	scratch your head and wonder why you care.  But being able to add new 
	CTCPs instead of requiring them to be written in C in a new version of
	epic is expected to help a lot of people.


*** News 01/16/2018 -- New status expando %{1}P ("status prefix") and variables
	The %{1}P value will expand to a "when window current" or "when window
	not current" value.  The idea is to put this at the start of your 
	/set status_format or /window status_format type variables.

	When a window is current, %{1}P will expand to either
		/window status_prefix_when_current
	or
		/set status_prefix_when_current

	When a window is not current, %{1}P will expand to either
		/window status_prefix_when_not_current
	or
		/set status_prefix_when_not_current

	You can use this all in your ~/.epicrc, like so:
		set status_format %{1}P%T [%R] %*%=%@%N%#%S%{1}H%H%B%Q%A%C%+%I%O%M%F%L %D %U %W
		set status_prefix_when_current ^C37,40
		set status_prefix_when_not_current ^C37,44
	which will make your status bar white-on-black when current,
	and white-on-blue when not current.


EPIC5-2.0
EPIC5-1.8

*** News 08/05/2016 -- EPIC5-2.0.1 released here (Commit id: 1869)

*** News 01/30/2016 -- EPIC5-2.0 released here (Commit id: 1864)

*** News 01/30/2016 -- EPIC5-1.8 released here (Commit id: 1862)

*** News 01/30/2016 -- /WINDOW LOGFILE and /SET LOGFILE more like /LOG FILNAME
	Historically, changing a logfile name (with /WINDOW LOGFILE and
	/SET LOGFILE) does not affect the log status.  This leads to 
	unexpected behavior if you do /WINDOW LOG ON LOGFILE foo.txt
	because /WINDOW LOGFILE only changes the filename the *next* time
	you open the log, not affecting the currently open log.

	The behavior of /LOG FILENAME is more in line with what people 
	said they expected.  If you change /LOG FILENAME while the 
	log is ON, then it will 1) close the existing log, 2) change
	the filename, and 3) re-open the log under the new name.

	The behavior of /WINDOW LOGFILE and /SET LOGFILE have been changed
	to match the behavior of /LOG FILENAME -- changing the logfile name
	while it is open will close the existing log and open a new one.

EPIC5-1.6

*** News 01/30/2016 -- EPIC5-1.6 released here (Commit id: 1854)

*** News 01/08/2016 -- Per-server vhosts now restrict protocol (ipv4/ipv6)
	Historically, the client tries to connect to the server using the
	addresses as they are returned in order.  (This is a great thing 
	for round-robin or geographic-aware dns resolvers.)

	However, if you have a per-server vhost, you probably intend that
	epic use that vhost to connect to the server.  But what happens if
	your vhost is ipv4 only or ipv6 only, and the first address to the
	server is to the other protocol?  Historically, epic will just go 
	ahead and connect without your vhost.

	You've been able to correct this behavior by specifying explicitly
	the protocol family:
		/server irc.foo.com:proto=v6:vhost=irc.leet6.com
	Some folks said it violated POLA, so here's a new rule:

	"If you set a per-server vhost, then that server can only be 
	 connected to if the vhost can be used.  If ths means that no
	 addresses can be used, then you will not be able to connect 
	 to the server until you clear the vhost."

	This rule does not apply to you if you're using /HOSTNAME but
	only if you are doing something like 
		/server irc.foo.com:vhost=irc.leet6.com

*** News 01/01/2016 -- Can now /query an exec process that doesn't exist
	Previously, you were forbidden from setting up a /query to an 
	/EXEC process that didn't exist.  That set up a race condition
	between running an /EXEC process and being able to corral its 
	output into a window via the query.  

	So now you can /query an exec process before you start it.  

	If you try to send a message to it before you do fire it up, 
	the user will see a diagnostic telling them that the message 
	could not be sent to a non-existing exec process.

*** News 01/01/2016 -- New /window operation, /window log_mangle
	This allows you to overrule /SET MANGLE_LOGFILES for logs that
	you create with /WINDOW LOG ON (only!)
	Example:
		window logfile "my.windowlog" mangle NORMALIZE,COLOR log on

*** News 01/01/2016 -- New /window operation, /window log_rewrite
	This allows you to overrule /SET LOG_REWRITE for logs that
	you create with /WINDOW LOG ON (only!)
	Example:
		window logfile "my.windowlog" rewrite "HOOBOO $1-" log on

*** News 01/01/2016 -- Refinement to $pad(), $[len]var, and $leftpc()
	These functions do not behave graciously since our conversion
	to UTF-8, since they count code points rather than columns.
	It just seems sensible to redefine the behavior of these 
	functions based on columns, which is what everybody probably
	expects them to do.

	Function: $pad(len char string)
	Summary: Extend, but do not truncate, do not justify 'string'
	Definition: 
		Return 'string' so that it takes up at least 'len' columns.
		If it is too short, it will be padded with 'char's until
			it is 'len' columns wide.
		If it is too long, it is NOT truncated.

	Function: $leftpc(len string)
	Summary: Truncate, but do not extend, do not justify 'string'
	Definition:
		Return the first 'len' columns of 'string'.  
		If it is too short, it will NOT be padded.
		If it is too long, it will be truncated on the right end.

	Function: $[len]var
	Summary: Extend, truncate, and justify $var
	Definition:
		Return $var so that it takes up EXACTLY 'len' columns.
		If it is too short, it will be padded with /set pad_char
			until it is 'len' columns wide.
		If it is too long, it is truncated.
		If 'len' is > 0, then the string is left justified, and
			padding (or truncation) happens on the right end.
		If 'len' is < 0, then the string is right justified, and
			padding (or truncation) happens on the left end.

	Function: $fix_width(len justify pad string)
	Summary: Extend, truncate, and justify 'string'
	Definition:
		Return 'string' so it takes up EXACTLY 'len' columns.
		If it is too short, it will be padded with 'pad'
			until it is 'len' columns wide.
		If it is too long, it is truncated.
		If 'justify' is "l" then the string is left justified,
			and padding (or truncation) happens at the right
		If 'justify' is "c" then the string is centered, and 
			padding (or truncation) happens equally at both
			ends.
		If 'justify' is "r" then the string is right justified,
			and padding (or truncation) happens at the left.

*** News 09/15/2015 -- New /ON, /ON RAW_IRC_BYTES
	This new /ON, /ON RAW_IRC_BYTES is the same as /ON RAW_IRC, 
	except $* is the _raw unmodified bytes_ received from IRC.
	Specifically, $* is not guaranteed to be a UTF-8 string, so 
	functions that expect a UTF-8 string won't work.  You should
	not try to /ECHO the $* from this /ON.

	Just like /on raw_irc, if you catch this hook, you will 
	suppress normal handling of the event:
		/on ^raw_irc_bytes * {echo nothing further happens}

EPIC5-1.4

*** News 08/29/2015 -- EPIC5-1.4 released here (Commit id: 1833)

*** News 08/25/2015 -- Improved automargin support 
	You can now use automargins to get better wrapping of long urls.

	1. Use a terminal emulator that supports automargins
	   (they pretty much all do)
	2. Set your TERM env variable to something that supports automargins
		export TERM=vt102am        
	   should do the job
	3. Restart EPIC after you've updated the TERM.  It's not enough 
	   to just change the TERM and re-attach screen.  It's a good 
	   idea to check this when you're upgrading.
	4. /SET -CONTINUED_LINE	
		to get rid of the + thingee
	5. /SET FIRST_LINE >   
		or something if you don't use /set output_rewrite 
		to prefix every line
	6. /SET WORD_BREAK<space><space><enter>
		URLs contain commas and dots and semicolons, and you 
		don't want epic to word break on anything other than 
		a space.

	That should do it!  Your display will now use the final column
	on the display, and urls should be unmangled when you copy them

	If you "forget" to /SET -CONTINUED_LINE or /SET WORD_BREAK and want 
	to rebreak your windows in order to take advantage of this after 
	the fact, you can always just do /window rebreak_scrollback.  

	*** IMPORTANT! ***
	Note that you _must_ be running with a TERM supporting 
	automargins or this will not change things!

	You can check this by doing /eval echo $getcap(TERM enter_am_mode)
	If it returns blank, then your TERM does not support automargins.
	*** IMPORTANT! ***

*** News 08/25/2015 -- New /SET, /SET FIRST_LINE
	For those of you who use /SET OUTPUT_REWRITE to prefix every line
	of output with something (like a timestamp), you can ignore this.

	If you /SET FIRST_LINE, the string will be prefixed before every
	logical line of output.  This is great if you /SET -CONTINUED_LINE
	so you can continue to tell what lines are what.

EPIC5-1.2 
EPIC5-1.1.11

*** News 08/09/2015 -- EPIC5-1.2 released here (Commit id: 1817)

*** News 07/20/2015 -- New operation: @serverctl(SET refnum UMODE ...)
	You've never been able to SET your UMODE, becuase that never
	made much sense.  But it was pointed out that the UMODE is used
	when you reconnect to establish your initial usermode, and they
	wanted to be able to control that.

	So now, when the server is disconnected, you can change its umode,
	which will be used when you next connect.  Try something like this:
		on server_lost * {defer @serverctl(SET $0 UMODE i)}

*** News 07/11/2015 -- EPIC5-1.1.11 released here (Commit id: 1810)

*** News 07/09/2015 -- Don't do /redirect @W<refnum> <command>
	You can /msg @W4 <stuff> for example to do an /xecho -w 4 <stuff>.
	But you can't use this with /REDIRECT because the redirection 
	itself causes another redirect, and it just gets stuck in an 
	infinite loop.  The client can't protect you from doing this
	without a rewrite, so, don't do this. ;-)

*** News 07/09/2015 -- New flag to /ENCRYPT: /ENCRYPT -REMOVE
	It was confusing to remove /encrypt's before -- you had to specify
	all the arguments but not the password:
		/ENCRYPT hop -blowfish
	And if you had a PROG crypto, you couldn't remove it at all!
		/ENCRYPT hop password program...
	(How do you not specify the password in this case?)

	Anyways, there is now /ENCRYPT -REMOVE which lets you unambiguously
	remove an encrypt for a target
		/ENCRYPT -REMOVE hop

*** News 07/09/2015 -- Only one /ENCRYPT per target now.
	Traditionally, ircII has only had one cipher type for /ENCRYPT.
	EPIC added more cipher types along the way, and it was possible
	to set up multiple /ENCRYPT sessions for the same person.
	However, as I reflect upon this, this isn't a reasonable thing to
	do, because if you do
		/ENCRYPT hop -blowfish password1
		/ENCRYPT hop -aessha password2
	And then you /msg hop, which one should it use?

	So I've changed it so when you change the encryption type (such as
	in the 2nd line above), it will REPLACE the first one -- you will 
	only be able to have one cipher session per target.

	If this is a problem -- if I broke something for you, please let me
	know so we can address your needs.  

*** News 04/14/2015 -- New function, $chankey(servref #channel)
	The $chankey() function returns the channel for a specified channel.
	I created this because $key() doesn't allow you to specify
	'servref' but rather uses from_server, which means you have to
	wrap it in an /xeval -s to use a non-default server. ick.

	Arguments:
	 $0	- A server refnum
	 $1	- A channel name 

	Return Value:
	 empty string - Either 1.) 'servref' not provided, or 
			       2.) '#channel' not provided, or
			       3.) You're not on '#channel' on servref, or
			       4.) #channel doesn't have a key
	 anything else - The mode +k key for #channel on servref.

*** News 04/14/2015 -- SSL info available via $serverctl(GET refnum SSL*)
	You can now get information about live SSL connections:

	$serverctl(GET refnum SSL_CIPHER)
		The encryption being used, something like "DHE-RSA-AES256-SHA"

	$serverctl(GET refnum SSL_VERIFY_RESULT)
		0 if the certificate was verified successfully.
		Any other value if it did not.  
		These values match up to the verify(1) man page.
		Most notable, error code 20 means that OpenSSL could not find
		your local CA file (see /SET SSL_ROOT_CERTS_LOCATION below)

	$serverctl(GET refnum SSL_PEM)
		This is the PEM (base64) format of the server's certificate.
		You could save this to see if it's changed. ;-)

	$serverctl(GET refnum SSL_CERT_HASH)
		This is the SHA1 digest of the server's certificate.
		It is converted into a byte string like AB:CD:EF:01:02:...

	$serverctl(GET refnum SSL_PKEY_BITS)
		This is the number of bits that the server's certificate
		said that the server's public key uses.

	$serverctl(GET refnum SSL_SUBJECT)
		This is the hostname of the subject of the certificate.
		In theory, this is supposed to be the server's hostname.
		This could be a wildcard string.

	$serverctl(GET refnum SSL_SUBJECT_URL)
		This is the SSL_SUBJECT, but passed through URL encoding.
		It's useful because the SSL_SUBJECT will have spaces, and
		this results in just one word.  

	$serverctl(GET refnum SSL_ISSUER)
		This is the Certificate Authority (CA) that issued the 
		server's certificate.

	$serverctl(GET refnum SSL_ISSUER_URL)
		This is the SSL_ISSUER, but passed through URL encoding.
		It's useful because the SSL_ISSUER will have spaces, and
		this results in just one word.

	$serverctl(GET refnum SSL_VERSION)
		This is the version of SSL you're using.  
		It should either be TLSv1 (good) or SSLv3 (bad).

	With all of the above information, I hope someone scripts a nice
	SSL executive script that caches the certificate information, 
	tells you whether the connection should be trusted, decides whether
	the ssl version is ok, the public key bits are ok, all that stuff.

*** News 04/10/2015 -- New /TIMER argument, /TIMER -SNAP
	A "snappable" timer fires off at the "top of the interval".
	That means it runs every time ($time() % <interval> == 0)
	If you snap to 60, it will run at the top of every minute,
	just like mail checking, clock updating, etc.  If you snap 
	to 3600, it will run at the top of every hour.

	Example:
		/timer -snap -repeat -1 -refnum hourly 3600 {
			echo I run at the top of every hour
		}

*** News 04/10/2015 -- New script, /LOAD find_ssl_root_certs
	This script is loaded by /load global, and you should load it 
	too, if you're not using /load global; or you should implement
	a similar functionality to get /SET SSL_ROOT_CERTS_LOCATION pointing
	to the right place for you.  Otherwise, your SSL certificates
	won't authenticate, and $4 in /on ssl_server_cert will always 
	be 0, even if the cert is actually legitimate.  Or maybe all
	of what I just said doesn't matter to you.

*** News 04/10/2015 -- New /SET, /SET SSL_ROOT_CERTS_LOCATION
	In order for SSL to verify certificates, it needs to have a copy
	of the root certificate authorities.  This is usually a file
	named "ca.bundle" or "ca-root-nss.crt".  You need to /set this
	variable to wherever your openssl compatable root ca authority
	certificates are.  It would help if I understood what I'm talking
	about more.  Anyways, the script /load find_ssl_root_certs
	tries to help you with this.

*** News 04/10/2015 -- Enhancements to /ON SSL_SERVER_CERT
	Apparently I've never documented /on SSL_SERVER_CERT.
	
	The /ON SSL_SERVER_CERT hook is thrown every time a successful
	connection to an SSL server is made.  For now, the only SSL 
	connections are to IRC servers, but some day I hope to support
	DCC as well.  For now, this refers only to server connections.

	  $0 - File Descriptor (need a way to convert to server refnum)
	  $1 - The "subject" of the certificate  -- the server name
	  $2 - The "issuer" of the certificate
	  $3 - How many bits are used by the certificate's public key
	  $4 - Did the certificate validate?
		0 = pass, anything else = fail.
		(For now, this the result of X509_get_verify_result())
		(This depends on /SET SSL_ROOT_CERTS_LOCATION (above))
	  $5 - What SSL type are we using? (TLSv1 or SSLv3)
	  $6 - What is the digest of the SSL Certificate?

	The idea is you could use $1 to cache the metadata about a 
	server, and use $2, $3, $4, $5, and $6 to see if anything
	changes from one connection to the next.

	Probably more additions will come later.  I'm especially interested
	in passing in the complete plain text certificate +url'd up.

*** News 04/10/2015 -- More robust certificate verification for SSL
	Based on a paper written by Roca He, who is doing a research
	project on improper use of OpenSSL API by open source software,
	the OpenSSL code in epic was reviewed and enhanced.  One point
	of interest was certificate verification -- epic wasn't doing
	any of that.  But now it is.  This requires OpenSSL to know
	where your root/trusted certificate authorities are, and there
	are more notes above about how to handle that.  The results of
	this verification are reflected in /on ssl_server_cert above.

*** News 04/10/2015 -- You can now /encode to /EXEC processes 
	The /EXEC system is now UTF-8 aware, and you can use the 
	/encoding command to recode between %procs targets now. yay!

	Example:
		/encoding %nonutf8prog iso-8859-15
		/exec -name nonutf8prog myprog
		<output from 'myprog' treated as iso-8859-15, 
		 converted to utf8 for epic's use>

*** News 04/10/2015 -- New scripts: /load sasl_auth, userlist, tmux_away
	Zlonix wrote these scripts.  I need to write blurbs about each.
	Until that time, read the scripts, they're well documented!

*** News 07/24/2014 -- New feature $windowctl(REFNUMS_ON_SCREEN <winref>)
	The $windowctl(REFNUMS_ON_SCREEN <winref>) will return all of the
	refnums on the screen that contains window <winref>.  They are
	returned in SCREEN ORDER, ie, from top to bottom.  

	This was because someone wanted to make the bottom window ONLY 
	double status bar, so he needed to know which one was on bottom 
	and which ones weren't.

	If <winref> is a hidden window, then it will return all of the 
	invisible windows, but in no guaranteed order.

	If <winref> is 0, then it will return the screen for the current
	window, of course.

	You MUST specify a window refnum of some sort, even if it's just 0.

*** News 04/19/2014 -- Support for tmux for /WINDOW CREATE
	If you run EPIC under tmux, you can now use /window create and it
	will create new screens running under other tmux screens, just like
	it does for gnu screen.  There is also /set tmux_options, but I did
	not really test that.  What EPIC does is run this command:

	  tmux new-window "<wserv_path> <tmux_options> localhost [port]"

*** News 04/17/2014 -- /SET TRANSLATION now retired
	The /SET TRANSLATION feature which has served us well for many
	years is now superceded by /ENCODING, and has been retired.

EPIC5-1.1.10
EPIC5-1.1.9

*** News 04/13/2014 -- EPIC5-1.1.10 released here (Commit id: 1781)

*** News 04/12/2014 -- EPIC5-1.1.9 released here (Commit id: 1780)

*** News 04/16/2014 -- $fix_width() now fully works, and UTF-8 aware.
	The $fix_width() function takes the following arguments:
		cols      $0    Number of display columns
		justify   $1    Justification ("l" left, "c" center, "r" right)
		fillchar  $2    Fill character (can be utf8 cchar)
		text      $3-   
	This fully supports UTF-8, so the result is a string that will
	take up "cols" columns, even if "fillchar" takes multiple columns.


*** News 04/16/2014 -- Many functions are now UTF8 aware
	These functions are now fully UTF8 aware.  This means that 
	their unit of operation is a unicode character, and not a byte.

		after		before		center		chop		
		chrq		curpos		fix_width	index		
		indextoword	insert		left		maxlen
		mid		msar		pad		pass
		rest		reverse		right		rindex
		rsubstr		sar		split		strip
		strlen		substr		toupper		tolower
		tr		wordtoindex	

	These things are also UTF-8 aware:
		/XTYPE -L
		$[num]VAR		(pad/truncate $var to num places)
		/FEC

	Additionally, case sensitivity is UTF8 aware, for all languages,
	not just English. (at least as far as I tested)


*** News 04/11/2014 -- New function, $encodingctl()
	The $encodingctl() gives you a lower level interface to the 
	encoding system.  

	The behavior of $encodingctl() is ugly and I regret several of
	the decisions I've made already, but that's the way it goes...

	As with other $*ctl() functions, if there was not information in
	the argument list to decide what you wanted to do, it will return
	the empty string.

	- $encodingctl(REFNUMS)
	Return all recode rule refnums.

	- $encodingctl(MATCH servref sender receiver)
	Decide which recode rule would be used for a message sent by
		"sender" to 'receiver' over the server 'servref'.
	If you are the sender, use $servernick().
	Return value:
	 empty_string - an argument is missing, or servref is not an integer
	 anything else - the rule that would be used.

	- $encodingctl(GET refnum <OPERATION>)
	Get an attribute of a recode rule:
	  -> Returns empty string if <refnum> is not a valid recode rule
	  -> Returns empty string if no <operation> specified.

	  - $encodingctl(GET refnum TARGET)
	  Return the complete "target" part of the rule; this is whatever
	  you passed to the /encoding command.

	  - $encodingctl(GET refnum ENCODING)
	  Return the complete "encoding" part of the rule; this is whatever
	  you passed to the /encoding command.

	  - $encodingctl(GET refnum SERVER_PART)
	  Return the 'server part' of the target.  This is the part before
	  the slash, or the empty string if there is no server part.

	  - $encodingctl(GET refnum TARGET_PART)
	  Return the 'target part' of the target.  This is the part after
	  the slash, or the empty string if there is no target part.

	  - $encodingctl(GET refnum SERVER_PART_DESC)
	  This returns nothing for now.  Maybe someday it will return a
	  server description (ie, host:port:...)

	  - $encodingctl(GET refnum MAGIC)
	   1 if this rule is a "system rule" and cannot be deleted.
	   0 if this is a user rule, and can be deleted.

	  - $encodingctl(GET refnum SOURCE)
	   1 - Set at boot-up by using your locale (CODESET)
	   2 - Set at boot-up from hardcoded defaults
	   3 - Set by the user

	- $encodingctl(SET refnum <OPERATION>)
	Change an attribute of a recode rule
	  -> Returns empty string if <refnum> is not a valid recode rule.
	  -> Returns empty string if no <operation> specified.

	  - $encodingctl(SET refnum ENCODING new-encoding)
	    Empty String - New-encoding was not specified
	    0 - Changed: The new encoding was set successfully
	   -1 - Not changed: The new encoding does not exist on your system
	   -2 - Not changed: The new encoding does not convert to UTF-8
	   -3 - Changed: The new encoding PARTIALLY converts to UTF-8

	- $encodingctl(DELETE refnum)
	Delete a recoding rule:
	 Empty String - <refnum> was not a valid recode rule.
	 0 - <refnum> is a magic rule and may not be deleted
	 1 - <refnum> was successfully deleted.

	- $encodingctl(CHECK encoding)
	Determine whether or not <encoding> could be used in recoding rules:
	  Empty String - <encoding> was not specified.
	  0 - Acceptable: The encoding is acceptable for use
	 -1 - Unacceptable: The encoding does not exist on your system
	 -2 - Unacceptable: The encoding does not convert to UTF-8
	 -3 - Partially Acceptable: The encoding PARTIALLY converts to UTF-8

	- $encodingctl(CREATE target encoding)
	Basically the same thing as /encoding target encoding
	Returns the refnum of the new rule.

EPIC5-1.1.8

*** News 03/11/2014 -- New command, /ENCODING
	The /ENCODING command is not really new, but it finally is in its
	final form.

	With /ENCODING you can specify rules that tell epic what encoding
	you think other people are using.  Whenever epic receives a non-utf8
	message, it will evaluate the rules to decide what encoding it should
	treat the non-utf8 message.  Whenever you send an outbound message
	to irc, epic will use the rules to decide if it should encode it 
	in something other than utf-8.

	This means you can talk to non-utf8 users, and their messages can
	be made utf8 for you; and your (utf8) messages can be non-utf8 for
	them.  This is full end-to-end recoding support.

	The rules look like this, and are evaluated with this priority:
		6. /ENCODING server/nickname
		5. /ENCODING nickname
		4. /ENCODING server/channel
		3. /ENCODING channel
		2. /ENCODING server
		1. /ENCODING irc		(the "magic" rule)

	In this case, "server" is anything that can be recognized:
		* A server refnum
		* A server "ourname"
		* A server "itsname"
		* A server group
		* Any server altname

	The client will evaluate each rule, and the "best match" is the
	first rule that lands highest on that first list.  Hopefully it
	should just be natural.  An example:

		# All "efnet" servers use ISO-8859-1	(level 2)
		/encoding efnet/ ISO-8859-1

		# Except #epic, which uses CP437	(level 3)
		/encoding #epic CP437

		# Except zlonix, who uses KOI8-R	(level 5)
		/encoding zlonix KOI8-R

	So if you get something non-utf8 message over an "efnet" server,
	it will be assumed to be ISO-8859-1.  Unless that message was 
	sent to #epic -- then it is assumed to be CP437.  Unless that 
	message was sent by zlonix, then it is KOI8-R.  In this way, you
	can set defaults for channels and overrule it by individual person.

	SYNTAX:
		If you do	/ENCODING <stuff>

	If <stuff> is a channel, then <stuff> is treated as a channel.
	If <stuff> is a number, then <stuff> is treated as a server refnum.
	If <stuff> contains a slash, then anything before the slash is the
		server part, and anything after the slash is the channel
		or nickname part.  *You can use a trailing slash to make
		it unambiguous you mean a server, or a leading slash to
		make it unambiguous you mean a channel/nickname.*
	If <stuff> contains a dot, then <stuff> is treated as a server.
	Otherwise, anything else is treated as a nickname.

	
*** News 03/06/2014 -- $cparse(%X) turns into a ^X
	You can use %X in $cparse() to inject a ^X which might make it 
	easier to handle 256 color support.  All the rules below still apply.

		/echo $cparse(%kone %rtwo %X80buckle %XFFmy %X32shoe)

*** News 03/05/2014 -- 256 color support (^X works like ^C)
	The ^X attribute allows you to set 256 colors (if your emulator
	supports that).

	The ^X attribute takes two hex digits to indicate a color between
	00 and FF (0 to 255).

		^X-1			Turn off color
		^X00			Turn on fg color 0
		....			....
		^XFF			Turn on fg color 255
		^X<number>,<number>	Turn on <fg>,<bg> colors
					The <fg> number can be omitted.

	The ^X attribute ALWAYS takes stuff after it.  You cannot use a naked
	^X or you risk forwards incompatability.  I *will* be adding more 
	stuff to ^X in the future so don't develop bad habits.

	Caveat:
	  256 color support isn't "standard" ansi so the client sends the 
	  hardcoded sequence ^[[38;5;<number>m  to your terminal.  If your 
	  terminal does not honor this way of doing 256 colors, then there's 
	  not much you can do about it...

*** News 03/05/2014 -- Italics support
	The highlight character ^P now toggles the "Italic" setting of
	your terminal emulator.  Mine doesn't support this so I couldn't
	test it very well.  Please report any bogons if you use it.

*** News 03/02/2014 -- Clarified behavior for /set lastlog 0
	It was pointed out that /set lastlog 0 did not do a reasonable
	thing with the unified scrollback buffer, so the behavior has
	been refined a bit.  

	Here is how /set lastlog <X> now works
	1. For each window, set /window lastlog <X>.
	2. For each window, rebuild each window's scrollback 
	   -- Which may throw away stuff!  
	   -- If you do /set lastlog 0, it throws away all scrollback
	      and does an implicit /window clear!
	3. For each window, if <X> is less than twice the window's size,
	   /window lastlog <twice its size> (24 lines -> /window lastlog 48)
	4. The final value of <X> will be twice the size of the biggest window.

*** News 02/14/2014 -- New flags, /LASTLOG -THIS_SERVER and -GLOBAL
	The /LASTLOG -THIS_SERVER flag will show all lastlog entries
	from any window belonging to the server server as this window.
	IE, it is a /lastlog that catch all of this server.

	The /LASTLOG -GLOBAL flag will show all lastlog entries from
	any window whatsoever.

*** News 02/11/2014 -- Auto-detect incorrect encodings lead to warnings
	If you have /ENCODING CONSOLE set to non-utf8, and then you type
	stuff that looks like UTF8, the client will tell you and suggest
	you switch.  This will end the problem with utf8 users seeing
	multiple garbage characters on the input line.

	If you have /ENCODING CONSOLE set to utf8 and you type something
	that is not utf8, the client will tell you and suggest you switch
	to something else.  Unfortunately it's not easy to know what you 
	are using, so it suggests ISO-8859-1.  

	If you are correctly setting LC_ALL (see below) then the above
	should never happen for you.

*** News 02/11/2014 -- Now honoring LC_ALL (locale charset settings)
	If you set your character set via locale environment variables,
	EPIC will now use your locale as the default character set for
	/ENCODING console.  If you do not set your locale variables,
	then epic will continue to default to ISO-8859-1.

*** News 02/11/2014 -- New /ENCODING target "scripts"
	Whenever you /load a script, epic needs to convert it into utf8.
	The normal way a script can declare itself is via /load -encoding
	(See the note from 11/17/2012)

        if (word(2 $loadinfo()) != [pf]) { load -pf -encoding CP437 $word(1 $loadinfo()); return; };

	If a script is not well-formed utf8 and it does not declare its
	own encoding, then it will be assumed to be whatever the value
	of /ENCODING scripts is.  The hardcoded default is "CP437".

	Naturally, if you /load a script that is not utf8 and is not
	CP437, it may not translate correctly.  But it seems most scripts
	use CP437, and we'll get everybody to declare/utf8-ify their scripts.

*** News 02/10/2014 -- EPIC users now utf8 stransparant to irc.
	As of right now, whether you are using utf8 or not, anything
	you send from epic will be sent to irc as UTF8.  Anything UTF8
	that anybody sends you will be displayed properly on your screen,
	even if you are not using UTF8.

*** News 02/10/2014 -- New command /ENCODING -- declare target encodings
	The new command /ENCODING is used to declare what string encoding
	a target is using.  At some point this will spawn into an all-
	encompassing feature, but for now, it's just used to declare the
	encoding of your console.

		/ENCODING console ISO-8859-1
	or	/ENCODING console UTF-8		(default)

	If EPIC detects that your input is illegal for the encoding you
	are using, it will ask you to change it.  If the stuff you type
	is not what you think you're typing, again, you might be using
	the wrong encoding.

*** News 02/09/2014 -- Unicode support for $chr(), new func $unicode()
	The $chr() function will now accept unicode descriptors
		$chr(U+0415 U+0420 U+20)
	The $unicode() function converts text to unicode descriptors
		$unicode() returns "U+0423"

*** News 02/08/2014 -- New server flag, "encoding"  -- WITHDRAWN
	Server descriptions now have an extra field "encoding" which is 
	used when you receive a non-utf8 string from the server.

	When you receive a non-utf8 string from the server, epic will 
	assume it is in this encoding and use iconv() to convert it to utf8.

	The default is "ISO-8859-1" for no particular reason.
	This will be supplanted by a /recode command in the future!

	*** This feature has been removed.  Do not use! ***

EPIC5-1.1.7

*** News 01/16/2014 -- New scripts (should document these!) from Zlonix
	   xmsglog	Encrypted logfiles
	   sasl_auth	SASL support (for some networks)
	   idlealert	Monitor friends' idle times (with lots of WHOISs)

*** News 01/02/2014 -- New status expando, %{4}S, full "itsname"
	Just to run this down, here are the %S expandos
		%S	Altname 0 (default: Shortened "ourname")
			(Only shown when you're connected to multiple servers)
		%{1}S	Altname 0 (default: Shortened "ourname")
		%{2}S	Full "ourname"
		%{3}S	Server Group
		%{4}S	Full Server "itsname"

*** News 01/02/2014 -- Add /input -- so you can stop arg processing
	If you wanted to be able to prompt starting with a hyphen, well,
	you couldn't do that before.  But now -- is honored and everything
	else is taken as the input prompt.

*** News 09/12/2013 -- New built in function $status_oneoff()
	This is a completely experimental function right now, which is
	helping me decouple the status bar generation from the windows.

	This function allows you to create your own status bar string,
	if you provide it a window and a status_format.
		$status_oneoff(winref ...status goes here...)
	As a simple example,
		$status_oneoff(1 %S)
	would return what %S would be on window 1.

EPIC5-1.1.6

*** News 07/31/2013 -- New action $LOGCTL(LAST_CREATED)
	The $logctl(LAST_CREATED) returns the log refnum of the most recent
	log that was created with /LOG NEW or $logctl(NEW).

*** News 07/31/2013 -- New action $LOGCTL(NEW)
	The $logctl(NEW) performs a /LOG NEW and returns the refnum of the
	newly created log.

*** News 07/31/2013 -- New /lastlog flag.
	/lastlog -regignore [regular-expression] avoids printing lines that
	would otherwise be printed without the regex.  Can be used in
	combination with -regex and -ignore and the other flags.

*** News 07/28/2013 -- $windowctl(REFNUMS_BY_PRIORITY) returns by current-ness
	The $windowctl(REFNUMS_BY_PRIORITY) operation returns all windows in
	the descending order that they have been the "current window". 

	This is based on your input screens -- if you have multiple windows 
	connected to multiple servers, this list doesn't care about that.  
	If you need that, iterate over the list and filter out the ones 
	for your server:

	fe ($windowctl(REFNUMS_BY_PRIORITY)) x { 
		if (windowctl(GET $x SERVER) == serverctl(FROM_SERVER)) {
			push results $x
		}
	};
	xecho -b Windows for this server in order of current-ness: $results


*** News 07/28/2013 -- FIXED-SKIPPED windows don't get channels on /window kill
	FIXED-SKIPPED windows (ie, /window fixed on skip on) are used to create
	status windows (see below).  They will no longer be given channels 
	from another window that is killed unless it is the last window 
	connected to the server.

*** News 01/09/2013 -- New /QUEUE flag, -RUNONE
	I can't believe I didn't think of this before!
	The /QUEUE -RUNONE flag will run the first command in a queue,
	leaving the rest of the queue alone.  This will work great with
	timer to create a FIFO queue that you can stagger commands through.
	For example, a command that slows down output to the server

		/TIMER -REPEAT -1 2 {QUEUE -RUNONE serverqueue}
		fe (#one #two #three #four) x {
			QUEUE serverqueue {join $x}
		}

EPIC5-1.1.5

*** News 11/28/2012 -- Lots of code quality improvements
	Ancient graciously set up epic to run under clang and its static
	analyzer, and we found lots of suggestions of things to fix.  

*** News 11/17/2012 -- New flag, /load -encoding
	You may now specify what a file's encoding is, and it will be
	converted automatically to utf8 (there will be a /set for this
	soon enough).  You can use this in your magic bootstrap:

	if (word(2 $loadinfo()) != [pf]) { load -pf -encoding iso-8859-1 $word(1 $loadinfo()); return; };

	This allows utf8 terminal users to load ascii art in 8859, and 
	it will Just Work!

EPIC5-1.1.4

*** News 8/5/2012 -- Anti-foot-shooting for $pad() and $repeat()
	Some had mentioned that you shouldn't be permitted to shoot yourself
	in the foot by asking for absurdly large output strings in $pad()
	and $repeat().  Normally I wouldn't agree to that, but I guess I'm
	getting soft in my old age....

*** News 8/5/2012 -- /xdebug no_color	- turn off all color unconditionally
	The experimental feature /xdebug no_color turns off color support
	at the lowest level of the client.  If you refresh your screen,
	any previously displayed color will be suppressed.  Turning off 
	this feature allows any color to be shown again.

*** News 8/5/2012 -- /LASTLOG -CONTEXT actually works correctly
	The /LASTLOG -CONTEXT feature shows you some lines before and after
	any lastlog match.  This has previously been broken and as of the
	time I write this, works properly, both normal and -REVERSE.

*** News 8/5/2012 -- /ON SET only thrown once when you type the exact name
	Previously, if you typed /SET <X> where <X> is the exact name of 
	any builtin SET, then /on set would be thrown twice.  That was never
	the intention, and this has been "fixed".

*** News 6/26/2012 -- Merge two windows together -- /WINDOW MERGE otherwin
	/WINDOW MERGE is like /WINDOW KILL except it moves everything from
	the current window to another window first.  This allows you to 
	"merge" two windows into one window.
	  * Window output
	  * Channels
	  * Logfiles
	  * Queries
	  * Timers
	If the current window can't be KILLed then everything will be moved
	away anyways, but you'll get an error message telling you that it
	can't be killed.  This isn't a bug.

*** News 6/26/2012 -- Expiring output -- /XECHO -E
	The /XECHO -E flag lets you create "expiring output" which will 
	disappear after however many number of seconds.
		/XECHO -E 10 This goes away in 10 seconds!
	This might be useful for status windows that you want to show new
	messages briefly.

*** News 6/24/2012 -- New $hookctl(CURRENT_IMPLIED_HOOK)
	When an implied /on hook is being expanded,
	$hookctl(CURRENT_IMPLIED_HOOK) is set to the name of that hook.

	I recommend wrapping this in an alias:
		alias ih {return $hookctl(CURRENT_IMPLIED_HOOK)
	Then you can use $ih() to get the hook.

	The use case is if you are using the same function for different
	implied hooks -- there was no way to tell the function which one
	it was working on.

*** News 6/24/2012 -- New $dccctl(SET refnum FULL_LINE_BUFFER [0|1])
	You can now set a dcc raw to "fully line buffered" mode.
	When this is turned on, /on dcc_raw will not trigger until
	a complete line is available.  You can turn this off by setting
	it to 0. 

	The corresponding $dccctl(GET refnum FULL_LINE_BUFFER) also works.

*** News 6/24/2012 -- New $dccctl(SET refnum PACKET_SIZE <bytes>)
	You can now set a dcc raw to "fully packet buffered" mode.
	When this is turned on, /on dcc_raw will not trigger until
	<bytes> bytes are available.  You can turn this off by setting
	it to 0. 

	The corresponding $dccctl(GET refnum PACKET_SIZE) also works.

*** News 06/24/2012 -- New option, /XEVAL -NOLOG 
	The -NOLOG option to /XEVAL suppresses logging for the command.
	The person who requested wanted to do something like this:
		alias ll { xeval -nolog {lastlog $*} }
	to be able to avoid logging /lastlog output.

*** News 06/24/2012 -- New option, /LASTLOG -IGNORE <pattern>
	The /LASTLOG -IGNORE <pattern> option allows you to display all of
	your lastlog EXCEPT whatever matches <pattern>.  In this way, it 
	acts as a reverse to the normal way.

EPIC5-1.1.3

*** News 03/24/2012 -- New status bar expando, %G (Network)
	The %G status bar expando shows the 005 NETWORK value for your server

*** News 06/09/2010 -- New semantics for /BIND TRANSPOSE_CHARACTERS
	The TRANSPOSE_CHARACTERS keybinding now has the following semantics:
	1. When the cursor is on the first character, swap the first and second
	   characters.  
	2. When the cursor is on a character (but not the first character), swap
	   the character under the cursor with the character before the cursor.
	3. When the cursor is at the end of the line (and not on a character), 
	   swap the last two characters on the input line.
	In all three cases, the cursor stays in whatever column it is in.

*** News 06/05/2010 -- New script: rejoin
	Stores channel/key on disconnect/part/kick. I hope it's useful!

	Allows you to rejoin all channels lost in a disconnect, by doing:
	/rejoin -all OR /rejoin -server

	See script for details. 

EPIC5-1.1.2

*** News 04/15/2010 -- New flags to $sar(), $msar(), case sensitivity
	In EPIC4, $sar() and $msar() were case sensitive.
	You could turn this off by using the 'i' flag.
	In EPIC5, $sar() and $msar() are case *INSENSITIVE*
	There has been no way to turn this off!
	You can now turn this off with the 'c' flag.

	Example:
		$sar(g/One/Two/one One one One)   -> "Two Two Two Two"
		$sar(cg/One/Two/one One one One/) -> "one Two one Two"

*** News 04/15/2010 -- New /on, /on unknown_set
	As a favor to howl, I've added /on unknown_set, which will be hooked
	whenever /set is called on a set that doesn't exist.
		$0  - The set that doesn't exist
		$1- - The value the user wanted to set.
	If you catch this, then the /on set that triggers for "unknown-set"
	will not be thrown.  If you don't know what I'm talking about, then
	you won't miss it.

*** News 04/01/2010 -- Can now backslash colons in server passwords
	Previously it was impossible to include colons in server passwords
	because colons are delimiters in server descriptions.  Now you can 
	backslash the colon and it will do the right thing.  Don't forget to
	backslash your backslashes!

		Real password		What you should use:
		-------------------	---------------------
		onetwothree		onetwothree
		one:twothree		one\:twothree
		one\two:three		one\\two\:three

*** News 03/25/2010 -- Can now modify servers by refnum (Fix to server descs)
	The /server command was broken in the epic5-1.1.1 release, and got 
	some extra work for the next release.  As part of this work, you can 
	now add change fields to a server refnum, like so:
		/server 1:type=irc-ssl
	Previously refering to a server refnum didn't support change fields.

EPIC5-1.1.1

*** News 3/19/2010 -- EPIC5-1.1.1 was released here

*** News 3/19/2010 -- The last value of /WINDOW SERVER is saved per window
	The last argument passed to /WINDOW SERVER is saved on a per-window
	basis, via $windowctl(GET x SERVER_STRING).  I added this because
	howl asked for it, although I don't know what he intended it for.

*** News 3/19/2010 -- Modifying server descriptions on the fly
	You may now modify server descriptions on the fly in the
	following situations:
		/SERVER -ADD <desc>
		/SERVER -UPDATE <desc>
		/SERVER <desc>
		$serverctl(READ_FILE filename)
		$serverctl(UPDATE refnum stuff)
		/WINDOW SERVER <desc>
	For example, let's say you created a server irc.foo.com, but
	you forgot that it used SSL.  Before it was a pain to "fix" that,
	but now you can fix it like this:

		/SERVER -ADD irc.foo.com:8855
	(oops, it uses ssl, i forgot!)
		/SERVER irc.foo.com:type=irc-ssl
	(aha! okie. now it will connect using ssl)
	
*** News 3/19/2010 -- You can log everything with /LOG SERVER ALL
	If you create a log like:
		/LOG NEW FILE myirc.log SERVER ALL ON
	that will log everything.
	Previously, you had to add each server individually by 
	refnum, but now you can just use the magic string "ALL"
	to refer to all servers.

*** News 3/19/2010 -- Rewrite /log support
	The /LOG command should work a lot better now.

*** News 3/19/2010 -- New target to msg a window, @E<winref>
	You can /echo to a window by /msg'ing its winref, as in:
		/msg @E3 This message will display in window 3.

*** News 3/19/2010 -- "global" now loads ambig and newnick scripts

*** News 3/19/2010 -- New /XEVAL -N flag, which reverses the ^ flag
	Normally if you run an alias or an on with ^, it will treat every 
	command as though it were prefixed with ^.  This suppresses the 
	output of many commands, which you may not want to do.  You can 
	negate the effect of ^ with /XEVAL -N.  For example:
		/on ^hook "chdir %" {
			xeval -n {cd $1-}
		}
	Normally the /CD command will output an error if it could not
	change the directory, but since ^ suppresses that error, in this
	example, you'd never know that the cd failed.  So you can wrap 
	commands whose error messages you want to see in /XEVAL -N.
	Don't forget!  Always wrap your commands in {} to avoid unintended
	back doors.

*** News 3/19/2010 -- Remember, $dccctl(GET refnum WRITABLE)
	If you combine $dccctl(FD_TO_REFNUM fd) with
	$dccctl(GET refnum WRITABLE) you can detect when a nonblocking
	connect has succeeded!

*** News 3/19/2010 -- New $dccctl(FD_TO_REFNUM <fd>)
	The $connect() function returns a file descriptor, which you can
	pass to $dccctl(FD_TO_REFNUM fd) to get the $dccctl() refnum,
	which you can use to do other stuff.

*** News 3/19/2010 -- New flag to /EXEC, -CLOSEOUT
	If you do /EXEC -CLOSEOUT it will close the stdin to the process,
	(ie, sends an EOF) which some processes need to decide that 
	they're supposed to do something.

*** News 3/19/2010 --  New scripts I should have documented
	These scripts have been added, but I never got around to documenting
	them.  That's a bummer.
		help.irc	history.rb	locale
		tabkey.sjh	logman		cycle
		set_color	ban		speak.irc

*** News 10/29/2009 -- Valgrind assistance
	If you want to run epic under Valgrind, you may want to pass the
	--with-valgrind  flag to configure, which will compile in some 
	additional assistance to help valgrind find memory leaks.  This
	support was graciously provided by caf, as were patches for the
	bugs he found using valgrind.

*** News 07/06/2009 -- The Fish64 xform actually works now (see below)
	The first implementation of FISH64 was not actually, to be technical,
	compatable with FiSH.  It is a strange thing and it took me a while
	to come up with an implementation of it that doesn't depend on how
	bits are stored in integers.  I have actually tested it against the
	real life FiSH implementation and it's correct now.

*** News 06/17/2009 -- New $xform, "FISH64"
	The FISH64 transform performs base64 encoding that is compatable
	with FiSH.  Fish64 uses the same algorithm as base64, but it uses
	a different character set.

*** News 06/08/2009 -- New $xform(iconv) functionality
	You can now refer to a pre-defined iconv encoding setup,
	instead of specifying encoding upon every use of
	$xform(iconv). Whereas you in the old days would do:

		echo $xform(iconv utf-8/ascii $stuff)

	which would take a lot of cpu time, as the client would have
	to do a lot of stuff to open, use, and then close up, the
	iconv stuff, you can now do as follows:

		@ id = iconvctl(ADD utf-8/ascii);
		echo $xform(iconv +$id $stuff);

	You can also do:

		echo $xform(iconv -$id $stuff)

	to reverse.
		Use /xdebug +unicode to debug iconv stuff!

*** News 06/08/2009 -- New control function: $ICONVCTL()
	This function works as follows:

		@ id = iconvctl(ADD fromcode/tocode[//option])
	
	This sets $id to a permanent identifier for doing encoding 
	from *fromcode* to *tocode*. (This may speed up encoding a 
	bit.)
		If the chosen encoding isn't accepted by iconv(),
	$iconvctl() returns empty.
	
		@ encoding = iconvctl(GET $id)

	This will return whatever you set the encoding $id to.

		@ iconvctl(REMOVE $id)

	This removes the $id from the table of encodings.

		@ iconvctl(LIST)

	This lists encodings.

		@ iconvctl(SIZE)

	And this returns the size of the iconv table.

	Do notice that identifers are re-used after removal.	

*** News 06/08/2009 -- Add USERINFO to /on hooks
	You can now add some information to "executing hooks".
	
		@ hookctl(USERINFO -1 stuff)

	This will set the USERINFO of the current executing hook 
	to "stuff". To get the userinfo of the current executing hook:

		echo $hookctl(USERINFO -1)

	This can be used in conjunction with changing the $* of a hook,
	to, for instance, add encoding information to a it.

*** News 06/06/2009 -- Can change $* in an /on hook now
	You can change the value of $* in an /on hook that will affect
	/on's with higher serial numbers.  
		@hookctl(ARGS <level> <new value of $*>)

	This is expected to be useful for things like iconv translation.  
	Please note carefully that the pattern matching of /on's against 
	$* is done *AFTER EACH ON IS RAN* so if you change $* you might
	affect which higher serial numbered /ons will run!

	Usually <level> is -1 and usually the new value of $* would be based
	on the current value of $*.  The change to $* takes place immediately.

	Example one:
		on #^hook -100 * {@hookctl(ARGS -1 >>>$0 $1<<< $2-)
		on ^hook * {echo $*}
		hook This is a test
	  would output
		>>>This is<<< a test
	  because the /on hook with serial number -100 changed the old value
	  of $*
		"This is a test"
	  to 
		">>>$0 $1<<< $2-"
	  which after expansion is:
		">>>This is<<< a test"
	  which is the value of $* in the /on hook with serial number 0.

	Example two:
		on #^hook -100 * {@hookctl(ARGS -1 $reverse($*))}
		on ^hook "ape" {echo APE! APE!}
		hook epa
	  would output
		"APE! APE"
	  because the first hook changes $* from its original value
		"epa"
	  to $reverse(epa) or
		"ape"
	  which matches the second hook.

*** News 04/10/2009 -- /WINDOW CHANNEL now outputs all channels in the window
	Previously, /window channel only output the window's current channel.
	It still does that, but now it will also output the full and complete
	channel list so you see the other channels in that window.

*** News 04/10/2009 -- /IGNORE user@host.com now works again
	Due to a really lame bug, /ignore user@host.com did not work 
	properly because the client thought it was a server name and did
	not fix it up to *!user@host.com which prevented it from matching
	anything which prevented it from working.  Sorry about that.

*** News 04/10/2009 -- Add permitted values for server desc "proto" field
	Previously, despite all of the documentation to the contrary, the 
	only permitted values were "0", "4", and "6" for "either", "ipv4 only"
	and "ipv6 only" respectively.  This has been increased so you can
	specify any of these values:

	For "either ipv4 or ipv6, I don't care" (the default)
		0
		any
		ip
		tcp
	For "ipv4 only, never use ipv6 for this server"
		4
		tcp4
		ipv4
		v4
		ip4
	For "ipv6 only, never use ipv4 for this server"
		6
		tcp6
		ipv6
		v6
		ip6

	Example to connect to an ipv6 server:
		/server irc.ipv6.foo.com:6665:proto=tcp6

	Example to connect to a server only using ipv4:
		/server irc.foo.com:proto=ipv4

EPIC5-1.0

*** News 12/25/2008 -- EPIC5-1.0 was released here.

EPIC5-0.9.1

*** News 12/12/2008 -- Configure will check for perl/ruby/tcl usability
	Up until now, configure would include perl/ruby/tcl as long as it
	existed and told us where its stuff was at.  That's bad if you
	don't install the dev packages, because linking against the 
	langauage library won't work if it's not there.  Configure will now
	try a test-compile to use the language embedding to see if it works
	and supports the api we expect.  Failures will cause that language
	to be turned off.  Be sure to re-run configure!

*** News 12/10/2008 -- New function, $chanlimit(#chan #chan #chan)
	The $chanlimit() command works just like $chanmode(), but it
	returns the +l argument -- the channel membership limit.
	This is by special request of fusion.

*** News 12/10/2008 -- Minor change to /SET NEW_SERVER_LASTLOG_LEVEL
	Previously, each time you connected to a server (received a 001 reply)
	the client would unconditionally assign all of the levels in 
	/set new_server_lastlog_level to the server's current window.
	This is rather annoying if you got disconnected from the server 
	because the default value is ALL,-DCC and that would clobber all of
	your window levels. 

	Having this brought to my attention, this has been changed to be 
	more reasonable.  These will now reclaim any unused levels, rather
	than unconditionally stealing them from other windows.  Thus, what
	/set new_server_lastlog_level ALL,-DCC means is, "each time I connect
	to a server, please put any levels that aren't being used by any 
	window connected to this server in the current window".   I apologize
	for the previous behavior which was stupid and shouldn't have survived
	as long as it did.

*** News 12/10/2008 -- Minor change to /SET OLD_SERVER_LASTLOG_LEVEL
	The same change applies when you /window server a window to a server
	that is already connected -- it has its window level changed to
	/set old_server_lastlog_level, but it will now NOT steal the level
	from any other window that already claims it.

EPIC5-0.9.0 (EPIC5-0.3.10)

*** News 11/24/2008 -- New /window operation, /WINDOW SCROLL_LINES
	The /WINDOW SCROLL_LINES operation overrules /SET SCROLL_LINES for
	one particular window.  The value may be -1 (which is the default, 
	and means use /SET SCROLL_LINES) or a positive number.

*** News 11/01/2008 -- New /SET, /SET DCC_CONNECT_TIMEOUT
	This set will control how long a nonblocking connect for a /dcc get
	or /dcc chat can go before the client decides to abandon it.  The 
	value is in seconds, and 0 turns this off (connects will not time out)
	The default value is 30 seconds.  This feature uses system timers,
	and you shouldn't change the value of this /set while a connect is 
	pending or you'll confuse things and your connects probably won't
	time out properly.

*** News 09/24/2008 -- New script, 'topicbar'
	The idea with this script is to use topline 1 of any window
	with a channel to display the topic of the given channel.

*** News 08/25/2008 -- /SET INPUT_INDICATOR_RIGHT now functional
	It was documented below, but the code for it wasn't finished until
	today.  So now it will start appearing on your input line.

*** News 07/01/2008 -- Add servers from file -- $serverctl(READ_FILE filename)
	You may now insert servers into the server list from a file using
	$serverctl(READ_FILE filename) where "filename" is the name of the 
	servers description file.  Note that the servers are appended to the
	end of the servers list!  The filename must be in the same format
	as the server description file that is loaded at startup.

EPIC5-0.3.9

*** News 06/25/2008 -- configure --without-wserv, job control
	Configure now checks your system to see if it has posix job control
	(which means you have setsid() and tcsetpgrp()) and if it does not,
	it turns off job-control features:
		/BOTMODE
		/EXEC
		/LOADing of compressed files
		$killpid()
		$exec()
		$open() of compressed files
		The -b command line option
		External crypto program support
		Asynchronous (nonblocking) DNS lookups
		Wserv support

	You can also use the --without-wserv flag to configure to turn off
	wserv support for a system that otherwise supports job control.
	There is nothing gained by omitting wserv support, only things
	removed.  Normally this flag wouldn't be added but I did it as a
	favor to someone.

*** News 06/25/2008 -- You can now /ignore a server
	Due to some networks (undernet) having annoying servers that 
	spam you 10-20 times a day with annoying messages you don't want
	to receive, it's now possible to /ignore a server:
		/IGNORE irc.server.com ALL

*** News 05/09/2008 -- Hitting ^C twice interrupts infinite loop
	Historically, if you hit ^C twice in a row and the client is stuck,
	it will send itself a SIGARLM.  In the past, this was because the
	client used blocking connect()s and stuff, and guarded them with
	alarm(3)s, so sending SIGARLM would cause an early interruption to 
	a blocking connect.

	Anyways, since we don't have any blocking stuff any more, this is no
	longer useful for its intended purpose.  You've been able to send the
	client a SIGUSR2 to raise a 'system exception" which attempts to 
	gracefully end an infinite loop in your script.  Hitting ^C twice
	in a row on a stuck client will send a SIGUSR2 which will cause an
	infinite loop in your script to terminate.

*** News 05/09/2008 -- New /SETs: INPUT_INDICATOR_LEFT, INPUT_INDICATOR_RIGHT
	This was written and contributed by fusion.  Thanks!

	The input line has been changed so the input prompt is always visible.
	When you reach the right or left side of the display, the input line 
	will still scroll side-to-side, but the input prompt will always be 
	visible, not just when you're at the start of the input line.

	Because it would otherwise not be obvious whether you are at the 
	beginning of the input line or not, there have been two new /set's
	added:
		/set input_indicator_left +
		/set input_indicator_right  +

	When there is more stuff on the input line than what is currently
	visible, if the extra stuff is off to the left, the first /set is
	used to tell you there is more in that direction.  If the extra 
	stuff is off to the right, the second set is used to tell you there
	is more in that direction.

	As of the time of this writing, the support for the second /set isn't
	ready yet, so there is no visual clue if you are at the end of the
	input line or not.  Keep watching for more info about this.

*** News 04/23/2008 -- Added new /on, /ON WINDOW_NOTIFIED
	This is hook is thrown when there's activity in a hidden window
	that is notified.
		$0 - The window refnum
		$1 - The level of the activity.
	Be careful with this hook, as output defered from it, may
	wreak havoc. 

EPIC5-0.3.8

*** News 04/10/2008 -- Added new /on, /ON SIGNAL
	You can hook signals with /ON SIGNAL
		$0 - The signal that was caught (a number)
		$1 - The number of times this signal has been caught
		     since the last time /ON SIGNAL was thrown
	Not every signal can be caught, and some signals are dangerous
	to catch.  For example, no matter what, you can't catch signals
	9 (KILL) or 13 (PIPE) 15 (TERM).  It's safe to catch 30 (USR1)
	and 31 (USR2), but everything else is entirely at your own risk.
	You should /defer anything you do within an /on signal to be safe.

*** News 04/10/2008 -- /USERHOST -FLUSH
	/USERHOST -FLUSH removes those userhosts which are "pending send"
	not those which are "pending receive".

EPIC5-0.3.7

EPIC5-0.3.6

*** News 03/10/2008 -- /NOTIFY list now applicable to local server.
	The notify list can now be updated on a per server basis.  This is done
	by placing the ":" nick before the list of local changes in the /NOTIFY
	command.  Everything on the /NOTIFY line UP TO the ":" is still
	applicable to every server.

	Examples:
	/NOTIFY : - [nicks]  # Clear local list and replace with [nicks].
	/NOTIFY - : [nicks]  # Clear all notify lists and add [nicks] locally.

*** News 01/28/2008 -- /ON WINDOW_COMMAND has command as $2        (kitambi)
	The command being executed is $2 in /on window_command.  If you do evil
	things with this, you may crash the client.  You Have Been Warned.

*** News 01/23/2008 -- New built in function $check_code(...)
	--- Warning --- This function is not really as useful as it looks
	because you would be unable to submit an invalid block statement
	or expression to the function without getting a warning from the 
	syntax parser in the first place.  I don't know how I will "fix"
	this, but maybe you might find the function interesting for now.

	The $check_code() function takes either a *block statement* 
	(surrounded by curly braces {}) or an *expression* (surrounded 
	by parenthesis ()) and tells you whether the item is well-formed
	and does not have any unmatched braces or parentheses.  It does 
	*NOT* tell you if the code or expresison inside the item is valid
	or even makes sense, it only tells you if it contains code that
	you could pass to /eval or to /@.

	Return values:
		 0 - The expression or block statement looks ok
		-1 - This is not an expression or block statement
		-2 - The expression/block statement is invalid, probably
		     because there is unmatched brace or parenthesis
		-3 - There is trailing garbage after the closing brace or
		     parenthesis.
	More return values will probably be added in the future as more
	errors become detectable.

*** News 01/22/2008 -- /SERVER listing now shows your vhost
	The listing of your servers from /SERVER now shows you the vhost
	that you're using (if any).  I forget who asked for this.

*** News 01/22/2008 -- Oper passwords no longer revealed with ^L
	Wjr pointed out that if you did /oper and typed a password that
	was hidden and hit ^L it would reveal the password.  This has 
	now been fixed.

*** News 01/22/2008 -- $ignorectl(SUSPEND) and $ignorectl(UNSUSPEND)
	Larne asked for a way to globally turn off /ignores for some 
	period of time.  So you can turn off all ignores globally 
	with $ignorectl(SUSPEND) and turn ignores back on again later
	with $ignorectl(UNSUSPEND).  A word of caution -- this is a counting
	queue, so each SUSPEND must be matched with an UNSUSPEND.  If you
	do two SUSPENDs and one UNSUSPEND, it will still be SUSPENDed.
	Use $ignorectl(RESET_SUSPEND) if you get the client totaly confused.

*** News 01/22/2008 -- You can /load executable files, with caution
	Crimedog said that all of his scripts on windows were executable 
	(+x) and epic wouldn't let him /load them, and so I've removed the
	restriction that you can't /load executable files.  I've replaced it
	with a warning that the file is executable and that /loading binary
	files yields undesirable results.

*** News 01/22/2008 -- /xecho -w -1 outputs to current window
	As a special favor to BlackJack, /xecho -w -1 will output to the
	current window, because this is what epic4 used to do, particularly
	when you did /xecho -w $winchan(#foo) and #foo was not a channel
	that you were on (so it returned -1) and it output to the curernt
	window.  In any other case but -1, /xecho -w to a window that 
	does not exist will drop the output.

*** News 01/22/2008 -- New built in function $strptime()
	Now you know the $strftime() function converts a $time() value into
	a string using a special format.  If you have the output of strftime
	and you have the format it was created with,t he $strptime() function
	will return the original $time() value it was created with.  This
	is probably useful by people who are parsing logfiles and want to 
	get a $time() value so they can do time math and see how long ago
	something occured.

	For the moment, this only works if you have strptime(3) on your
	system, and not everybody does.  Very soon, a compat version of 
	strptime() will be shipped with epic to ensure minimum functionality.

*** News 01/05/2008 -- You can now use arglists with /input (fusion)
	You can now use arglists with input, like so:

	input "Enter command and arguments: " (cmd, args) {
		xecho -b You entered [$cmd] and [$args]!
	}

*** News 01/03/2008 -- $info(o) values for libarchive, iconv support
	If the binary supports libarchive, $info(o) will include 'r'.
	If the binary supports iconv, $info(o) will include 'v'.
	Libarchive support is required to /load from a .zip file
	Iconv support is required to be able to do character set translation.

*** News 11/29/2007 -- New function $fix_width()
	The $fix_width() function takes the following arguments:
		cols      $0    Number of columns
		justify   $1    Justification (must be "l" for left justify)
		fillchar  $2    Fill character (a dword, so use " " for space)
		text      $3-   

	This function returns <text> formatted so that it takes up exactly 
	<cols> number of columns on the display.  It does this by adding
	<fillchar> to the string on either the left or the right or both.

	If <text> is already wider than <cols> then it is truncated to <cols>.

	<Justify> must be either "l" for left justify, "c" for center, or "r"
	for right justify.  Only left justify is supported.  The others are
	for future expansion.

	This function is intended for creating full width reverse toplines:
		@ :cols = word(0 $geom())
		@ :str = fix_width($cols l " " blah blah blah blah)
		window topline 1 "^V$str"

	You will probably want to call $fix_width() in a separate statement
	from the /window topline in order to avoid the syntactic confusion
	with passing a double quoted word to /window and passing a double
	quoted word to $fix_width() (the space).  Trust me.  Don't go there.

	In the future, support will be added for right justify and centered.
	Please keep watch out in this document for more info.

*** News 11/29/2007 -- Support for ZIP files from libarchive
	Support for loading files from .zip files has been added.  This first
	round of implementation just adds the raw ability, but it's not 
	totaly ready to be used yet.  You're welcome to start playing with
	it and reporting any problems you have.

	You can $open() a file for reading or /load it from a zip file:
		/load foo.zip/file
	and
		@fd = open(foo.zip/file R)

	If you /load a zip file, it will load the file ".ircrc" in the top
	level directory.  This might be enahanced or changed in the future:
		/load foo.zip
	acts like
		/load foo.zip/.ircrc

	Some operations cannot be performed on zipped files, such as $fseek()
	and $frewind() and so forth.  This might change in the future.

	All of this is based on 'libarchive' being installed.  You will need
	to re-run configure in order to pick up libarchive support after you
	do a cvs update.

*** News 10/22/2007 -- New flag to /lastlog, /lastlog -window
	The /lastlog -window flag lets you grep the lastlog from a different
	window.  The output will still go to the *current* window, however!

*** News 09/19/2007 -- Some built in functions now 'builtin' aliases.
	Several functions that have been deprecated by $xform() have been
	demoted from built in functions to aliases in the 'builtins' script.
		encode		decode		b64encode	b64decode
		urlencode	urldecode	sedcrypt	sha256

EPIC5-0.3.5

*** News 09/14/2007 -- New built in function: $splitw(<delim> <string>)
	This function takes a <string> which has sections delimited by
	<delim>.  The <delim> argument can only be one character.  An
	obvious example of this is $PATH which is a <string> that uses
	the colon as <delim>.  The <delim> is a dword, so you can use the
	space as a delimiter if you needed to.

	This function unconditionally converts <string> into a dword list.
	You will need to xdebug dword to iterate over the return value, or
	you can use $unsplit(" " $splitw(<delim> <string>)) to collapse it
	to a uword list (although this is probably pointless)

	Example:
		@ directories = splitw(: $PATH)
	might return
		/bin /sbin /usr/bin /usr/sbin /usr/local/bin

	Example:
		@ foo = splitw(: one:two:three a berry:four:five)
	returns
		one two "three a berry" four five

*** News 09/13/2007 -- New /SET, /SET STATUS_HOLDMODE
	This is the value that %{1}H expands to.  The default is " (Hold)".
	If you don't like the "(Hold)" in your status bar when your window
	is in hold mode but not holding anything, unset this variable 
	entirely:
		/set -status_holdmode
	and it will disappear.  Or remove %{1}H from your status format.

*** News 09/13/2007 -- New status expando, %{1}H, hold mode indicator.
	The %H status expando expands when your window is in hold mode
	*and* there is something being held.  But if your window is in
	hold mode, but nothing is held, you can't tell just by looking.
	So the %{1}H status expando will expand whenever the window is in
	hold mode *except* when %H will expando.  This is so you can put
			%{1}H%H 
	in your status format and one or the other (but not both) will 
	expand at all times hold mode is on.

	This expando has been added to the client's default status format.
	The value of %{1}H is controlled by /set status_holdmode, and the
	default value of that is " (Hold)"

*** News 09/13/2007 -- Remember, *0 is an rvalue, but *var is an lvalue
	The deref operator ("*") converts a token into an rvalue and then
	uses that as an lvalue.  Example:
		assign foo bar
		@ *foo = [testing]
		echo $bar
	in the above (*foo) is the same as "bar".  But numbers are different.
	Derefing a number yields an rvalue:
		alias oofda { @ foo = *0 }
		offda one
	To convert an argument into an lvalue, deref it a second time:
		alias booya { @ *(*0) = 'testing' }
		booya varname
		echo $varname
	Does that make it clear?

*** News 09/02/2007 -- New function, $is8bit(string)
	This function find the first character of string that has the
	eight bit set. Useful to discover Unicode-strings or other 
	non-eight bit characters.

*** News 08/22/2007 -- New xform, $xform(ICONV "from/to" text)
	If your binary is built with iconv support (re-run configure before
	you tell me it doesn't work!) then you will be able to use iconv()
	to translate strings from one character encoding to another.  This
	might be useful to experimentally convert to and from utf8 while 
	you wait for the unicode-enabled input line to be written.

XXX Help files updated to here XXX

*** News 08/22/2007 -- Checks for iconv in configure: --with-iconv
	Configure will now check for libiconv support.  Normally it will
	look for libiconv.a and iconv.h in the --prefix/lib and 
	--prefix/include directory, or in /usr/local or /opt or /usr/opt.
	If your iconv support is not in any of these places, then you must
	supply the --with-iconv=/path/to/dir where that directory contains
	"include/iconv.h" and "lib/libiconv.a"

*** News 08/22/2007 -- Checks for alt place in configure: --with-localdir
	A lot of times you'll have software installed but it won't be in
	the places the compiler looks for it.  Usually this is /usr/local
	or /opt or something.  You can use this flag to tell configure to
	look in this directory in addition to the default directories.  This
	can aid in ensuring optional dependancies don't get turned off because
	they're in a directory the compiler doesn't look in.

*** News 08/13/2007 -- New /SETs, /SET DEFAULT_REALNAME and DEFAULT_USERNAME
	Per a discussion on the list, it was decided to introduce two new
	sets, /SET DEFAULT_REALNAME and /SET DEFAULT_USERNAME.  These control
	what realname and username should be sent to the server each time you
	connect.  Although I can't stop you, you shouldn't unset these 
	variables if you know what is good for you.

	/SET DEFAULT_REALNAME replaces /SET REALNAME, which confused people
	who though that /setting it changed their realname immediately

	/SET DEFAULT_USERNAME replaces /ircuser which confused people
	because they thought it was going to be a /set.

*** News 07/20/2007 -- New /SET, /SET LASTLOG_REWRITE
	This provides the default value to /lastlog -rewrite whenever you
	don't use the /lastlog -rewrite flag.  If you always want the 
	/lastlog command to timestamp each line, try this:
		/set lastlog_rewrite $strftime($1 %T) $8-

	(Remember, if you use this in a pf loaded script to double up $'s...
		/set lastlog_rewrite $$strftime($1 %T) $$8-		)

*** News 07/20/2007 -- New flag to /LASTLOG, /LASTLOG -REWRITE
	The /LASTLOG -REWRITE flag rewrites each lastlog line with the 
	following values for $*
		$0  - The lastlog item's unique refnum
		$1  - Timestamp (suitable for use with $strftime())
		$2  - Window refnum
		$3  - Output level
		$4  - Reserved for future use
		$5  - Reserved for future use
		$6  - Reserved for future use
		$7  - Output target
		$8- - The logical line of output
	Not all of these field are intended to be useful -- but I don't
	know what sort of imaginative things people might come up with.

	Example:  To put a timestamp before every line...
		/lastlog -rewrite "$strftime($1 %T) $8-"

*** News 07/04/2007 -- New $windowctl() option, $windowctl(GET refnum CHANNELS)
	You can now fetch all of the channels in a window by using
		$windowctl(GET <refnum> CHANNELS)
	There is no defined order to the channels returned.

*** News 07/02/2007 -- Clarification of single-indirection implied hooks
	Earlier, I said...
	   You may surround <string> with {}s if you wish, to avoid quoting
	   hell.  Match sure to keep your {}s matched up if you do so.
	   See the "loadformats" info above for how to practically use this.

	But due to a mistake, this never worked correctly.  This is now 
	fixed.  If you compare the normal two-expansion version:
		@ var = [format_send_public]
		@ fmt = '<%W$N%n> $1-'
		addset $var int
		@ hookctl(set list send_public implied \\$var);
		set $var $fmt
	which ties /on send_public dynamically to the value of 
	/set format_send_public.

	If you wanted to do it directly, and not tie the implied hook to a
	variable, you can surround the format in curly braces, like so:
		@ hookctl(set list send_public implied {<%W$N%n> $1-})
	Remember that curly braces protect the insides from $-expansion,
	so for all purposes the inside of {}s is a literal string that is
	not expanded except each time the /on is thrown.

*** News 07/02/2007 -- Clarification on using $'s in expression parser
	The old expression parser used to allow you to use expandos as
	lvalues in order to indirectly assign to variables, for example:
		alias inc {
			@ $0 += 1
		}
	But this is not supported in the new math parser.  Instead you 
	have to use the deref operator, like so:
		alias inc {
			@ *0 += 1
		}
	which does the same thing.

*** News 06/25/2007 -- New flag to /XECHO, /XECHO -AS
	The /XECHO -AS flag will output a message to all windows on the
	current server.  You can combine this with the -S flag if you want 
	to output to another server.  This might be good for blasting a 
	message when you're disconnected from the server.  For another
	example, see below.

*** News 06/25/2007 -- New window level, SYSERR
	The SYSERR window level will now be used for all of those layered
	"INFO --" system errors ("syserrs").  The client will make its best
	effort to ensure that these messages go to the correct server's
	windows, and that /on yell is hooked in the correct server context.

	You can combine this with /xecho -as  [see above]...
		on ^yell "* INFO -- *" {xecho -as $*}

EPIC5-0.3.4

*** News 06/02/2007 -- New tranformer: $xform(ALL)
	The $xform(ALL) transform ignores the text and returns a list of
	all supported tranformers.  If the user didn't compile SSL support,
	then you won't be able to use the strong crypto transforms, so this
	is a great way to check before trying to use crypto.

	Since this is a regular old transform, you can further transform it
	any way you want to (if you want to).

*** News 06/01/2007 -- More $xform()s added to the list below
	Please re-read the list immediately below, it's been updated!

*** News 06/01/2007 -- Totaly rewritten $xform(), now actually useful!
	Here's the plan -- we're going to do this over again a second time.

	$xform("<transformations>" "meta" "meta" text)
	     
	Where the <transformation>s are supported by transform_string().
	At the time i write this, they are:

	Reversable encodings that convert between binary and printable data
	and do not require a meta:
	     URL             URL encoding           
	     ENC             Base16 encoding        
	     B64             Base64 encoding        
	     CTCP            CTCP encoding          
             NONE            Straight copy -- data not changed

	Reversable encryptions that require a meta value (a password):
	     SED             Simple Encrypt Data  
	     BF              Blowfish-CBC         
	     CAST            CAST5-CBC            
	     AES             AES256-CBC
	     AESSHA          AES256-CBC with SHA256 digest of the meta
	     DEF             Default encryption     (NOT IMPLEMENTED YET!)

	Irreversable digest operations that do not require a meta:
	     SHA256          SHA256 message digest

	You can string together multiple transformations.  Any transformation
	that requires a meta value (ie, a cipherkey) should be supplied after
	the transformations *in the correct order*.  After this should be the
	plain text.  To apply a transformation, prefix its name with a plus sign
	("+") and to remove a transformation, prefix its name with a minus sign
	("-").  For example, +URL means url encode, and -URL means url decode.

	The transformations is a dword (must be surrounded by double quotes if
	it contains a space, which it will if you do multiple transformations).
	The meta values are dwords (must be surrounded by double quotes if 
	they contain a space).  These two things make this function behave 
	differently than functions normally do, so this is a documented 
	deviancy!

	Examples:
	     URL-encode a string     $xform(+URL this is a string)
	     URL-decode a string     $xform(-URL this%20is%20a%20string)
	     SED-cipher a string     $xform(+SED password this is a string)
	     SED-decipher a string   $xform(-sed password <whatever>)

	More practical examples:
	1) Read binary data from a file, encrypt it, and url encode it again.
	     @fd = open(file.txt R)
	     @data = read($fd 1024)
	     @cipher = xform("-CTCP +SED +URL" password $data)
	     @close($fd)
	     msg someone $cipher

	Why does this work?
	 -- $read() returns ctcp-enquoted data, so -CTCP removes it
	 -- Now we have binary data, so +SED will cipher it
	 -- Now we have ciphertext, so +URL will url encode it.

	We can send this to someone else, and they can put it in $cipher...

	     @newfd = open(newfile.txt W)
	     @newdata = xform("-URL -SED +CTCP" password $cipher)
	     @writeb($newfd $newdata)
	     @close($newfd)

	We did the reverse of the above:
	 -- We -URL to recover the binary data
	 -- We -SED to decrypt it using the password
	 -- We +CTCP to put it in a form we can use with $writeb().

	Viola!

*** News 05/31/2007 -- New math operator, Unary ** operator
	The ** unary prefix operator treats its operand as an unexpanded
	string, permitting it to be expanded again.  This should remove
	most any need to use /eval.  Example:
		@ foo = 'this is a string with $var in it'
		@ var = 'one'
		@ val1 = **foo
		@ var = 'two'
		@ val2 == **foo
	So $val1 is "this is a string with one in it" and $val2 is
	"this is a string with two in it".  You can apply this operator
	to any rvalue (natch):
		@ foo = **[booya $foo booya!]
	becomes
		@ foo = **[booya this is a string with $var in it booya!]
	becomes
		@ foo == [booya this is a string with two in it booya!]
	Yeah!

*** News 05/17/2007 -- New script, 'dcc_ports'
	This script adds two new sets:
		/set dcc_port_min <number>
	and
		/set dcc_port_max <number>
	which create a port range which will be used by your dcc's.  Each
	number in the range will be chosen sequentially (this will be 
	enhanced further in the future) and any time you change either of 
	these values you reset the sequence and it goes back to the minimum
	port again.  When the port range is exhausted, it will cycle back
	again to the minimum port.

*** News 05/17/2007 -- Default ports for dcc, recovering from ports-in-use
	There are two new features, that sort of complement each other, and
	sort of overlap each other.  Whether to use one or the other depends
	on how your script is set up and what your needs are.

	When you use /dcc chat or /dcc send, but do not use the -p flag to
	specify a port, you can create a callback to your script to create
	a default port:
		@dccctl(DEFAULT_PORT <string>)
	The <string> will be expanded with $* being the dcc refnum of the
	dcc needing a port number.  The expanded value of <string> will be
	used to set the WANT_PORT value (see below).  If a port cannot be 
	used because it is already in use by someone else, this value will
	be expanded repeatedly until it generates a port we can use.

	Whenever you use /dcc chat or /dcc send and use -p to bind a port,
	and that port is already in use, then /on dcc_lost will be thrown
		$0	The person you're sending it to
		$1	The dcc type (ie, send or chat)
		$2	The filename (url encoded)
		$3	The refnum of the dcc
		$4	The port that we tried to use that is in use.
		$5-	The literal string "PORT IN USE"
	It is expected that you will catch this hook and then do
		@dccctl(SET $4 WANT_PORT <new port>)
	and then the client will try the new port again.  If a port cannot
	be used because it is already in use by someone else, this hook will
	be thrown repeatedly until you set a port we can use, or you don't
	change the port

	In either case, if a port is in use, and neither the DEFAULT_PORT
	nor /ON DCC_LOST results in the WANT_PORT value being changed, then
	the dcc will be considered a failure, and it will be abandoned, 
	just as it has always been until this change.

*** News 05/15/2007 -- New dccctl() value, "WANT_PORT"
	You can now get and set the "WANT_PORT" value for a dcc, which 
	is the value that the -p flag sets.  You can therefore use this
	to change the -p flag for the user.  Because this value is only
	used at the time the client tries to bind() an inbound address,
	changing this value after the dcc has already been offered will
	have no effect: it is too late at that point to change the port.

*** News 05/15/2007 -- Normalization of the /SAY and /SEND commands
	The old behavior:
	- /SEND (the command used when you type text) sends a message to 
	       1) Current Query
	    or 2) Current Channel
	    or 3) Nowhere
	- /SAY (the "empty" command to overrule a query) sends a message to 
	       1) Current Channel
	    or 2) Nowhere

	The new behavior:
	- /SEND sends a message to 
	       1) Current Query
	    or 2) Current Channel
	    or 3) Nowhere
	- /SAY sends a message to 
	       1) Current Channel
	    or 2) Current Query		<--- this is the change here
	    or 3) Nowhere

	So /SAY will now send a message to the current query if you are
	not on a channel, rather than silently failing.  Until this change,
	there was no way to do this at all.

*** News 05/13/2007 -- ***IMPORTANT*** Removal of /SET REVERSE_STATUS_LINE
		*** IMPORTANT *** IMPORTANT *** IMPORTANT ***
	The /set reverse_status_line variable has been removed.  This means
	that your /set status_format* variables, and your /window status_format
	variables will not be auto-prepended with ^Vs like they have always
	been since time immemorial.  You ***MUST*** prepend your status_formats
	with ^Vs (control-Vs, ie, the reverse character) if you want them to
	appear in reverse. 

EPIC5-0.3.3

*** News 04/11/2007 -- New argument to /xecho, -TARGET
	Usually if you wanted to output to a channel's window, you would do
		/xecho -w $winchan(#channel) ....
	but someone asked (nicely) for a -target flag to /xecho because they
	thought that looked nicer.  So now you can do
		/xecho -t #channel ...

*** News 04/11/2007 -- New serverctl operation "ALLGROUPS"
	The $serverctl(ALLGROUPS) operation will return a unique list of all of
	the group names used in all your servers.  Do not pass any arguments 
	to this operation, since that's reserved for future expansion.

*** News 04/11/2007 -- New gettable serverctl value, "FULLDESC"
	This value returns the fully qualified server description for a
	server, suitable for writing into a servers file, or for passing
	to /server -add or any old place.  You can only get this value; 
	you cannot set it, although that's planned for the future.

*** News 01/27/2007 -- New serverctl value, "AUTOCLOSE"
	Normally a server is automatically closed when the last window
	to it is disconnected (or killed).  Many people hate this behavior
	and would like to have a server's connect persist even if there
	are no windows connected to it.  You can now control this behavior.
		@serverctl(SET refnum AUTOCLOSE 0)
	The default is 1 (natch).  The behavior of turning this off is 
	not well understood.  Use with caution, tell me of any troubles.

*** News 12/09/2006 -- New configure flag, --with-threaded-stdout
	EJB has graciously provided us with a patch to do threaded output.
	This code will hopefully work around problems with gnu screen
	blocking, causing pinging out of your servers.  You can turn this
	code on by using the '--with-threaded-stdout' flag to configure.
	This code is sort of experimental at this time, but this is an
	officially supported feature, so please do report any problems you
	have with it.
	
	If you turn this option on, then $info(o) will include 'o'.

*** News 11/17/2006 -- New flag to /userhost, /userhost -extra
	The /USERHOST -EXTRA flag causes the present value of a variable 
	to be appended to the $* value of the results of a userhost query.

	Remember that /on 303 and /userhost -cmd are hooked with:
		$0	Nickname
		$1	+ if oper, - if not oper
		$2	+ if away, - if not away
		$3	Username
		$4	Hostname
	and now,
		$5-	Contents of the -extra variable

	If you do not hook /on 303 or use /userhost -cmd, the contents of the
	-extra variable are appended to the output.

	Usage:
		@ foo = [extra stuff]
		USERHOST hop -extra foo -cmd {xecho -b $0!$3@$4 ($5-)}
	might output
		*** hop!jnelson@epicsol.org (extra stuff)

	Remember that the argument to the -extra flag is *an lvalue*
	(the name of a variable) and is not literal text!  The lval is 
	expanded (yeilding an rval) and *that* value is saved, and used
	later on.  You can freely change the value of the variable and 
	that will not affect anything:

	Example:
		@ foo = [extra stuff]
		USERHOST hop -extra foo -cmd {xecho -b $0!$3@$4 ($5-)}
		@ foo = [blah blah]
	still will output:
		*** hop!jnelson@epicsol.org (extra stuff)
	because the value is saved at the time you do /userhost.

EPIC5-0.3.2

*** News 11/04/2006 -- $shift() and $unshift() changed like $push() and $pop()
	In the same way that $push() now only takes two arguments, $unshift()
	now takes only two arguments, an lval, and an unquoted dword:
		@ foo = [one two]
		@ unshift(foo three four)
			($foo is one two "three four")

	In the same way that $pop() now only takes one argument, $shift() 
	now only takes one argument, an lval:
		@ foo = [one two three]
		@ booya = shift(foo)
			($booya is one and $foo is two three)

*** News 11/04/2006 -- The /xdebug command now takes a block argument
	You can temporarily change the /xdebug value for a block of
	statements, and have the original value restored, all in one
	operation:

		xdebug dword { @push(bar "one two") }
		@ foo1 = leftw(1 $bar)
			($foo1 is "one)
		xdebug dword { @ foo2 = leftw(1 $bar) }
			($foo2 is "one two")

*** News 11/04/2006 -- New built-in function, $curcmd() (nullie)
	With this one you know what command is currently executing. For
	instance, you might want to not timestamp lines generated by
	/lastlog to avoid having doubled timstamps, or prepend them with
	something else than timestamps.

	Example:

	/set output_rewrite ${ curcmd() == [lastlog] ? [Lastlog:] : Z} $1-

*** News 11/4/2006 -- Changes to how $push() works with dwords
	Up until this time, the $push() function treated the arguments
	as a list of words, rather than a single word that needed to be
	added.  This caused confusion because people didn't have a way
	to add a word atomically.  For example:

		xdebug dword
		assign foo one two
		@push(foo three four)
		echo $foo                 returns   one two "three four"

	So from now on, $push() takes two arguments, an lvalue, an a
	single *unquoted* dword.  If the dword contains spaces, then it
	will be quoted before added to the variable (as above)

*** News 11/4/2006 -- Changes to how $pop() works (along with dwords)
	Up until this time, the $pop() function has had two conflicting
	behaviors, that ambiguously overlapped:
		$pop(word word word word)
		$pop(lval)
	But what happens if you have a word list with only one word?
	For this reason, $pop() wasn't really useful, and people avoided
	using it.  So I've removed the ambiguity by removing the first
	case, which isn't used.

	If you want to $pop() off a word list, use $rightw(1 <word list>)
	which has the proper behavior.

	Further, to differentiate $rightw() from $pop(), the $pop()
	function returns an *dequoted string*, as in this example:

		assign foo one two "three four"
		xdebug dword
		echo $rightw(1 $foo)       returns "three four"
		echo $pop(foo)             returns three four

*** News 10/25/2006 -- New field for server descriptions, 'vhost'
	You may now specify per-server virtual hostnames ("vhosts")
	by using the "vhost" field in a server descrpition.  

	Example:
		/server irc.foo.com:vhost=my.other.hostname.com

	Obviously the value you set for vhost must be a value that 
	you could use for /hostname; that is, a hostname or p-addr
	that resolves to an address your machine will let you bind().
	Specifying a vhost for the wrong protocol will malfunction.
	 (For example: /server irc.foo.com:pr=4:vhost=[ff::ff] )

*** News 09/24/2006 -- New /ON, "WINDOW_SERVER" (nullie)
	This new hook will trigger every time window's server is
	changed. This might be useful for purging window-related data
	such as scripted window-to-channel associations. The params are:
	
	$0 - affected window's refnum
	$1 - old server
	$2 - new server

*** News 09/19/2006 -- New /ON, "UNKNOWN_COMMAND" (nullie)
	The hook is thrown whenever an unknown command is entered on the
	command line or specified in a script. This way, you can
	complete a partial command name with an ambiguous hook. The
	params are:

	$0 - command being unknown
	$1 - parameters, if any

*** News 09/19/2006 -- New /ON, "CHANNEL_LOST" (nullie)
	This hook is thrown whenever the client leaves a channel for any reason.
	It is meant as a generalized way for clearing channel-related structures
	in scripts, instead of having to hook on several other ones (KICK, PART
	and SERVER_LOST). The parameters are:

	$0 - server refnum where a channel is being destroyed
	$1 - channel name

	Have one thing in mind, though. This hook isn't necessarily
	thrown in current server's context. Instead of $servernick(),
	you need to do $servernick($0).

*** News 09/18/2006 -- New keybinding, "RESET_LINE"
	The "RESET_LINE" keybinding is intended to be used by tabkey and
	history recall scripts that need to be able to replace the contents 
	of the input line without affecting the cut buffer.

	The RESET_LINE keybinding takes an argument (naturally) that is
	the new value of the input line.  If you don't provide an argument
	then the input line is simply cleared.

	Example:
	   /alias oofda {parsekey reset_line this is your new input line}
	   /bind ^I parse_command {oofda}
	   (type something and then hit tab)

*** News 09/18/2006 -- New field in server description, "proto"
	A new (seventh) field, has been added to server descriptions.
	This field is named "protocol" (see below) and can be abbreviated
	as "pr".  The field restricts what socket protocols you want epic
	to use for this server.  Possible values are;
		tcp4    or   4		IPv4 only
		tcp6    or   6		IPv6 only
		tcp     or   any	IPv4 or IPv6, doesn't matter
	Naturally, "any" is the default.

*** News 09/16/2006 -- Enhancements to server descriptions
	You are permitted to skip fields in the server description.  Each
	field has a name that you refer to it by:
		host	port	pass	nick	group	type

	You can change the field by "assignment", like so:
		/server irc.foo.com:group=efnet
	or
		/server irc.foo.com:7005:type=irc-ssl

	The server description actually skips to the name field, so
	you don't have to specify them in order, you can skip around
	however it suits you.  If you skip to a field and then don't
	specify the name of the following field, whatever field would
	naturally follow is assumed:

		/server irc.foo.com:7005:group=efnet:irc-ssl

	In this case, you skipped to the 'group' field, and the field
	that follows that is 'type', so that is what "irc-ssl' is 
	assumed to be.

	This can be used any place server descriptions are taken, 
	from /server, to /window server, to ircII.server, to the
	command line, everywhere.

*** News 09/14/2006 -- WINDOW command will fail if given invalid window refnum
	Consider the behavior of
		/WINDOW foo KILL
	in the two cases where the window 'foo' does and does not exist.

	Up until now, if 'foo' existed, it will killed, but if 'foo' did
	not exist, this was not an error, and the current window was killed.
	Ignoring window refnums that didn't exist was intended as a courtesy
	but actually is a hassle, since most people assume that any further
	commands will operate only on that window, or not at all.

	So this behavior has been changed.  In the above case, if 'foo' does
	not exist, then the /window command simply fails at that spot and
	all further actions in that statement are ignored.

*** News 09/14/2006 -- Some functions no longer support double quoted words
        The work was done by nullie.

        The following functions have previously supported double quoted
        words, but due to popular sentiment, this support is being removed
        to make the functions useful with untrusted data (ie, irc stuff)

        match           rmatch          userhost        word
        remw            insertw         chngw           pattern
        filter          rpattern        pop             findw
        findws          splitw          diff            sort
        numsort         uniq            remws           getsets
        getcommands     getfunctions    prefix          rfilter
        copattern       corpattern      cofilter        corfilter

        The following functions didn't support dwords unless you did
        /xdebug extractw.  Nothing has changed for these functions.

        leftw           rightw          midw            notw
        restw           insertw         chngw           beforew
        tow             afterw          fromw           splice

        The /FE and /FOR I IN (list) commands previously supported
        double quoted words, but now they will not.

        You can revert to the epic4 behavior (for now) with /xdebug dword

*** News 08/31/2006 -- New serverctl value, $serverctl(GET refnum ADDRSLEFT)
	This returns the number of addresses left from the last dns lookup
	for the server "refnum".  It's intended to be used like this:
		on server_status "% % CLOSED" {
			if (serverctl(GET $0 ADDRSLEFT)) {
				# epic will be reconnecting...
			}
		}
	This will help nullie write his reconnect script to not trap until
	epic is finished with what it wants to do.

*** News 08/17/2006 -- Add support for OPERWALL, /ON OPERWALL, OPERWALL level
	OPERWALL is an efnet server command which servers send amongst
	themselves.  The local server very graciously sends it to you 
	wrapped in a WALLOPS (big ups to whoever was considerate enough
	and made this decision, you made it much easier to support!)
	so we peer at WALLOPS and route "OPERWALL - " messages to 
	/ON OPERWALL and a new OPERWALL window level.  Now black will 
	stop asking me to add this.

	This is all opt-in.  If you don't hook /on operwall, then they
	continue going to /on wallops and WALLOPS as they have always done.

*** News 08/17/2006 -- Two new server states CREATED and DELETED
	When a server is added to the server list, it's initial state is
	"CREATED".  Once it is initialized, it is switched to the "RECONNECT"
	stage.  So you can trap new servers with:
		/on server_status "% CREATED %" {xecho -b New server is $0}

	Servers that are deleted are switched to the DELETED state:
		/on server_status "% % DELETED" {xecho -b Server $0 going away}
	No matter what you do, you cannot stop the deletion of a server.
	It is recommended you /defer any changes you make to the server list
	from within an /on server_status that traps a server deletion.

EPIC5-0.3.1

*** News 07/01/2006 -- Revamped /encrypt command
	The /encrypt command, which is a legacy ircII command, has been 
	revamped and a lot of functionality has been added (see below).
	The new format looks like this:

	/encrypt
		See the cipher list

	*** Ways to specify the other person (TARGET is a NICK or CHANNEL)
	/encrypt TARGET key
		Trade messages with 'nick' on all servers using 'key'.
	/encrypt 0/TARGET key
		Trade messages with 'nick' only on server 0 using 'key'
	/encrypt SERVERNAME/TARGET key	
		Trade messages with 'nick' on server 'name' using 'key'.
		'name' can be "ourname", or "itsname"
	/encrypt SERVERGROUP/TARGET key
		Trade messages with 'nick' on group 'group' using 'key'.
	/encrypt ALTNAME/TARGET key
		Trade messages with 'nick' on servers with 'altname' using
		'key'.

	*** Ways to specify the cipher you want to use
	/encrypt nick key
		Trade SED messages (default)
	/encrypt nick key /path/to/program
		Trade messages ciphered by an external crypto script
	/encrypt -SED nick key
		Trade SED messages
	/encrypt -SEDSHA nick key
		Trade SED messages, using the SHA256 hash of 'key'.
	/encrypt -CAST nick key
		Trade CAST5-CBC messages 
	/encrypt -BLOWFISH nick key
		Trade BLOWFISH-CBC messages 
	/encrypt -AES nick key
		Trade AES256-CBC messages 
	/encrypt -AESSHA nick key
		Trade AES256-CBC messages, using the SHA256 hash of 'key'.

	Compatability and availability of all these is discussed in the 
	following news bulletin.

	AESSHA is the best cipher, but if it's not available, then SEDSHA
	is better than SED, but SED is universally compatable.  If you must
	use SED, use a very long key string (40 chars)

	For backwards compatability, messages ciphered by external crypto
	scripts are always sent as SED messages, even though they are not,
	technically, ciphered with SED.

*** News 06/26/2006 -- CAST5, Blowfish, AES, and AES-SHA encryption support
	This only took me 8 years to catch up with ircII!  Yah!
	All of this support is provided by the EVP api in OpenSSL.
	If you don't have OpenSSL, this won't be supported.

	The /ENCRYPT command now takes an argument after the nickname,
	which can be either -CAST, -BLOWFISH, -AES, or -AESSHA
		/ENCRYPT nick -CAST key
		/ENCRYPT nick -BLOWFISH key
		/ENCRYPT nick -AES key
		/ENCRYPT nick -AESSHA key

	The -CAST support is fully compatable with ircII (as of the last 
	time I tested it) and -BLOWFISH isn't compatable with anything 
	but your fellow bleeding edgers (since nobody supports that yet).
	Same thing with -AES (only supported by EPIC for now).

	The Blowfish support is *NOT* compatable with FiSH, because FiSH
	has a couple of idiosyncracies (non-standard Base64, supports keys
	that are longer than what openssl supports).  It would probably be
	more likely that FiSH needs bag-on-the-side support than to be 
	mainlined into the crypto support...

	AES uses a fixed 256 bit key.  This is 32 characters.  If your key
	is not 32 characters, it is padded out with nuls (ascii 0).  If your
	key is more than 32 characters, it is truncated.

	AESSHA runs your key through SHA2 to generate a 256 bit key for AES.
	In any other way, it works the same as AES.

	SEDSHA runs your key through SHA2 to generate a 256 bit key for SED.
	In any other way, it works the same as SED.  Unlike the others, this
	does not depend on openssl and is always available.

	Please remember, when you /ENCRYPT a target, everything you send to
	that target is encoded.  This includes /CTCP requests (/DCC offers).
	Because ircII does not do CTCP-over-CTCP, you won't be able to send
	/DCC SEND or /DCC CHAT offers to ircII users, but it will work with 
	EPIC users.

*** News 06/24/2006 -- New ON, /ON NUMERIC
	The /ON NUMERIC hook is thrown for all numerics that you do not 
	catch with a more specific /ON <number>.  For example:
		/ON 318 * {echo Numeric 318 -- $*}
		/ON NUMERIC * {echo Not Numeric 318 -- $*}
	This can be used to unilaterally replace the default format for 
	numerics that you don't otherwise catch, without having to actually
	catch all of them!  The expandos are:
		$0	The numeric being thrown
		$1	The server who sent the message
		$2-	The message

*** News 06/24/2006 -- New script, 'reconnect' from nullie
	The script causes the client to reconnect to disconnected servers after
	a certain period (/SET reconnect_time, measured in minutes), cycling
	between servers in a given servergroup. To register given servergroup
	for its use, use "/NETWORK ADD group". With conjunction with the
	autojoin scripts, allows the user to automatically rejoin channels he
	was on before a disconnect.

*** News 06/24/2006 -- New script, 'renumber' from nullie
	The script adds an alias (/RENUMBER), filling "gaps" in window numbers.
	If /SET renumber_auto is set (defaults to OFF), this will be performed
	automatically after a window is closed.

*** News 06/24/2006 -- New script, 'floodprot' from nullie
	This script buffers all outgoing data sent by the client before
	transmitting it to the remote server, to avoid "Excess Flood"
	disconnection, e.g. pasting a lot of text or sending a lot of /MODE
	commands.

	/SET floodprot_burst decides how many lines can be sent immediately
	after an idle period.

*** News 06/24/2006 -- New script, 'autoquery' from nullie
	The script creates new windows for incoming private messages, one per
	each sender.

	/SET autocreate_split_queries (defaults to OFF) decides whether windows
	will be split (ON) or hidden (OFF).

*** News 06/24/2006 -- New script, 'autojoin' from nullie
	The script maintains a list of channels to join upon connecting to a
	server. For user-defined channels that are joined after every connect,
	create an ~/.irc/channels file with one line for each channel to join 
	in the specified format:

	<channel>   <winnum>  <servergroup> [key]

	For instance:

	#epic       2         efnet
	#secretchan 3         somenet       qwerty

	If the reconnect script is loaded and given server's group is added 
	via /network add, the script will automatically rejoin channels on 
	which user was before a disconnect, remembering the keys.

*** News 06/24/2006 -- SEND_LINE does SCROLL_FORWARD in HOLD_MODE
	For those who use HOLD_MODE, if you accidentally scroll back, you
	might find yourself trapped in a place where hitting <enter> (the
	SEND_LINE keybinding) does not get you back current.  To avoid this
	unfortunate situation, if you are in HOLD_MODE, then SEND_LINE 
	keybinding (the <enter> key) will now always advance you forward a 
	page, even if you are in scrollback mode.

*** News 06/24/2006 -- Support for MAILDIR maildrops, /SET MAIL_TYPE
	If you /SET MAIL_TYPE MAILDIR, epic will check your maildir maildrop
	for email instead of your mbox maildrop.  You can /SET MAIL_TYPE MBOX
	to switch back to mbox.
	
	You must set the MAILDIR (or MAIL) environment variable to point to
	your maildrop.  For MAILDIR, your maildrop is the directory that 
	contains the 'new' and 'cur' subdirectories.

	EPIC will check your 'new' directory for new emails.  This support
	works the same way for maildir as it does for mbox:

		/on mail  (maildir)		/on mail (mbox)
		$0 -- New emails in 'new'	$0 -- New emails in 'mbox'
		$1 -- Total emails in 'new'	$1 -- Total emails in 'mbox'

	The maildir support does not look in 'cur' or anywhere else. 

*** News 06/24/2006 -- Nonblocking SSL negotiation
	Connections to ssl-enabled servers have an SSL protocol negotiation
	stage between the nonblocking connect() and irc protocol negotation.
	Traditionally, this SSL negotation has been blocking, but this can
	take a long time, during which the client is locked up.  So now 
	SSL negotiations are nonblocking.  Yay.

	Previously, the server state goes from CONNECTING to REGISTERING.
	But if it has to take a detour for nonblocking ssl negotation, it
	will go from CONNECTING to SSL_CONNECTING and then to REGISTERING.

*** News 06/24/2006 -- New script, 'nickcomp' from Blackjac

*** News 06/06/2006 -- Much improved configure support for perl/tcl/ruby
		*** VERY IMPORTANT FOR PACKAGE MAINTAINERS ***

	The code in the 'configure' script that decides how to integrate
	perl, tcl, and ruby has been rewritten.  For the most part, it 
	should automagically work.  Here's the details:

	PERL:
		--with-perl
		--with-perl=yes
		--with-perl=/usr/local/bin/perl
		--with-perl=no
		--without-perl
	By default, unless you specify --without-perl (or --with-perl=no), 
	configure will look for the 'perl' program in your path (or use the 
	perl you specify) and will interrogate the extmod package for details 
	how to compile sucessfully against perl.  If the default 'perl' binary
	is fine, you need do nothing at all.  Otherwise, all you need know 
	is what 'perl' binary you want to use.

	RUBY:
		--with-ruby
		--with-ruby=yes
		--with-ruby=/usr/local/bin/ruby
		--with-ruby=no
		--without-ruby
	Same deal as with perl.  Configure will use the 'ruby' program (or
	the one you specify) to get all the details necessary to get support
	for ruby.  For most people, you need do nothing at all.

	TCL:
		--with-tcl
		--with-tcl=yes
		--with-tcl=/usr/local/lib/tclConfig.sh
		--with-tcl=no
		--without-tcl
	TCL is a little different.  It looks for a 'tclConfig.sh' file.
	These ones are used by default (in this order)
		/usr/lib/tclConfig.sh
		/usr/local/lib/tclConfig.sh
		/usr/local/lib/tcl8.4/tclConfig.sh
	If your tclConfig.sh is not in one of these places, or if you want
	to use a specific tcl version, you must specify the path to the 
	tclConfig.h file to --with-tcl.

	Note to package maintainers:
	Please throw away all of the hacks and patches you have in place
	to support the old nasty way to set these flags.  I tried to make
	this as painless as possible.  If the user has ruby/perl installed, 
	then ruby/perl support will be included.  The hard part is tcl, since
	every system puts their tclConfig.sh in an odd place.

*** News 06/06/2006 -- /SET INPUT_ALIASES and /SET CMDCHARS deprecated
	Due to recent optimizations in the epic command parser, these two
	sets no longer have any function (although they have not yet been
	removed).  If you desparately need the use of these /set's, please
	let me know.  Otherwise someday they will vanish into the ether.

*** News 06/06/2006 -- Support for RUBY: $ruby() and /ruby
	Support for ruby (1.6.4 and 1.8.4 are tested) has been added.  The 
	support is similar to perl and tcl.  You can call out to ruby using 
	either the $ruby() function or the /ruby command.  

	The following callbacks to epic are available from within ruby:

	EPIC.echo(stringval)		Same as /echo
	EPIC.say(stringval)		Same as /xecho -b
	EPIC.cmd(stringval)		Run "stringval" without $-expansion
	EPIC.eval(stringval)		Run "stringval" with $-expansion
	EPIC.expr(stringval)		Return value of epic expression
	EPIC.call(stringval)		Return value of epic function call

	All of these functions take one argument (a string) and return one
	value (also a string).

	Some examples:
	To export an epic variable to ruby (as a String object):
		@ ruby(rubyvar='$epicvar')
	or
		ruby {rubyvar = EPIC.expr("epicvar")}

	To export a ruby variable to epic:
		@ epicvar=ruby(rubyvar)
	or
		ruby {EPIC.expr("epicvar=[#{rubyvar}]")}

	To iterate over each item in a ruby collection, passing each
	value as the argument to an epic command:
		ruby {
			something.each {|x|
				EPIC.cmd("epiccmd #{x}")
			}
		}
	(Think of using this to do database access)

	The EPIC.cmd and EPIC.eval commands do not start off a new atomic
	scope within epic.  Consequently, you can read (and write) local
	variables that are otherwise in scope:
		alias booya {
			@ :epiclocal = 5
			ruby {
				rubyvar = EPIC.expr("epiclocal").to_i
				rubyvar = rubyvar + 1
				EPIC.expr("epiclocal = '#{rubyvar}'")
			}
			echo $epiclocal
		}
	It is *IMPORTANT TO REMEMBER* that because EPIC is untyped, everything
	returned to ruby is a String object.  If you have a number in EPIC,
	it will be a String in ruby, and you need to explicitly convert it
	to an integer (as in the above example).

	It is *ALSO VERY IMPORTANT TO REMEMBER* that if you use the /ruby
	command, and you use it in a script that is loaded with the standard
	(legacy) loader, the loader will insert semicolons into your ruby code
	in the places where it would belong if it were ircII code.  Needless
	to say, THIS MAY OR MAY NOT ALWAYS BE CORRECT.  The best solution is 
	to get with the program and switch over to the pf-loader where this 
	is entirely under your control.

*** News 06/06/2006 -- New commands: /PERL and /TCL
	Support for perl and tcl has been via the $perl() and $tcl() 
	functions.  Regretably, calling a function submits the argument
	list to quoting hell, which can be particularly painful for perl.
	To remedy this somewhat, you can now call a block of perl or tcl
	code with these new commands.  The block of code must be surrounded
	by curly braces, which protect the inside from quoting hell.
	Usage:	/perl { <perl code goes here> };
	Usage:	/tcl { <tcl code goes here> };

	See the above warnings (for ruby) about using these commands in a 
	script that is loaded by the standard loader.  Semicolons may be
	a problem.  Use the PF loader instead.

*** News 06/06/2006 -- Removal of /set highlight_char
	The long-threatened removal of /set highlight_char has occurred.
	Please use the 'highlight' script for its replacement.

*** News 06/06/2006 -- New script, "chanmonitor"

*** News 06/06/2006 -- Extra support for 64 bits, if you have it.
	Most built-in-functions that take an integer argument will 
	now support 64 bit integer values.  

	Further, the following functions can output 64 bits whereas
	they never could before:
		fseek		numsort		strtol
		tobase		tobase		stat

EPIC5-0.2.0
	EPIC5-0.2.0 was released somewhere in here

EPIC5-0.0.8

*** News 12/09/2005 -- New window verb, /WINDOW INDENT (ON|OFF)
	The /window indent value permits you to tweak the /set indent
	value on a per-window basis.  However, if you do /set indent,
	it will overwrite all of your per-window indent values (I'm 
	open to discussion on this), to retain backwards compatability.

*** News 12/09/2005 -- New built in function, $xform()
	    *** OBSOLETE *** THIS INFORMATION IS OLD *** OBSOLETE ***
	This information is no longer useful.  Do not rely on it.  Please
	            check out the new documentation up above.
	    *** OBSOLETE *** THIS INFORMATION IS OLD *** OBSOLETE ***

	The $xform() function does (symmetrical) string transformations.
	What is a symmetrical transformation?  It is one in which all of
	the bits in the original string are present in the result string,
	and the original string can be recreated from the result.  This is
	distinct from the idea of hashing, which is closer to being a lossy
	compression algorithm.

	The general format of the $xform() function is:
		$xform(TYPE DIRECTION KEY text)
	whereby TYPE is one of the following:
		URL	Convert non-printable chars to %HH equivalent
			Equivalent to $urlencode() and $urldecode()
		ENC	Base16 encoding, for creating variable names
			Equivalent to $encode() and $decode()
		B64	Base64 encoding, for sending over email/http
			Equivalent to $b64encode() and $b64decode()
		NONE	No changes (just copies the string)
		CTCP	Mangle things they way CTCP ought to be
			No equivalent

	and DIRECTION is either "E" (for encoding) or "D" (for decoding),
	and KEY is ignored except for SED, where it is the cypher used in 
	the SED bit-twiddling, and DEF where it is the nickname of the 
	person who it should be transformed for

	At the time I write this, SED, CTCP, and DEF are not implemented
	yet, but they will be soon.  Watch for more info.

	More transformations are expected to be supported in the future
	(including real encryption routines).  It will probably be possible 
	to add your own at some point.

*** News 11/28/2005 -- New /ON, /ON KEYBINDING
	This on goes off whenever a keybinding is activated:
	  $0 - The keybinding that is activated
	  $1 - Length of the sequence (future expansion -- always 0 for now)
	  $2 - The ascii number of the key that activated it.

	Note that $2 only contains the last character in the bound sequence.
	In the future, $2- will change to be a word list containing all of
	the characters in the sequence.  When this change is made, $1 will
	be changed to a positive value.  But for now, $1 is always 0, and 
	$2 is always just the last character.

	If you hook this silently, you will suppress the keybinding!
	If you only want to spy on a keybinding, hook it quietly.
	The purpose for which this was requested was to be able to trap 
	everything bound to SELF_INSERT without having to rebind them.

	Examples:
	To do something every time capital-A is pressed:
	   on -keybinding "SELF_INSERT % 65" {echo You pressed A}
	To keep the user from using the SWITCH_CHANNELS keybinding.
	   on ^keybinding "SWITCH_CHANNELS % *" #

*** News 10/30/2005 -- New function, $dbmctl() [hash table support]
	*** Notice *** This function uses a custom implementation of SDBM.
	The file format it generates should be compatable with $perl() but
	is not compatable with ndbm or gdbm.

	The $dbmctl() function is an interface to the unix DBM API:
	    $dbmctl(OPEN type filename)
		Open a DBM file for read and write access.
	    $dbmctl(OPEN_READ type filename)
		Open a DBM file for read-only access.
	    $dbmctl(CLOSE refnum)
		Close a previously opened DBM file
	    $dbmctl(ADD refnum "key" data)
		Insert a new key/data pair.  Fail if key already exists.
	    $dbmctl(CHANGE refnum "key" data)
		If key already exists, change its data.  If it doesn't exist, 
		add it.
	    $dbmctl(DELETE refnum "key")
		Remove a key/data pair
	    $dbmctl(READ refnum "key")
		Return the data for a key.
	    $dbmctl(NEXT_KEY refnum start-over)
		Return the next key in the database
	    $dbmctl(ALL_KEYS refnum)
		Return all keys -- could be huge! could take a long time!
	    $dbmctl(ERROR refnum)
		Return the errno for the last error.

	"Type" must always be "STD" for now.  Reserved for future expansion.
	"Filename" must be a filename that doesn't include the ".db" extension!
		This is a requirement of the DBM api, and not an epic thing.
	"Refnum" is the integer value returned by OPEN or OPEN_READ
	"Key" is a hash table key value
	"Data" is a hash table data value
	"Start-over" is 1 if you want to fetch the first key in the table, and
		is 0 if you want to fetch the next key.  You must call this
		with 1 before you call it with 0, according to the API.
	ALL_KEYS does a "start-over" and you need to do another "start-over"
		after using it.

*** News 10/29/2005 -- New built in function, $levelctl()
	You may now add new window/lastlog/ignore/flood levels at runtime!
	Once you add a level, you can use it in your windows, and you can use
	it in /xecho -l.  Since ignore and flood control are hardcoded features
	adding a level will show up in their types, but will never be used.

	The $levelctl() function permits you to add and query levels:
	  $levelctl(LEVELS)
		Return a space-separated list of all canonical level names.
		This does not return any alias names (see below)
	  $levelctl(ADD new-name)
		Add "new-name" as a new canonical level name.  This new name
		is automagically considered part of ALL.  There is no way to
		remove levels (yet).  This returns the new level's refnum,
		which is a small integer value.  If you try to add a level
		that already exists, it is not added again, but it's existing
		refnum is returned.
	  $levelctl(ALIAS old-name new-name)
		Add "new-name" as an alias for "old-name".  An alias name is
		permitted anywhere that level names are accepted, but an alias
		name is never output by the client -- it is silently converted
		into the canonical name.  For example, "OTHER" is the canonical
		level name, and "CRAP" is its alias.  If you do 
			/WINDOW LEVEL CRAP
		It will tell you
			*** Window level is OTHER
	  $levelctl(LOOKUP level-name)
		Returns the refnum for a level-name.  The level-name can either
		be a canonical level name or an alias level name.
	  $levelctl(LOOKUP refnum)
		Returns the canonical level name for a refnum.
	  $levelctl(NORMALIZE level-name(s))
		Given a comma-and-space separated list of level names (like
		what you specify to /window level), this returns a canonical
		form which is suitable for /save'ing or displaying to the user. 
		For example,
			ALL,-CRAP
		would return
			PUBLIC MSG NOTICE WALL WALLOP OPNOTE SNOTE ACTION DCC
			CTCP INVITE JOIN NICK TOPIC PART QUIT KICK MODE USER1
			USER2 USER3 USER4 USER5 USER6 USER7 USER8 USER9 USER10

EPIC5-0.0.7

*** News 10/12/2005 -- The %D status bar expando (dcc activity) improved
	Previously, the %D status bar expando showed you "packets" of 
	activity.  A packet was 2k of data, so it looked like this:
		(      hop: 99 of 2350 (4%))
	Now this is all done in reasonable units, like so:
		(      hop: 198Kb of 4.7Mb (4.2%))
	Perhaps in the future I might remove the "b" after "Kb" and "Mb".

*** News 10/12/2005 -- New script 'highlight', removal of highlight ignores
	This is kind of experimental for now, but there is a new script 
	'highlight' which you can /load, which will implement the features
	that were previously available via "highlight ignores".  I haven't
	yet moved /set highlight_char over to the script, but I will do that
	for EPIC5-0.0.8.  If you want to help me make this script more robust,
	drop me a line or just /msg me.  If you have no idea what I'm talking
	about, don't worry, you're not missing anything.

*** News 10/12/2005 -- New built in functions $b64encode() and $b64decode()
	The $b64encode() function takes an arbitrary string and returns an
	expanded string using the Base64 encoding, which uses all of the 
	characters A-Z, a-z, 0-9, and + and /.  If you count them up, this
	is 64 distinct characters (hence the name base64), which represent
	six bits of information.  Thus, 3 bytes of data (24 bits) can be 
	transformed into 4 characters which are safe to send through data
	channels that can only handle text (such as the irc server, or 
	DCC CHAT, email, or a web server)

*** News 10/12/2005 -- Can now set both level and target with with /XECHO
	Maybe it would help if I just re-documented /XECHO's options:
	XECHO options can always be abbreviated as long as they're not 
	ambiguous.  For example, /XECHO -CURRENT can be used as /XECHO -C,
	and /XECHO -WINDOW can be used as /XECHO -W.

	 -CURRENT
		The same as -W 0  (output to current global window)
	 -LINE <number>
		Must be used in conjunction with -WINDOW.  This allows you to
		replace the <number>th line on a specified window.  This is 
		the feature of the old "scratch windows".  All windows behave
		as scratch windows these days.
	-LEVEL <window-level>
		The window levels are of course NONE, CRAP, PUBLICS, MSGS,
		NOTICES, WALLS, WALLOPS, OPNOTES, SNOTES, ACTIONS, DCCS,
		CTCPS, INVITES, JOINS, NICKS, TOPICS, PARTS, QUITS, KICKS,
		MODES, and USER1 through USER10.  If you use this option 
		without using -WINDOW, then the output will be sent to the
		window that claims the corresponding level for the current
		server.  If you use this option together with -WINDOW, then
		it will be sent to that window anyways.  This is new, because
		previously you could not tag output to a window with a level
		that didn't belong to that window.
	-VISUAL
		Output to a visible (non-hidden) window.  If window 0 (the
		global current window) is visible, it is used.  If window 0
		is hidden, then another visible window is chosen.
	-WINDOW <refnum>
		Output to a particular window, instead of the normal one.
		Remember that window refnum 0 is special, and represents the
		global current window which **may or may not** be connected
		to the current server.  If you need the current server's 
		current window, use $serverwin().
	-ALL
		Output to each and every window.
	-BANNER
		Prefix the output with the /SET BANNER value.
	-RAW
		Send the output directly to the terminal and do not attempt
		to pre-process it.  This is used to send special control 
		sequences to your terminal that EPIC does not support (such
		as sequences to change your character set)
	-NOLOG
		Do not allow the line of output to be written to the global
		logfile, a window logfile, or to a /LOG logfile.
	-SAY
		Use say() to output the line, which makes it subject to 
		output suppression rules.  An example will help:
			alias loud { xecho -b This is always displayed }
			alias noloud { xecho -b -s This is not always }

		In the first case, you could do
			loud
		or
			^loud
		and you would still see the output either way.  There is no
		way to "suppress" the output.  In the second case
			noloud
		shows the output but
			^noloud
		does not.
	-X
		("Extended" output?)  Ignore /SET DISPLAY_MANGLE for just 
		this one line, and pretend it was /SET to NORMALIZE instead.
		The most obvious use for this would be to output something in
		color even when the user has used /set display_mangle to strip
		color.
	-F
		Do not allow the line of output to trigger hidden-window-
		notification.  Normally output to a hidden window will cause
		that window's refnum to show up in the %F status bar expando.
		Using this keeps that from happening.
	--
		End of argument processing: there are no more arguments after
		this and everything else should just be output directly
			xecho -- -f <- the -f here is output!

*** News 10/05/2005 -- Automatic Scrollback rebuilding, new /WINDOW operation
	Whenever the width of a window changes (either because you're 
	swapping it in, and the screen is a different size than when you 
	swapped the window out, or because it's visible on a screen that 
	is changing size), the window's scrollback will be rebuilt.  This is
	accomplished by clearing the scrollback and then rebreaking the 
	lastlog.  If you have a large scrollback and a small lastlog, this
	may result in some data lossage.

	Although the rebuilding process attempts to keep the top of the window
	the same before and after the rebuild, it's possible for what you were
	seeing on your window to be gone after the rebuild.  If this occurs,
	the window will be set to the very top of the scrollback (ie, as far
	back as it can go).

	Generally, it is advisable from now on to have a lastlog that is 
	at least as large as your scrollback, if you want to avoid any
	chance of problems when your windows resize.

*** News 10/05/2005 -- DNS Helper (hostname lookups) now nonblocking by default
	EPIC5-0.0.6 shipped with a dns helper, but it was synchronous (it 
	did not fork off a subprocess).  It has been debugged and now it is
	turned on, and it is asynchronous.  This DNS helper is only used for
	server connections.  This is the final stage in making connections to 
	server fully nonblocking. yay!  However, forking off a child process
	to do the nonblocking dns lookup can cause issues with resource
	limitations.  (Typically you are only permitted to run a certain
	number of processes at a time, and this would count against that limit.)

*** News 10/05/2005 -- New /SET, /SET MANGLE_DISPLAY, many sets removed
	There is a new /SET, /SET MANGLE_DISPLAY, which mangles all output
	being sent to your display, naturally.  It works in the same way
	that /set mangle_inbound, mangle_outbound, mangle_logfiles works,
	and its default value is "NORMALIZE".  You should always specify
	either "NORMALIZE" or "MANGLE_ESCAPE" -- if you turn off both, then
	it will be possible for remote people on irc to send raw escape
	sequences to your display, and that is bad!

	The following /SETs have been superceded by recent changes, and have
	been removed:
		alt_charset 	blink_video 	bold_video
		color		display_ansi 	display_pc_characters
		inverse_video	underline_video

*** News 10/05/2005 -- Functions that changed because of unified mangler
	$leftpc(<COUNT> <TEXT>)
		This function performs a $stripcrap(NORMALIZE <text>) now.
	$numlines()
		This function performs a $stripcrap(NORMALIZE <text>) now.
	$printlen(<TEXT>)
		This function performs a $stripcrap(NORMALIZE <text>) now.
	$stripansicodes(<TEXT>)
		This function is the same as $stripcrap(NORMALIZE <text>)
	$stripc()
		This function is the same as $stripcrap(COLOR <text>)
	$stripcrap()
		See below for important information.

*** News 10/05/2005 -- Unified string mangler/normalizer
	Forget everything you thought you knew about the old mangler and
	normalizer (sorry!).  This is (believe it or not) much less complicated
	than before, and certainly more well documented than before!

	EPIC5 now includes a unified mangler/normalizer that is used by the
	following features:
		(The above functions)
		The input prompt	
		/lastlog -mangle
		/log mangle		
		/set mangle_inbound
		/set mangle_outbound	
		The status bar

	All characters are grouped into one of 9 "types":
	0	Normal chars 		32-127, 160-255
	1	High bit control chars	128-159
	2	Escape char 		^[
	3	Color char 		^C
	4	Highlight toggle	^B ^E ^F ^O ^V ^_
	5	Unsafe char		^M (\r)
	6	Control char		^@ ^A ^D ^H ^K ^L ^N ^P ^Q 
					^R ^T ^U ^W ^X ^Y ^Z ^\ ^] ^^
	7	Beep			^G
	8	Tab			^I
	9	Non-destructive Space	^S

	There are the 12 following mangle types:
	MANGLE_ESCAPES	NORMALIZE	STRIP_COLOR		STRIP_REVERSE
	STRIP_UNDERLINE	STRIP_BOLD	STRIP_BLINK		STRIP_ND_SPACE
	STRIP_ALT_CHAR	STRIP_ALL_OFF	STRIP_UNPRINTABLE	STRIP_OTHER

	The mangle types transform the characters, according to this table:
	-----------------------------------------------------------------------
        A = Character or sequence converted into an attribute
        M = Character mangled (ie, ^A into ^VA^V)
        S = Character stripped, sequence (if any) NOT stripped
        X = Character stripped, sequence (if any) also stripped
        T = Transformed into other (safe) chars
        - = No transformation

						Type
				0    1    2    3    4    5    6    7    8    9
	(default)               -    -    -    -    A    -    -    T    T    T
	NORMALIZE               -    -    A    A    -    X    M    -    -    -
	MANGLE_ESCAPES          -    -    S    -    -    -    -    -    -    -
	STRIP_COLOR             -    -    -    X    -    -    -    -    -    -
	STRIP_*                 -    -    -    -    X    -    -    -    -    -
	STRIP_UNPRINTABLE       -    X    S    S    X    X    X    X    -    -
	STRIP_OTHER             X    -    -    -    -    -    -    -    X    X
	(/SET ALLOW_C1)         -    X    -    -    -    -    -    -    -    -
	-----------------------------------------------------------------------

	There are only *three* ambiguous cases:
	* Type 2:
		MANGLE_ESCAPES has first priority, then NORMALIZE, and 
		finally STRIP_UNPRINTABLE
	* Type 3:
		STRIP_UNPRINTABLE has first priority, then NORMALIZE, and
		STRIP_COLOR.  You need to use both NORMALIZE and STRIP_COLOR
		to remove color changes in escape sequences
	* Type 6:
		STRIP_UNPRINTABLE has first priority over NORMALIZE.

*** News 10/05/2005 -- "ANSI" mangle type linked to "NORMALIZE"
	The mangler type "ANSI" has been renamed to "NORMALIZE".  You can 
	continue to use "ANSI", but it will be silently changed to "NORMALIZE"
	internally.

*** News 10/05/2005 -- New behavior for $mask() [jm]
	Jm needs to write me some documentation for this...

*** News 08/30/2005 -- De-support of 7-bit-only terminal/emulators
	Once upon a time, long long ago, on a planet far away, there
	used to be terminals and terminal emulators that did not know
	how to handle characters with the 8th bit set.  Thankfully,
	nobody uses these any more.  There has been rudimentary support
	in ircII clients to support this old hardware, but it is now
	unnecessary and only causes to torment people who can't figure
	out how to input their 8 bit characters!

	/SET EIGHT_BIT_CHARACTERS has been removed, and its behavior is
	now hardcoded to the previous "ON" value.  If you really need
	7-bit support, epic4 will always have it...

*** News 08/30/2005 -- Simplification of display mangling, part one
	There are 15 /set's that control how output to the display is
	prepared ("mangled").  That's about 14 too many.  The following
	/set's are at best historical curiosities and do not serve any
	modern purpose.  Their function has been eliminated.

	    BEEP_MAX	TAB	TAB_MAX		ND_SPACE_MAX

*** News 08/23/2005 -- New /window verb, /WINDOW FLUSH_SCROLLBACK
	The /WINDOW FLUSH_SCROLLBACK option, which I already regret the
	naming of, deletes all of the items in your scrollback buffer.
	Because the scrollback buffer is used to paint your window, your
	window will be cleared when you do this.

*** News 08/10/2005 -- Mangle level "ALL" does not include "UNPRINTABLE"
	This was a mistake, and "ALL" should never have included the
	UNPRINTABLE mangle level.  So if you want to do them both, then
	just do ALL,UNPRINTABLE.  Although you would never want to do
	that, since UNPRINTABLE is more encompassing than ALL.

EPIC5-0.0.6

*** News 08/08/2005 -- Updated behavior for /DCC GET
	Based on a request from larne, the following syntaxes are now
	supported for /DCC GET:

	* /DCC GET nick file1 file2 ... fileN /directory
	This will download one or more files to /directory.

	* /DCC GET nick offered-file not-offered-file
	This will do a /DCC RENAME get nick offered-file not-offered-file
	and then a /DCC GET not-offered file.  You cannot combine this with
	the previous syntax.

	* /DCC GET nick
	Download ALL files that "nick" have offered you.  You can combine
	this with the first one, (/DCC GET nick /directory) to download 
	all files offered to a single directory.

*** News 08/06/2005 -- Generalized status bar activity handling
	There are three new windowctls, and /window commands to follow
	later on.  They allow you to control a new status bar expando %E.
	This will take a bit of explanation.

	Each window has 11 "activity levels".  An "activity level" is a 
	format and a $* value.  The activity levels are numbered 0 to 10,
	and 0 is special.

	Each window has a 'current activity level' which you can set with:
		$windowctl(SET <winref> CURRENT_ACTIVITY <number>)
	When you set the current activity level to a number 1 to 10, then 
	the %E status bar expando will replace with that activity level's
	format expanded against that activity level's data.  The activity 
	level 0 is not usable, but instead stores default values that are
	used if you don't define something else for the activity level.

	To set a window's activity level format:
		$windowctl(SET <winref> ACTIVITY_FORMAT <number> <string>)
	Remember that $'s in <string> need to be doubled, in order to protect
	them from being expanded at $windowctl() time.  You do not need to
	(and you should not) put quotation marks around <string>.  If you do
	not set an activity format for a given level, then the activity format
	you set for level 0 will be used as a default.

	To set a window's activity level data:
		$windowctl(SET <winref> ACTIVITY_DATA <number> <string>)
	This is the value of $* that will be used to expand the activity 
	format each time epic redraws the status bar.  If you do not set 
	an activity data for a given level, the activity data you set for 
	level 0 will be used as a default.

	So to put this all together, if you set the current_activity value
	to 0, then the %E expando will expand to nothing.  If you set the
	current_activity level to any value 1 to 10, then it will expand to
	the "activity_format" value, which is expanded on the fly using the
	"activity_data" value as $*.  If you use a current_activity level that
	is missing an activity_format or activity_data value, the values you
	set for level 0 are used as defaults.

	Example:
		$windowctl(SET 1 ACTIVITY_FORMAT 1 ^C3$$*)
		$windowctl(SET 1 ACTIVITY_FORMAT 2 ^C4$$*)
		$windowctl(SET 1 ACTIVITY_FORMAT 3 ^C7$$notifywindows())
		$windowctl(SET 1 ACTIVITY_FORMAT 4 ^C12$$*)
		$windowctl(SET 1 ACTIVITY_DATA 0 booya)
		$windowctl(SET 1 ACTIVITY_DATA 2 hazmat)

	Now if I
		$windowctl(SET 1 CURRENT_ACTIVITY 1)
	then %E will expand to a green "booya" because the format is
		^C$$* and the value of $* is "booya" 
				(from activity_data 0)
	If I then
		$windowctl(SET 1 CURRENT_ACTIVITY 2)
	then %E will expand to a bold red "hazmat" because the format is
		^C4$$* and the value of $* is "hazmat" 
				(cause I set activity_data 2)

	If I then
		$windowctl(SET 1 CURRENT_ACTIVITY 3)
	then %E will expand to a yellow $notifywindows() that will dynamically
	update, because there are two $'s in front of it.  It will ignore the
	"booya" because it does not reference $*.

	This feature is intended to be used to create a %F workalike that
	gives you full and utter control over all aspects of how the expando
	will show up, since you get 10 levels, plus a default level, and you 
	get to control which ones use the default and which ones get their
	own custom $* and you control the formats -- you control it all.
	Have fun!

*** News 08/05/2005 -- New flag to /LASTLOG, /LASTLOG -MANGLE
	The /LASTLOG -MANGLE flag takes a mangle description (the thing that
	$stripcrap() takes) and mangles each line in the lastlog before 
	matching against your pattern.  THIS IS VERY EXPENSIVE so you should 
	not do this unless you really want to.

*** News 08/05/2005 -- *** IMPORTANT *** /EXEC -OUT changed
	/EXEC -OUT will now output to the window's current target, instead
	of to the window's current channel.  The current target is $T, and
	is the window's current query (if there is one) or the window's 
	current channel (if there isn't a query).

*** News 08/05/2005 -- Text following unmatched braces/parens/brackets
	If an unmatched brace/bracket/paren is found, about 20 chars after
	the character will be output in the error message, to help you find
	it.  Unfortunately I can't give you line numbers yet.

*** News 08/05/2005 -- New server status state, "ERROR"
	The "ERROR" status state is used in the following cases:
	* DNS lookup failed (DNS -> ERROR -> CLOSED)
	* Connect failed (CONNECTING -> ERROR -> CLOSING -> CLOSED)
	* Socket write failed (<ANY> -> ERROR -> CLOSING -> CLOSED)

*** News 08/05/2005 -- The script "altchan.bj" is renamed to "altchan" ...
	... Because we deleted the old (lame) altchan script.

*** News 08/05/2005 -- New /set, /SET OLD_MATH_PARSER
	To use the old math parser, you need to /SET OLD_MATH_PARSER ON.
	This is a promoted version of /xdebug old_math, which you should
	no longer use, it will no longer work.  Please make plans to 
	migrate away from the old math parser in epic5 scripts.

*** News 08/05/2005 -- New mangle type, "UNPRINTABLE"
	You can now mangle "UNPRINTABLE", which closely models the 
	/SET NO_CONTROL_LOG variable.  When you mangle UNPRINTABLE, all
	characters that are not printable (isgraph() || isspace()) will
	be removed, which includes all of the highlight characters.

*** News 08/05/2005 -- New script 'ison' from jm
	Jm has written a new 'ison' script which implements an ison queue,
	and acts as the backend for the 'notify' script which he has also
	revamped for this release.  This notify script will soon take over
	the notify duties for epic, so it'd be a good thing to test it 
	before the builtin notify goes away...

*** News 08/05/2005 -- New operators, === and !==
	The new math parser now supports two new operators, which do case
	sensitive string comparisons.  === returns true if the two strings
	are the same, and !== returns true if the two strings are not the
	same.

*** News 08/05/2005 -- *** IMPORTANT *** Case insensitive changes
	All case insensitive string comparisons should now be handled in
	the CASEMAPPING=ascii sense.  This means '{' and '[' are not equal,
	and | and ~ and ] and } are not equal.

*** News 08/05/2005 -- 005 CASEMAPPING value now handled (sort of)
	Some networks use CASEMAPPING=ascii, which means that the characters
	{|} are not the same as [~] as they are when CASEMAPPING=rfc1459.
	So in order to keep track of nicks correctly on the former servers,
	we now track CASEMAPPING and use it as best we can.

*** News 08/05/2005 -- Flexible on patterns can match against themselves
	When you do /ON TYPE '<pattern>', the value of $* in the pattern is
	the $* of the ON, so you can essentially do self-references.  For
	example, this now works...
		/on general_privmsg '% $0 *' {
			... you sent a message to yourself ...
		}

*** News 08/05/2005 -- Can set NOTIFY_NAME via $windowctl()
	You can $windowctl(SET <refnum> NOTIFY_NAME <stuff>) now, instead
	of having to battle with quoting hell with /window notify_name "..."

*** News 07/27/2005 -- Extended $userhost() behaviour.
	Given a channel name in place of a nick, $userhost() will return the
	unknown userhost as it did previously, but as a side effect, it will
	attempt to resolve the remaining list of nicks in the given channel
	only.  This is a somewhat esoteric feature which should only be useful
	for people who use the $serverctl() maxcache feature.

*** News 07/27/2005 -- New function $shiftbrace().
	Given a variable name, $shiftbrace() will remove the braced expression
	at the beginning of the variables value and return it.  This is a
	somewhat experimental scripting feature.

	/assign qwer {asdf zxcv} {zxcv asdf qwer}
	/eval echo $shiftbrace(qwer)  =>  asdf zxcv
	/eval echo $shiftbrace(qwer)  =>  zxcv asdf qwer

*** News 06/21/2005 -- Many scripts desupported
	Many scripts that shipped with epic4 are now unsupported in epic5.
	They will always be available at 

	ftp://ftp.epicsol.org/pub/epic/old-scripts/epic5-old-scripts.zip

	but they won't be supported for use with epic5, unless someone 
	makes a motion to re-support them.  Here are the scripts:

	alias		altchan		autokick	autoop
	away		basical		columns		dcc_spacefix
	dcc_timeout	deban		dig		dns
	edit		efnext		environment	events
	events.hop	fake-dcc	fe		fe.pf
	genalias	hybrid6		imap		ip-to-int
	ircprimer	keybinds	killpath	kpstat
	langtrans	list		ls		meta
	mkpdir		more		mudirc		netsplit.env
	newformat	nicks		old-dcc		prefix
	recursion	repeat		scandir		sdiff
	silent		sound		starutils	stat
	status_lag	tabkey.th	tabkey		tc
	time		tls		vi-binds	webster
	window

*** News 06/03/2005 -- New script, "newnick" (blackjac)
	The "newnick" script, which is loaded automatically by the 'global' 
	script [and which you need to load from your startup script if you 
	like this functionality] implements the functionality of the old 
	/SET AUTO_NEW_NICK feature, which has recently been removed.

	It exposes four /SETs (default values):
	  /SET AUTO_NEW_NICK ON
		When ON, automaticaly mangle the nickname to ensure that you
		can connect to the server without delay.
	  /SET AUTO_NEW_NICK_CHAR _
		Then character to append to your nickname in the mangling 
		process.
	  /SET AUTO_NEW_NICK_LENGTH 9
		The maximum nickname length that can be mangled.
	   /SET -AUTO_NEW_NICK_LIST
		A list of nicknames to try to use before mangling begins.

*** News 06/03/2005 -- New /TIMER flags, and execution contexts.
	[ About timer domains ]
	Each timer now belongs to one of three domains:
		1) Server timers
		2) Window timers
		3) General timers
	A "server timer" is created when you run /TIMER down-wind from
	an /ON caused by a server event.  A "window timer" is created
	when you run /TIMER at any other time.

	[ Forcing a particular type of timer ]
	You can manually force a timer to be in a particular domain
	with a flag:
		/TIMER -SERVER <refnum>		Create a server timer
		/TIMER -WINDOW <refnum>		Create a window timer
		/TIMER -GENERAL			Create a general timer.

	[ When timers go off, they change the current window and server ]
	When a timer goes off, if it is a server timer, then it will 
	switch the default server to its server, and switch the current
	window to that server's current window.  If the timer is a window
	timer, it switches the current window to its window, and switches
	the current server to that window's server.  If the timer is a 
	general timer, it leaves the current window alone, and switches
	the server to the current window's server.

	[ When timers go off and their window or server is gone ]
	So each server and window timer is "attached" to something, either
	a server refnum or a window refnum.  If that thing it is attached to
	no longer exists when the timer goes off, then the timer has to either
	turn into a general timer (and bind to the current server) or simply
	not execute at all.  If you use the -CANCELABLE flag, the timer is 
	cancelable and will not execute in this case.  The default is to be
	non-cancelable and to be treated as a general timer if the window or
	server disappears.

*** News 06/03/2005 -- The SIGUSR2 signal now supported
	If you send epic the SIGUSR2 signal (kill -USR2 pid) from an
	outside shell, EPIC will throw a "system exception" which makes
	epic give up on whatever it is doing and fully unrecurse.  This
	can be used to break out of an infinite loop.

*** News 06/03/2005 -- Several builtin /set's removed, now scripted
	The following builtin sets which no longer are used in the base
	executable have been formally removed.  Many of them have been
	implemented by the 'builtins' script.

		AUTO_NEW_NICK		AUTO_RECONNECT	
		AUTO_RECONNECT_DELAY	AUTO_REJOIN_CONNECT
		CONNECT_TIMEOUT		HELP_PAGER
		HELP_PATH		HELP_PROMPT
		HELP_WINDOW		MAX_RECONNECTS
		SWITCH_CHANNEL_ON_PART

*** News 06/02/2005 -- Your startup script is loaded before connecting
	Historically ircII has always loaded your startup script (~/.epicrc
	or ~/.ircrc) after you successfully completed your first connection 
	to a server.  This was to allow you to do irc commands in your startup
	script, such as /JOIN and /NICK and /UMODE, etc.  This has now been
	changed so your startup script is always loaded at startup before 
	epic has attempted to connect to a server.

	If you need or prefer the prior semantics, please wrap your startup
	script in a wrapper like so:

		on #^connect -1 * {
			... your old ircrc goes here ...
			@hookctl(remove -1)
		}

	What this does is tell epic that you want the body of your script 
	(the "... your old ircrc goes here ..." part) to be loaded after
	you connect to a server, and the @hookctl(remove -1) part makes sure
	that the /on deletes itself after it goes off.  This exactly effects
	the prior behavior.

	The -B command line option continues to be supported, but is ignored
	since its behavior is now the default.

*** News 06/02/2005 -- Changes to how server descriptions w/o ports get handled
	As you may or may not know, epic handles all server descriptions
	in a unified way.  Server descriptions are the things that look like:
		<hostname>:<port>:<password>:<nick>:<group>:<protocol>

	The places you can use a server description ("hostname") are:
		/SERVER -ADD <hostname>
		/SERVER -DELETE <hostname>
		/SERVER +<hostname>
		/SERVER -<hostname>
		/SERVER <hostname>
		/DISCONNECT <hostname>
		$serverctl(REFNUM <hostname>)
		/XEVAL -S <hostname>
		/XQUOTE -S <hostname>
		/MSG -<hostname>/target
		$winchan(#chan <hostname>)
		/LOG SERVER <hostname>
		/WINDOW SERVER <hostname>
	(Every other place in epic requires a server refnum)

	This change makes the following backwards-incompatable change:
	When you add the first server for a given <hostname>, the port you
	use on that server becomes the default port for that <hostname> in 
	all other places in epic.  This is true even if you later add other 
	servers for <hostname> using other ports!

	Example 1:
			/server -add irc.host.com:6666
	adds a new server, we'll call server refnum 0.  So now any place you 
	use "irc.host.com" (see above) will default to port 6666, and thus 
	"irc.host.com" will refer to server refnum 0, rather than being an 
	error.  Furthermore, let's say you do:
			/server -add irc.host.com:6667
	and this becomes server refnum 1.  The "irc.host.com" continues to 
	refer to server 0, even though this server uses port 6667.

	Example 2:
			/server -add irc.host.com:6666
	adds a new server, server refnum 0.  Then later on you do 
			/server irc.host.com
	This will connect to server refnum 0, on port 6666, and WILL NOT 
	create a new server on port 6667 and connect on port 6667!

*** News 06/02/2005 -- /QUIT with no arguments uses per-server quit messages
	If you do /QUIT with no arguments, then any per-server quit message
	that you have previously set with $serverctl() will be used, and will
	not be clobbered by /SET QUIT_MESSAGE.  Thus, /SET QUIT_MESSAGE is 
	only used as a fallback for servers that do not have their own quit 
	message.

*** News 06/02/2005 -- Preliminary support for nonblocking dns for server
	DNS lookups for server connections are now done in a nonblocking way.  
	This code has not been tested very much so it may be unstable.  If you
	try to break it, you probably will.  Let me know how you did it.

	Nonblocking DNS lookups require the use of a "dns helper" process,
	which is fork()ed for every lookup.  Since you only have a limited 
	number of processes available to you, this means you shouldn't try
	to connect to an unreasonable number of servers concurrently.  If you
	do manage to do that, you may receive errors.

	It should be safe to use /server to cancel a very long nonblocking
	dns request.

*** News 06/02/2005 -- /WINDOW SERVER shows server list instead of error
	Previously /window server without an argument displayed an error
	which isn't terribly friendly.  So now it will display the server
	list, as if you had done /server without any arguments.

*** News 05/10/2005 -- Nickname rejections handled differently now
	When the server rejects your NICK change request after you have
	already connected and registered with the server, no further
	action will take place because there's nothing that can be done
	to stop a nick collision, and nothing that needs to be done if 
	your new nick is taken.

	If the server rejects your NICK during the registration process,
	epic will no longer "fudge" your nickname by adding _'s and/or
	rotating your nickname, and will no longer "reset" your nickname
	by prompting you for a new one and stopping the world until you
	give it a new one.  Rather, EPIC will throw /on new_nickname and
	if you don't use it, will warn you that you need to use /NICK
	to choose a new nickname (and /server +<ref> to reconnect if you
	ping out)

	This also means that /SET AUTO_NEW_NICK is gone (soon to be replaced
	by a script, which will implement the old hardcoded behavior), and
	handling nickname change failures is fully under control of script!

*** News 05/10/2005 -- New window verb, /WINDOW KILLABLE (default on)
	When you /WINDOW KILLABLE OFF, the following changes occur:
	  * /WINDOW KILL will fail[1] with an error message.
	  * /WINDOW KILL_ALL_HIDDEN will not kill the window.
	  * /WINDOW KILL_OTHERS will not kill the window.
	  * /WINDOW KILLSWAP will fail[1] with an error message.
	[1] A failure of *any* /window operation causes any subsequent 
	    operations in the same statement to be discarded, to avoid
	    accidentally operating on the wrong window.  This behavior
	    is not a special-case.

*** News 05/07/2005 -- New $windowctl(SET <ref> TOPLINE <num> <stuff>)
	You can now set a topline literally (without wrangling with the
	/window command) this way.  The <stuff> is not subject to any
	sort of double-quoting shenanigans.  BTW, don't get too excited,
	I haven't implemented any other $windowctl(SET *) operations,
	but this one needed to be done right away.

*** News 05/02/2005 -- New $serverctl(), $serverctl(GET <ref> LOCALPORT)
	This returns the port used by the local side of a connection to 
	the server.  If the server is not connected, it returns 0.

*** News 04/28/2005 -- *** IMPORTANT *** /ON LEAVE changed to /ON PART
	This is a hard cut-over breaking of backwards compatability.
	There is no way to "go back", but there are means you can take 
	to compensate:
		@type = (info(i) < 1224) ? [LEAVE] : [PART]
		on $type * { ..... }
	This is the last vestige of the senseless substitution of the
	non-irc word "LEAVE" for the irc word "PART".  Good riddens to it
	and may it never return again.

*** News 04/25/2005 -- /SET -CREATE removed (use /ADDSET)
	The /set -create command which was first marked for deletion 6 months
	ago has now been removed.  You can use the /addset alias provided by
	the "builtins" script to get the same functionality:
			/ADDSET <name> <type> {
				... callback ...
			}
	Don't forget to /SET <name> <value> after you do an /addset to give 
	it an initial value!

*** News 04/25/2005 -- *** IMPORTANT *** NEW MATH PARSER NOW DEFAULT
	The new math parser, which first appeared in 1998 in EPIC4pre2.001,
	and has been stable since early 1999, has now finally been made the
	default math parser.  IT IS POSSIBLE THAT VERY BADLY WRITTEN SCRIPTS
	OR SCRIPTS THAT DEPEND ON BUGS IN THE OLD MATH PARSER MAY BREAK AS A
	RESULT OF THIS CHANGE!  

	If you need to continue using the old math parser, you can turn it 
	back on with:
			/xdebug old_math
	The old math parser won't be going away any time soon, but it will
	never again be the default math parser (barring rioting crowds with
	effigies and pitchforks at my door...)

	The operation /xdebug new_math is now a no-op, and doing it will have
	no effect.  YOU CANNOT TURN OFF NEW MATH BY DOING /XDEBUG -NEW_MATH 
	ANY MORE.  YOU MUST DO THE ABOVE /xdebug old_math.

EPIC5-0.0.5

*** News 04/18/2005 -- New $windowctl() stuff
	$windowctl(GET <win> DISPLAY_SIZE)	to replace $winsize()
	$windowctl(GET <win> SCREEN)		to replace $winscreen()
	$windowctl(GET <win> LINE <num>)	to replace $winline()

*** News 04/18/2005 -- /SET -OLD_SERVER_LASTLOG_LEVEL turns that feature off.
	If you /set -old_server_lastlog_level, then the window's level is not
	changed when it is connected to an existing server.  This is probably
	a bad idea as it can lead to level duplication, the effects of which
	are poorly-defined

*** News 04/18/2005 -- /SET -NEW_SERVER_LASTLOG_LEVEL turns that feature off.
	If you /set -new_server_lastlog_level, then the window's level is not
	changed when it is connected to a new server.  This allows you to 
	completely opt-out of this feature. 
	
*** News 04/18/2005 -- New $serverctl() stuff
	So $serverctl(LAST_SERVER) and $serverctl(FROM_SERVER) have been
	added and are used by the builtins script

*** News 04/18/2005 -- Translations now fixed
	I apologize to you RUSSIAN_WIN translation users for it being 
	broken.  This release fixes it up, and the fix was back-ported 
	to EPIC4.

*** News 04/18/2005 -- /SET STATUS_* subformats no longer size limited
	Previously, the effective size that /SET STATUS_* values you could
	expand to (especially if you have /SET STATUS_DOES_EXPANDOS ON)
	as limited on a per-expando basis.  This was inconveniently small 
	for some people who were using lots of characters doing color 
	markup, etc.  So all of those limits have been removed.

*** News 04/18/2005 -- "ERROR --" changed to "INFO --", not used for exec, dcc
	The "ERROR --" io messages were alarming people so they were changed
	to "INFO --" in this release.  Further, EXEC and DCC connections do 
	not output these diagnostic messages.

*** News 04/18/2005 -- /XECHO -v outputs to current window if it's visible
	/XECHO -v has traditionally output to the first visible window (the
	first window on the first screen), but it was agreed that it is more
	reasonable to output to the current window if that is visible.

*** News 04/18/2005 -- Ways to keep dcc xfers from swamping your cpu...
	There are three new ways to keep high speed dcc xfers from swamping
	your cpu with unnecessary status bar redraws:

	$dccctl(UPDATES_STATUS [0|1])
	   Turns off (on) whether or not the %D status bar value should be
	   updated whenever any activity occurs.  If you turn this off, then
	   %D is frozen until you turn it back on.  The argument is optional,
	   and if you don't supply it, it returns the current value.  If you
	   do supply it, the old value is returned.

	$dccctl([SET|GET] <refnum> UPDATES_STATUS [0|1])
	   Turns off (on) whether or not a particular dcc should update the 
	   %D status bar expando when it has activity.

	/ON ^DCC_ACTIVITY * #
	   Hooking (and suppressing) /ON DCC_ACTIVITY will prevent the status
	   bar from being redrawn as a result of this activity.  This won't
	   stop %D from being updated, but it does stop the status bar from
	   thrasing uncontrollably.
 
*** News 04/18/2005 -- The "builtins" script (Blackjac)
		*** IMPORTANT *** IMPORTANT *** IMPORTANT ***
	As noted below in "notes for forward compatability", there is a new
	script called "builtins" which is loaded by "global" which is not 
	loaded by default if you have a startup script.  This means it is
	really important that you add either /LOAD global or /LOAD builtins 
	to your ~/.ircrc (or ~/.epicrc) file, or you will notice that the 
	following features have "disappeared":
		*** IMPORTANT *** IMPORTANT *** IMPORTANT ***

	Commands:
		BYE, DATE, EXIT, HOST, IRCHOST, IRCNAME, LEAVE, REALNAME,
		SAVE, SIGNOFF, WHOWAS
	Functions:
		LASTSERVER, SERVERGROUP, SERVERNAME, SERVERNICK, SERVERNUM,
		SERVEROURNAME, SERVERTYPE, WINBOUND, WINCURSORLINE, WINLEVEL,
		WINLINE, WINNAM, WINNICKLIST, WINNUM, WINQUERY, WINREFS,
		WINSCREEN, WINSCROLLBACKSIZE, WINSERV, WINSIZE, WINSTATSIZE,
		WINVISIBLE
	Sets:
		AUTO_REJOIN, AUTO_REJOIN_DELAY, AUTO_UNMARK_AWAY, AUTO_WHOWAS,
		BEEP_ON_MSG, COMMAND_MODE, DCC_TIMEOUT, FULL_STATUS_LINE,
		NUM_OF_WHOWAS, REVERSE_STATUS_LINE, SHOW_END_OF_MSGS,
		SHOW_WHO_HOPCOUNT, VERBOSE_CTCP

	Furthermore, the following new commands have been added:
		ADDSET, DELSET
	which allow you to add a scripted built in /set variable (which
	is used extensively by this script).

		*** IMPORTANT *** IMPORTANT *** IMPORTANT ***
	THESE FEATURES NO LONGER EXIST AS HARDCODED EPIC5 FEATURES AND ARE 
	SOLELY IMPLEMENTED AS SCRIPT FEATURES VIA "builtins"!  IT IS NOT A
	BUG THAT THEY ARE "MISSING".  The way to fix this is to /load global 
	or /load builtins in your startup script.
		*** IMPORTANT *** IMPORTANT *** IMPORTANT ***

*** News 04/18/2005 -- New script, "loadformats" (fudd)
	This script presents the usable interface to implied on hooks.
	See the explanation for implied on hooks below.

	/ADDFORMAT <type> [value]
	   This binds the implied on hook for /ON <type> to a new builtin set
	   that is called /SET FORMAT_<type>.  Remember that implied on hooks 
	   are only used if you don't hook serial number zero!

	/DELFORMAT <type>
	   This removes the implied on hook for /ON <type> and removes the
	   builtin set /SET FORMAT_<type>.

	/DUMPFORMATS
	   This does a /DELFORMAT for all possible values of <type>.  You
	   will not have any implied on hooks after this is done.

	/LOADFORMAT <filename>
	   Given a file that contains lines that look like this:
		<TYPE> <VALUE>
	   Run /ADDFORMAT <TYPE> <VALUE> for each line.

	/SAVEFORMATS <filename>
	  Save all implied hooks to a file so it can be loaded later with
	  /LOADFORMAT.

*** News 03/29/2005 -- Just a note about mangling of /on patterns...
	In EPIC5, your /ON patterns are no longer "filled out" to the expected
	number of words in $*.  An example:
		EPIC4:	/ON PUBLIC "hop" 	becomes	
			/ON PUBLIC "hop % *"
		EPIC5:	/ON PUBLIC "hop"	stays as it is.
	Note that the EPIC5 version *WILL* *NOT* *EVER* *MATCH* any real
	/on public's because it doesn't match any valid $* values.

	Why was this change made?  Because /on is a general purpose function,
	there were problems where some /on's sometimes did not always have
	the minimum number of words in $* that they were supposed to, and it
	would lead to spaces in odd places, preventing matches.

	It was decided that we would leave "filling it out" as a task that 
	the scripter would need to be responsible for.  In the above case,
	you can simple solve the problem with:

		EPIC5:	/ON PUBLIC "hop *"

*** News 03/20/2005 -- History recall moved into a script (Blackjac)
		*** IMPORTANT *** IMPORTANT *** IMPORTANT ***
	History recall (BACKWARD_HISTORY and FORWARD_HISTORY) are now
	scripted features, implemented by 
			/load history
	instead of hardcoded into the client.  If you want to be able to
	use cursor up and down to recall input history you need to add
			/load history
	to your ~/.epicrc or ~/.ircrc file.  Remember, if you do not put 
			/load history
	in your startup file, then you will find that input history recall
	will no longer be available.  So just put 
			/load history
	in your startup file, and everything will be cool!

EPIC5-0.0.4

*** News 03/14/2005 -- New option to $line(), $line(<num> -TIME)
	You can now provide the "-TIME" option to the $line() function
	which returns the timestamp when that line was added to the lastlog.
	You can use -TIME together with -LEVEL, and if you do, the level is
	always the second-to-last word, and the timestamp is the last word.
	You can pass the timestamp to $strftime() for conversion.

	Example:
		$line(<num> -LEVEL -TIME)
	might return
		"Some line goes here CRAP 1110000000"

*** News 03/11/2005 -- Implied ON hooks ($hookctl(SET LIST <type> IMPLIED <str>)
	A lot of people create a large number of on hooks that look like:
		ON ^TYPE * { echo $cparse($format[type]) }
	for each TYPE.  Since all of these ONs are otherwise identical,
	they clutter up the /on list, and managing all of the variables
	is a hassle.  So there is now a feature to allow you to set an
	"implied" on hook that does nothing more than the above.

	You can now set a format with 
		$hookctl(SET LIST <type> IMPLIED <string>)
	and if you don't have an /ON <type> *, then epic will do
		echo $cparse(<string>)
	and suppress the normal output.

	You may surround <string> with {}s if you wish, to avoid quoting
	hell.  Match sure to keep your {}s matched up if you do so.
	See the "loadformats" info above for how to practically use this.

	You can compile out this feature if it offends you by #undef'ing
	IMPLIED_ON_HOOKS in config.h

*** News 03/03/2005 -- New status bar expandos, %{2}+ and %{3}+
	These two status exapndos act like %+ and %{1}+ respectively,
	except that %{2}+ and %{3}+ will contain only the mode string,
	and not any subsequent arguments.  Specifically, %{2}+ and %{3}+
	don't include the channel's key or user limit (if any).

*** News 03/03/2005 -- New configure options, --with-multiplex
	You can choose which multiplex function you want epic to use with
	a new configure option, "--with-multiplex".  These are the values:
			--with-multiplex=select
			--with-multiplex=poll
			--with-multiplex=freebsd-kqueue
			--with-multiplex=pthread
	If you don't choose a multiplexer, or you choose one that isn't
	supported by your system, it will use select instead.  Note that
	the kqueue() support is only tested on freebsd-4 and freebsd-5.

	If you use --with-multiplex=pthread and --with-ssl together,
	and your openssl version was not compiled to be thread-safe, then
	you won't be able to use ssl with pthreads (it would crash if 
	you tried).

*** News 03/03/2005 -- New $hookctl() operation, "GET HOOK <ref> STRING"
	The $hookctl(GET HOOK <ref> STRING) operation returns a string that
	is suitable for passing to /eval {....} to recreate the hook.  This
	will be used by scripts that want to "/save" an /on.  If you write
	these values to a file, you will be able to /load it later.

*** News 03/03/2005 -- Changes to how error messages are displayed
	Traditionally, ircII clients have tried to emit one error message
	for each error event, putting all of the information into one line.
	This has its advantages and disadvantages.  The main disadvantage is
	that if information is difficult to pass around, it is usually 
	discarded, and the user may be given a vague or useless error message.

	So now error reporting has been revamped in epic.  Whenever an error
	is actually generated (or first noticed), the information that is 
	available will be displayed to the screen.  Then a failure condition
	will be sent up to the higher levels of epic, and each higher level	
	will give you whatever information it has available.  Thus, while an
	error message may take up 3 or 4 lines of display, each line has a 
	little piece of information the others don't, and taken together, you
	get the full picture of what happened.

	Each error line looks like this:
			*** ERROR -- <message>
	where "***" is the value of /SET BANNER.  These errors are yells, 
	so if you want to hide them, use /ON YELL.


*** News 02/28/2005 -- !!! IMPORTANT !!! "global" SCRIPT NO LONGER AUTO-LOADED
		This is a backwards incompatable change!

	A lot of people who run script packs find that the "global" script 
	which is automatically and unconditionally loaded by epic at startup
	is a nuisance which must be eliminated.  After much deliberation, it
	seems agreeable to most parties that *by default*, the "global" script
	will only be loaded for those people who do not have their own startup
	script (~/.ircrc or ~/.epicrc)

	If you *have* your own startup script, and you *like* having the global
	script loaded at startup then you need to make a one-time change to 
	the top of your startup script:
			load global
	which will give you old backwards-comaptable behavior.  If you do not
	make this change, then some /on's, /alias's, and other things that
	you may have had access to will disappear until you make this change.

	For those of you who do not have your own startup script, you will not
	notice any changes.

	For those of you who have wished the 'global' script to die a horrible
	fiery death, you have got your wish.

*** News 02/28/2005 -- New built in command, /SUBPACKAGE
	The /SUBPACKAGE command is probably temporary, so don't get too
	attached to it yet.  This command should be used in scripts that
	are loaded from other scripts, who have used the /PACKAGE command,
	to set up a "nested" package name.  For example:

	File "one":
	  /PACKAGE one
	  /LOAD two
	  /ALIAS frobnitz {echo hello!}

	File "two"
	  /SUBPACKAGE two
	  /ALIAS oofda {echo hi!}

	In this example, "frobnitz" will be in the package "one" and 
	"oofda" will be in the package "one::two".

*** News 02/28/2005 -- The invite command changed here
	This was never documented for some reason until June 2007.

	You can now invite people more flexibly:
		INVITE #chan nick1 nick2 nick3
	or	INVITE nick #chan1 #chan2 #chan3
	or	INVITE nick

	and it will just figure out the right thing to do.

*** News 02/04/2005 -- Clarification on /EXEC process output.
	Until this time, partial lines of output from /EXEC processes
	were silently ignored, could not be redirected, and callbacks
	and /on's were not executed in the correct window.  Now the
	partial-line and full-line handler for /exec processes has been
	unified, and they both work like this:

	1) Set up the correct server and window
	   Note: Yes, I am aware this may need further refactoring.
	2) Increment the line output
	   Note: Partial lines now count as a line of output for the
		 -LIMIT counter.
	3) If you used -MSG or -NOTICE or -OUT, send the the text to 
	   the target.
	   Note: This usually hooks an /ON SEND_* hook (see below)
	4) If you used -LINE or -LINEPART or -ERROR or -ERRORPART, call
	   the appropriate callback
	-or-
	4a) If you didn't use them, hook the appropriate /ON.
	4b) If you didn't hook the /ON, and you aren't redirecting the
	    output to another target, output it to the screen.

	Note that (4b) and (3) work in complement to each other.  If you do
		/EXEC -MSG nick w
	then the output will be handled through /ON SEND_MSG (either from
	your own on, or from the default output).  If you do not, then the 
	output will be handled by /ON EXEC. 

*** News 02/02/2005 -- New $serverctl() attribute, "PROTOCOL"
	$serverctl(GET <refnum> PROTOCOL) now either returns "IRC" or 
	"IRC-SSL" depending on whether you're doing a regular irc, or an 
	ssl-enabled irc connection, eh!

*** News 02/02/2005 -- Automatic creation of ALTNAME for servers, %S changes.
	When you create a new server (with /server or /server -add or 
	/window server), the server will automatically be given its first
	"altname" (see "Alternate server names" below).  

	Each server's first altname is constructed as:
	* If the server name is an IP address, the whole name, 
        * Otherwise, the first segment that does not start with "irc".

	This is the value you have historically seen at %S on your status
	bar.  This value is now automatically created as your server's
	first altname, and you can change it!

		@serverctl(SET <refnum> ALTNAMES <whatever>)

	%S, and %{1}S have been changed to use the "first altname" for 
	the server, whatever you choose that to be.

	Please note:  If you clear your server's altname list, then %S and
	%{1}S will have no choice but to show the server's full name.  This
	is intentional and not a bug.  %{2}S will continue to show your
	server's full name all the time.

*** News 01/27/2005 -- New status expando, %{3}S
	%{3}S is just like %{2}S, but instead it shows the server's group
	instead of the server's name.

*** News 01/26/2005 -- Notes for forward compatability
			*** /SET -CREATE DEPRECATED ***
	EPIC5-0.0.3 includes /SET -CREATE, but this is deprecated, and will
	be removed by EPIC5-0.0.5.  It is suggested that you start using
	$symbolctl() [see below] in scripts that support EPIC5-0.0.3 and up.

			*** THINGS ARE GOING TO BE CHANGING ***
	Starting before EPIC5-0.0.5, there will be a signficant shift to try
	to script as many /set's and builtin commands as possible.  This
	stuff is all intended to be put into the 'builtins' script which is
	loaded from 'global' (ie, it's automatically loaded).  Some of you do
	not load 'global' and so by EPIC5-0.0.5, if you do not load builtins,
	you will find a lot of features gone.  I won't consider this a "bug",
	but the natural progression of EPIC5's development...

	Builtin commands that will be moved to 'builtins'
		/BYE		/DATE		/EXIT
		/HOST		/IRCHOST	/IRCNAME
		/LEAVE		/REALNAME
		/SIGNOFF

	Builtin functions that will be moved to 'builtins'
		$serverhost()

	Builtin SETs that will be moved to 'builtins'
		AUTO_REJOIN	AUTO_REJOIN_DELAY	AUTO_UNMARK_AWAY
		AUTO_WHOWAS	NUM_OF_WHOWAS		BEEP_ON_MSG
		COMMAND_MODE	FULL_STATUS_LINE	REVERSE_STATUS_LINE
		SHOW_CHANNEL_NAMES		SHOW_END_OF_MSGS
		SHOW_WHO_HOPCOUNT		VERBOSE_CTCP

EPIC5-0.0.3

*** News 01/25/2005 -- New built in function, $symbolctl()
	Here's the plan.  An all-encompasing low-level symbol manipulation 
	thingee.  This interface is not intended to replace $aliasctl(), but 
	rather to supplement it.  No, $aliasctl() will never be removed.  
	Yes, much of its functionality (but not all) is duplicated here.

	$symbolctl(TYPES)
	     Return all of the types supported in this version of EPIC:
	     ALIAS		ASSIGN			BUILTIN_COMMAND
	     BUILTIN_FUNCTION	BUILTIN_EXPANDO		BUILTIN_VARIABLE

	$symbolctl(PMATCH <type> <pattern>)
	     Return all symbols of type <type> that match <pattern>.  You can 
	     use the special value "*" for <type> to get symbols of all types.

	$symbolctl(CREATE <symbol>)
	     Ensure that <symbol> exists in the global symbol table.  When 
	     symbols are first created, they do not contain any actual values, 
	     but rather act as a placeholder in case you want to set any.  
	     You must ensure that a symbol exists before you try to change 
	     its values.  CREATEing a symbol that already exists is harmless; 
	     feel free to do it.

	$symbolctl(DELETE <symbol>)
	$symbolctl(DELETE <symbol> <type>)
	     Delete all of the values of a particular symbol, or just one type.
	     Example: $symbolctl(DELETE booya ALIAS) is the same as 
				/alias -booya
	     Warning: You can delete built in variables/functions/etc with this!
		      There's no way to restore them back if you do!  Caution!

	$symbolctl(CHECK <symbol>)
	     Inspects <symbol> to see if it has any values left.  If there are 
	     no values left for <symbol>, it is removed from the global symbol 
	     table.  You must then CREATE it again if you want to use it later.

			*** IMPORTANT NOTE ABOUT "LEVELS" ****
	In order to "get" or "set" a symbol's values, the symbol needs to exist.
	If you try to "get" or "set" a symbol that doesn't exist, $symbolctl() 
	will return the empty string to tell you that it failed.  You need to 
	use the CREATE operation above to bootstrap a new symbol before using 
	it.

	Now, /STACK PUSH and /STACK POP work by manipulating "levels" in the 
	symbol table.  By rule, <level> == 1 always refers to the "current" 
	value of a symbol.  If you do /STACK PUSH, then the value you pushed 
	will be copied to <level> == 2.  If you /STACK PUSH something else, 
	that values moves to <level> == 3.  So what you can do is use 
	"GET x LEVELS" to find out how many levels a symbol has, and then use 
	"GET x <num>" to find out if there is a symbol type that interest you 
	at that level.  IN THIS WAY you can directly manipulate the /stack push
	values without having to actually use the /stack command.

	In general, <level> is always 1 for everything you want to do, unless 
	you are intentionally monkeying around with your /stack values.

		 *** NOW BACK TO YOUR REGULARLY SCHEDULED HELP ***
	$symbolctl(GET <symbol> LEVELS)
	     Return the number of levels of <symbol> that are /STACKed.  This
	     value is always 1 unless you have /STACK PUSHed something.
	$symbolctl(GET <symbol> <level>)
	     Return all of the <type>s that are defined for <symbol> at <level>.
	     If <level> is 1, it gets the current value(s).  If <level> is > 1,
	     it starts looking at the /STACK PUSHed values.

	$symbolctl(GET <symbol> <level> ALIAS VALUE)
	$symbolctl(GET <symbol> <level> ALIAS STUB)
	$symbolctl(GET <symbol> <level> ALIAS PACKAGE)
	$symbolctl(GET <symbol> <level> ALIAS ARGLIST)
	     Retrieve one of the values for one of your aliases

	$symbolctl(GET <symbol> <level> ASSIGN VALUE)
	$symbolctl(GET <symbol> <level> ASSIGN STUB)
	$symbolctl(GET <symbol> <level> ASSIGN PACKAGE)
	     Retrieve one of the values for one of your assigns

	$symbolctl(GET <symbol> <level> BUILTIN_COMMAND)
	$symbolctl(GET <symbol> <level> BUILTIN_FUNCTION)
	$symbolctl(GET <symbol> <level> BUILTIN_EXPANDO)
	     Returns 0 if these types are not in use (ie, if there is not a 
	     built in command), and returns non-zero if there is.  If you're 
	     smart, you won't try to do anything with the non-zero value.

	$symbolctl(GET <symbol> <level> BUILTIN_VARIABLE TYPE)
	$symbolctl(GET <symbol> <level> BUILTIN_VARIABLE DATA)
	$symbolctl(GET <symbol> <level> BUILTIN_VARIABLE BUILTIN)
	$symbolctl(GET <symbol> <level> BUILTIN_VARIABLE SCRIPT)
	$symbolctl(GET <symbol> <level> BUILTIN_VARIABLE FLAGS)
	     Retrieve information about a /SET.  
	     The "TYPE" is one of "STR", "INT", "BOOL", or "CHAR"
	     Generally, either "BUILTIN" or "SCRIPT" is set, but not both.

	$symbolctl(SET <symbol> <level> ALIAS VALUE <string>)
	$symbolctl(SET <symbol> <level> ALIAS STUB <string>)
	$symbolctl(SET <symbol> <level> ALIAS PACKAGE <string>)
	$symbolctl(SET <symbol> <level> ALIAS ARGLIST <string>)
	     Change one of the values for one of your aliases.  If you omit
	     the <string>, it will clear the value.

	$symbolctl(SET <symbol> <level> ASSIGN VALUE <string>)
	$symbolctl(SET <symbol> <level> ASSIGN STUB <string>)
	$symbolctl(SET <symbol> <level> ASSIGN PACKAGE <string>)
	     Change one of the values for one of your assigns.  If you omit
	     the <string>, it will clear the value.

	$symbolctl(SET <symbol> <level> BUILTIN_VARIABLE)
	     Create a new user-created /SET with default values (type == BOOL,
	     data == OFF, script is <empty>.)
	$symbolctl(SET <symbol> <level> BUILTIN_VARIABLE TYPE <set-type>)
	$symbolctl(SET <symbol> <level> BUILTIN_VARIABLE DATA <string>)
	$symbolctl(SET <symbol> <level> BUILTIN_VARIABLE BUILTIN)
	$symbolctl(SET <symbol> <level> BUILTIN_VARIABLE SCRIPT <code>)
	$symbolctl(SET <symbol> <level> BUILTIN_VARIABLE FLAGS)
	     Change one of the values for one of your /set's.  You cannot 
	     change values for system /set's, sorry.  Setting the TYPE value 
	     changes the DATA value to a default (<empty> for strings, 0 for 
	     everything else) so always set DATA after setting TYPE. 
	     Yes, you can change the TYPE of a /set after you create it!
	     It's probably a bad idea to set FLAGS for the present.

*** News 01/13/2005 -- New $logctl() feature, $logctl(CURRENT)
	$logctl(CURRENT) can return one of these values:

		-1	Nothing is being logged right now
		 0	Something is being logged to the global log, or to
			a window log
		>0	Something is being logged to the given log refnum.
			You can use this log refnum with $logctl().

*** News 01/11/2005 -- New /ON, /ON NEW_NICKNAME
	This allows you to create your own nickname mangler whenever EPIC
	is resetting your nickname to register to a server.  Remember that
	your nickname is (usually) not reset if /SET AUTO_NEW_NICK is ON, so
	you need to /SET that to OFF before this can be used.

	Each time that your nickname is "reset" (see next entry), the
	/ON NEW_NICKNAME hook will be thrown.
		$0 - Server whose nickname is being reset
		$1 - Your current nickname ("*" if you're unregistered)
		$2 - The nickname you tried to change to ("*" if none)
	You should somehow use this information to generate a new nickname
	and then do a /NICK <newnick> operation.  You should avoid trying
	to do anything else, particularly asynchronous things like /wait,
	because the status of the connection is unknown when this /ON is
	thrown (ie, you may not be connected).  Just stick to /NICK.

	If you do not hook the /ON, or you do not do a /NICK within the
	/ON, then epic will prompt you for a new nickname in the way it
	has always done (so this change is opt-in and backwards compatable).

*** News 01/11/2005 -- Changes to how nick change errors are handled
	Until this point, EPIC sometimes (rather aggressively) forced the
	user to provide a new nickname, when it wasn't necessary.  This 
	was called "resetting" the nickname.  The variable /SET AUTO_NEW_NICK
	existed to try to avoid needlessly annoying the user with requests
	for nicknames when a new one could be whipped up.

	Well, anyways, this has all been refactored.  If you are already 
	connected to the server, epic will no longer "reset" your nickname 
	when it recieves one of the many numerics that indicate that your 
	nickname is not acceptable.  "Resets" will only occur when you are
	unregistered (when you do not have a nickname yet).

	As with before, if you have /SET AUTO_NEW_NICK ON, then your nickname
	will never be "reset" unless the nickname mangler is unable to come
	up with an alternate nickname for you.

*** News 01/11/2005 -- Change to how /set new_server_lastlog_level works
	In EPIC4, when you do /window server, the window that is changing
	server has its window level changed to /SET NEW_SERVER_LASTLOG_LEVEL.
	This can result in two windows with level ALL if the new server can't
	be connected to.  Starting with now, this change will not occur until
	after we have registered with the server (when we get the 001 numeric).

	This means if you do /window new server foo.com, that the new window
	will be level NONE until the connection to "foo.com" is successful.
	Furthermore, only the "current window" for "foo.com" will be changed.
	If you move multiple windows to a new server at the same time, only
	one of them is the "current window" and only that window ever has
	its window levels changed.  The others stay at "NONE".

*** News 01/06/2005 -- Can now set ipv4/ipv6 vhosts separately
	The /hostname (/irchost) and -H command line option may now take 
	two hostnames, separated by a slash.  Example:

	/hostname foo.bar.com		Use "foo.bar.com" as the vhost for 
					both ipv4 and ipv6
	/hostname foo.bar.com/faz.bar.com
					Use "foo.bar.com" as the vhost for
					ipv4, and "faz.bar.com" as the vhost
					for ipv6.
	/hostname foo.bar.com/		Use "foo.bar.com" as the vhost for 
					ipv4 only -- don't change the ipv6
					vhost
	/hostname /foo.bar.com		Use "foo.bar.com" as the vhost for
					ipv6 only -- don't change the ipv4
					vhost

*** News 01/06/2005 -- New serverctl attr: $serverctl(GET <num> ADDRFAMILY)
	The $serverctl(GET <num> ADDRFAMILY) value returns either
		ipv4		if the server connection is ipv4
		ipv6		if the server connection is ipv6
		unix		if the server connection is a filename

*** News 01/06/2005 -- Old (undocumented) function: $servports()
	There has been a function $servports() around for a very long time,
	but it's never been documented.  It returns two values, the first
	one is the remote port (what you connected to the server with), and
	the second one is the local port (which is not always useful).

*** News 01/01/2005 -- Argument lists for hooks, and $fix_arglist() (howl)
	$fix_arglist(arglist) returns how argument list arglist will be parsed
	by epic.

	It is now possible to supply argument lists to hooks, just like
	one would do for aliases:
	/on hook "*" (a,b,...) {echo $a $b $*;};

	They are optional, of course.

	/ON lists the argument list as part of its output, and does also include
	the hook's userial (unique serial)

*** News 01/01/2005 -- New function: $hookctl() (howl)
    This function presents a low-level interface to the /on system

    === ADDING AND REMOVING HOOKS ===
    $hookctl(ADD [#][!][']<noise><TYPE> [<serial>] <pattern> 
							[(<arg list>)] <code>)
	Compare this to:
	   /ON [#]<noise><TYPE> [<serial>] [!]<pattern> [(<arglist>)] <code>

	Where [#] is used to indicate a <serial> number should be used,
	where [!] is used to indicate that the ON is a "negative" ON
	where ['] is used to indicate that the ON is a "flexible" ON
	where [<noise>] is one of "?", "^", "-", "+", "%", or nothing.
	where [<TYPE>] is one of the ON types (ACTION, MSG, PUBLIC, etc)
	where [<serial>] is the ON's serial number (NOT the refnum!)
	where [<pattern>] is the ON's wildcard pattern (the "nick") that is
		matched against $* each time the ON is checked
	where [<code>] is the ircII code that is executed each time the ON
		is run.

	ADD registers a new /ON and returns the /ON's new <refnum>.  This
	<refnum> can be used in other $hookctl() operations.

	Example: $hookctl(ADD ^MSG * (nick, msg) {echo msg from $nick: $msg})
	   is the same as
		/ON ^MSG * (nick, msg) {echo msg from $nick: $msg}
	   except it returns the new /on's <refnum> (of course).

    $hookctl(REMOVE <refnum>) 
	Delete the given /ON.  If <refnum> is -1, it removes the currently
	executing /ON.

    === WORKING WITH THE ONs YOU'VE CREATED ===
    -- In these GET HOOK operations, the values in <> are the values that
       were originally provided to the ADD operation (or the /ON command)
       <Refnum> is always allowed to be -1, and refers to the currently
       executing /ON.

    $hookctl(GET HOOK <refnum> ARGUMENT_LIST)
	Return an /ON's <arglist> if it has one.

    $hookctl(GET HOOK <refnum> FLEXIBLE)
	Return 1 if /ON was created with ['] (flexible -- <pattern> is 
		expanded each time the /ON is checked)
	Return 0 if not.

    $hookctl(GET HOOK <refnum> NICK)
	Return an /ON's <pattern>.

    $hookctl(GET HOOK <refnum> NOT)
	Return 1 if /ON was created with [!] (negative on)
	Return 0 if not.

    $hookctl(GET HOOK <refnum> NOISE)
    $hookctl(GET HOOK <refnum> NOISY)
	Return an /ON's <noise>.

    $hookctl(GET HOOK <refnum> PACKAGE)
	Return an /ON's </package> value.

    $hookctl(GET HOOK <refnum> SERIAL)
	Return an /ON's <serial>.  If one wasn't given, 0 is the default.

    $hookctl(GET HOOK <refnum> SKIP)
	Return 1 if the /ON is being "skipped" 
		("skipped" == ignored -- treated as if it had been deleted)

    $hookctl(GET HOOK <refnum> STUFF)
	Return an /ON's <stuff>
		("stuff" == the ircII code when the /ON goes off)

    $hookctl(GET HOOK <refnum> TYPE)
	Return an /ON's <TYPE>

    -- In these SET HOOK operations, if you attempt to change a value so it
       clobbers (duplicates) an existing ON's value, then the operation will
       either fail, or it will replace the existing ON.  I'll have to ask
       howl how he handled this.

    $hookctl(SET HOOK <refnum> ARGUMENT_LIST)
	Clear an /ON's argument list (so it no longer takes one)

    $hookctl(SET HOOK <refnum> ARGUMENT_LIST <list>)
	Replace an /ON's argument list

    $hookctl(SET HOOK <refnum> FLEXIBLE [0|1])
	Clear (or set) an /ON's flexible-pattern attribute

    $hookctl(SET HOOK <refnum> NICK <pattern>)
	Change an /ON's <pattern>.  Warning -- You can only have one /ON
	per serial number with the exact same <pattern>.

    $hookctl(SET HOOK <refnum> NOT [0|1])
	Clear (or set) an /ON's "negative on" attribute

    $hookctl(SET HOOK <refnum> NOISE <noiseref|noise>)
    $hookctl(SET HOOK <refnum> NOISY <noiseref|noise>)
	Change an /ON's <noise> value using either a noise-refnum or
	the noise's name itsself
	Example: $hookctl(SET HOOK 147 NOISE SILENT)
	     and $hookctl(SET HOOK 147 NOISE 1) do the same thing.
	
    $hookctl(SET HOOK <refnum> PACKAGE <string>)
	Change an /ON's </package> value.  This is used by /unload.

    $hookctl(SET HOOK <refnum> SERIAL <number>)
	Change an /ON's <serial> value.  Warning -- You can only have one 
	/ON per serial number with the exact same <pattern>.

    $hookctl(SET HOOK <refnum> SKIP [0|1])
	Clear (or set) an /ON's skippable attribute.  When an /ON is being
	"skipped", it cannot ever be executed; it is treated as if it were
	deleted. 

    $hookctl(SET HOOK <refnum> STUFF <ircII code>)
	Change the ircII commands executed when an /ON goes off.


    === GETTING INFORMATION ABOUT ON TYPES ===
    $hookctl(LIST)
    $hookctl(LIST LISTS)
	Return all of the valid <TYPE>s

    $hookctl(LIST LISTS <pattern>)
	Return all of the valie <TYPE>s that match the <pattern>
	Ex: $hookctl(LIST LISTS g*) returns "GENERAL_NOTICE GENERAL_PRIVMSG"

    $hookctl(LIST POPULATED_LISTS)
	Return all of the valid <TYPE>s that have an /ON registered for them.

    $hookctl(LIST POPULATED_LISTS <pattern>)
	Return all of the valid <TYPE>s that match the <pattern> that have an
	/ON registered for them.

    $hookctl(LIST HOOKS)
	Return all of the registered <refnum>s

    $hookctl(LIST HOOKS <pattern>)
	Return all registered <refnum>s for <TYPE>s that match <pattern>.
	Ex: $hookctl(LIST HOOKS MSG) returns the refnums of your /ON MSG's

    $hookctl(FIRST_NAMED_HOOK)
	Return the number such that
		$word($hookctl(FIRST_NAMED_HOOK) $hookctl(LIST))
	returns the first non-numeric /ON type.

    $hookctl(NUMBER_OF_LISTS)   
    $hookctl(COUNT)
    $hookctl(COUNT LISTS)
	Return the number of items in $hookctl(LIST)

    $hookctl(COUNT LISTS <pattern>)
	Return the number of items in $hookctl(LIST LISTS <pattern>)

    $hookctl(COUNT POPULATED_LISTS)
	Return the number of items in $hookctl(LIST POPULATED_LISTS)

    $hookctl(COUNT POPULATED_LISTS <pattern>)
	Return the number of items in $hookctl(LIST POPULATED_LISTS <pattern>)

    $hookctl(COUNT HOOKS)
	Return the number of items in $hookctl(LIST HOOKS)

    $hookctl(COUNT HOOKS <pattern>)
	Return the number of items in $hookctl(LIST HOOKS <pattern>)

    $hookctl(GET LIST <TYPE> NAME)
	This just returns <TYPE>, since the name of any <TYPE> is itsself.

    $hookctl(GET LIST <TYPE> PARAMS)
	Return the mininum number of words in $* for any /ON of this type.
	Remember that your <pattern> is expected to match a $* that has 
	AT LEAST this number of words.  If your <pattern> doesn't, the /ON
	will never go off.

    $hookctl(GET LIST <TYPE> MARK)
	Return the number of invocations of this /ON type are pending.
	For example, the first time an /ON MSG event is thrown, then 
	$hookctl(GET LIST MSG MARK) is 1.  If your /ON does something 
	funky like a /WAIT and another MSG comes in before your /ON is
	finished, then $hookctl(GET LIST MSG MARK) is 2.

    $hookctl(GET LIST <TYPE> FLAGS)
	This is an internal bitmask value.  The only defined bit is 1, which
	is used to prevent an /ON from going off recursively.  One such ON
	is /ON INPUT.  If $hookctl(GET LIST INPUT MARK) is 1, then another
	/ON INPUT event is thrown, no /ON's will actually be executed; the
	ON is considered unhooked.  This allows you do perform certain
	commands (like /sendline) from within certain /ON's (like /on input)
	where without this flag that would result in infinite recursion
	(and crash)

    === GETTING INFORMATION ABOUT NOISE TYPES ===
    $hookctl(DEFAULT_NOISE_LEVEL)
	This always returns "NORMAL" for now.  This is the <noisetype> whose
	VALUE (see below) is the null character.

    $hookctl(NOISE_LEVELS) 
	This returns all of the <noisetype> values.
	Ex: $hookctl(NOISE_LEVELS) returns "SILENT QUIET NORMAL NOISY SYSTEM"

    $hookctl(NOISE_LEVELS <pattern>)
	This returns all of the <noisetype> values that match <pattern>
	Ex: $hookctl(NOISE LEVELS s*) returns "SILENT SYSTEM"

    $hookctl(NOISE_LEVEL_NUM)   
	This returns the highest <noiseref> value.

    In the GET NOISE operations, <noisetype> is the name of a noise 
    flag.  This is one of "SILENT", "QUIET", "NORMAL", "NOISY", and 
    "SYSTEM".  <noiseref> is a refnum that uniquely identifies each
    of the noise types.  The above are numbered 1, 2, 3, 4, and 5 
    respectively.

    $hookctl(GET NOISE <noisetype|noiseref> NAME)
	Get the name of the noise type.  One of "SILENT", "QUIET", 
	"NORMAL", "NOISY", or "SYSTEM"

    $hookctl(GET NOISE <noisetype|noiseref> DISPLAY)
	Returns 0 if the noise type does a /SET DISPLAY OFF while
		executing the /ON body.
	Returns 1 if /SET DISPLAY is not changed when the /ON goes off.

    $hookctl(GET NOISE <noisetype|noiseref> ALERT)
	Returns 0 if you are not told when the /ON is executed.
	Returns 1 if you are told whenever the /ON is executed.

    $hookctl(GET NOISE <noisetype|noiseref> SUPPRESS)
	Preface:  With most /ON's, if you do not have any /ON's that 
		are appropriate to run, then some "default" action will
		be taken.
	Returns 0 if executing the /ON does not cause the "default" action
		to be suppressed.
	Returns 1 if executing the /ON causes the "default" action to be
		suppressed.

    $hookctl(GET NOISE <noisetype|noiseref> VALUE) 	(refnum)
	Returns the refnum of the noise type.
	Example: $hookctl(GET NOISE SILENT VALUE) returns "1".

    $hookctl(GET NOISE <noisetype|noiseref> IDENTIFIER)
	Returns the <noise> value to use when you want to use this noise type.
	Example: $hookctl(GET NOISE SILENT IDENTIFIER) returns "^".

    $hookctl(GET NOISE <noisetype|noiseref> CUSTOM)
	This always returns 0 for now.

    === MISCELLANEOUS OPERATIONS ===
    $hookctl(EXECUTING_HOOKS)
	This returns the refnums of all of the hooks that are currently 
	pending (executing).  Since /ONs work like a LIFO queue, the first
	word is the current /ON, and the second word is the /ON that is waiting
	for the first one to finish, etc.   Obviously you can use this to 
	operate on an /ON from within itsself whenever it goes off.

    $hookctl(HALTCHAIN <refnum>)
    $hookctl(DENY_ALL_HOOKS)
    $hookctl(DENY_ALL_HOOKS 1)
    $hookctl(DENY_ALL_HOOKS 0)
    $hookctl(EMPTY_SLOTS)
    $hookctl(HOOKLIST_SIZE)
    $hookctl(LAST_CREATED_HOOK)
    $hookctl(PACKAGE <package> [<type>])
    $hookctl(SERIAL <sernum> [<type>])
    $hookctl(RETVAL)
    $hookctl(RETVAL <value>)
    $hookctl(LOOKUP <type> <pattern> <serial>)
    $hookctl(MATCH <type> <pattern>) 

*** News 01/01/2005 -- New function: $hookctl() (howl)
    This new function, $hookctl(), lets the users do pretty much whatever
	they

	$hookctl() arguments:
   ADD <#!'[NOISETYPE]><list> [[#]<serial>] <nick> [(<argument list>)] <stuff>
       Argument list not yet implemented for $hookctl()
   ADD <#!'[NOISETYPE]><list> [[#]<serial>] <nick> <stuff>
       - Creates a new hook. Returns hook id.
   COUNT
       - See COUNT/LIST
   HALTCHAIN <recursive number>
       - Will set the haltflag for eventchain. May halt the current chain,
         or any chain currently being executed.
         Returns 1 on success, 0 otherwise.
   DEFAULT_NOISE_LEVEL
       - returns the 'default noise level'. It is not currently possible
         to change the current noise level, and probably never will be.
   DENY_ALL_HOOKS <arguments>
       - this sets the deny_all_hooks flag, or gets it's value. If set,
         to anything non negative, all hooks will be "ignored", and the
         default action of any event will be taken. Similar to a /DUMP ON
         but doens't actually remove any hooks.
   EMPTY_SLOTS
       - will return a list of empty slots in the hook-list.
   EXECUTING_HOOKS
       - will return a list of the current executing hooks. This is a
         'recursive' list, listing the current hook first.
   FIRST_NAMED_HOOK
       - returns FIRST_NAMED_HOOK
   HOOKLIST_SIZE
       - will returns HOOKLIST_SIZE
   LAST_CREATED_HOOK
       - returns the value of LAST_CREATED_HOOK
   LIST
       - See COUNT/LIST
   NOISE_LEVELS <pattern>
       - Returns a list of 'noise-types'. If <pattern> is specified only
         noise levels matching pattern will be returns.
   NOISE_LEVEL_NUM
       - Returns NOISE_LEVEL_NUM
   NUMBER_OF_LISTS
       - Returns NUBER_OF_LISTS
   PACKAGE <package> [<list>]
       - Returns a list of hooks of the given package. If <list> is
         specified, it will return only hooks in list <list>
   RETVAL <recursive number> [<new value>]
       - If recursve number isn't specified, 0 (the current) is specified.
         Will either return the value of retval for the given hook, or
         set it.
   SERIAL <serial> [<list>]
       - Works exactly like PACKAGE.

   GET <type> <arg>
       - See GET/SET
   LOOKUP <list> <nick> [<serial>]
       - Returns hook matching given parametres.
   MATCH <list> <pattern>
       - Returns a list of matching hooks.
   REMOVE <hook id>
       - Removes the hook with the given hook ID. Returns 1 on success,
         0 otherwise.
   SET <type> <arg>
       - See GET/SET

   * GET/SET usage
   GET gettype <arguments>
       - will return 'gettype'
   SET gettype <arguments>
       - will set 'gettype' or similar, and return 1 on success. Not all
         'gettypes' may be set, and not all gettypes will silently ignore
         being set.

   It is very important to remember that GET won't ever SET anything(!!!)
   Won't, and shouldn't.

   * GET/SET types:
   HOOK <argument>
       - More info on this under GET/SET HOOK
   LIST <arguments>
       - More info on this under GET/SET LIST
   NOISE <argument>
   NOISY <argument>
       - More info on this under GET/SET NOISE/NOISY
   MATCHES <argument>
       - More info on this under GET/SET MATCHES

   * GET/SET HOOK usage:
       GET HOOK <hook id> <prop> <arg>
       SET HOOK <hook id> <prop> <arg>

       <prop> may be one of the following:
       ARGUMENT_LIST
           - Returns or sets the argument list, if SET and <arg> is empty,
             it will be set to NULL, and therefore not used.
       FLEXIBLE
           - Returns or sets the value of flexible
       NICK
           - Sets or gets the hook's nick. The position of the hook will
             be changed if needed, and it is not possible to change this
             to a "crashing nick"
       NOT
           - Sets or gets the value of NOT.
       NOISE
       NOISY
           - Sets or returns the value of noisy.
       PACKAGE
           - Returns or sets the hook's packagename
       SERIAL
           - Returns or sets the serial for the hook. The hook's position
             in the list will be changed if necesarry, and it is not
             possible to set the serial to a crashing serial.
       SKIP
           - Returns or sets the value of skip.
       STUFF
           - Returns or sets the value of stuff.
       TYPE
           - Returns or sets the type.

   * GET/SET LIST usage:
       GET LIST <listname> <prop>
       SET LIST <listname> <prop> - not functional

       <prop> may be one of the following:
       COUNT
           - Returns count of hooks
       FLAGS
           - Returns flags
       MARK
           - Returns mark
       NAME
           - Returns name
       PARAMETERS
       PARAMS
           - Returns value of params



   * GET/SET NOISE/NOISY usage:
       GET NOISE <noisename> <prop>
       SET NOISE <noisename> <prop> - not functional

       <prop> may be one of the following:
           ALERT
               - returns value of alert.
           CUSTOM
               - returns value of custom.
           DISPLAY
               - returns value of display.
           IDENTIFIER
               - returns value of identifier.
           NAME
               - returns name.
           SUPPRESS
               - returns value of suppress.
           VALUE
               - returns value of value. d'oh!

   * GET/SET MATCHES:
       - This function is not ready yet, and will currently RETURN_NULL.

   * COUNT/LIST usage:
      COUNT / LIST work doing the same, the only difference is that
      COUNT will return the count of lists/hooks, while list will return
      a list

      The following options are permitted:

          LISTS <pattern>
          - Will either return all lists available, or only the
            matching ones.
          POPULATED_LISTS <pattern>
          - Works _just_ like LISTS, but will only return "populated"
            lists
          HOOKS <pattern>
          - Will either return all the hooks on the system, or all
            the hooks in the matching lists



*** News 10/29/2004 -- New /ison features.
	To go with the -d and -f switches, the following switches have been
	added to exploit the new queueing mechanism:

		-n                  # Prioritise this request.
		-s                  # Send the next ison request now.
		-len number         # Change the number of nicks per request.
		-oncmd {commands}   # Run these commands for online users.
		-offcmd {commands}  # Run these commands for offline users.

	The descriptions of these switches are simplistic and a little
	inaccurate.  Some clarifications follow.

	-len will change the maximum length of an ISON request from 500 to the
	given number.  This number will be used for _all_ requests from then on
	including those from the notify system.  In practice it may be
	necessary in some cases to tune this value downwards to avoid the
	server dropping some names off the list when they are all online.

	-oncmd and -offcmd will run the given lines of code with $* set to all
	the users that a reply indicates are online or offline, respectively.
	Note that a single /ison request can generate multiple replies.  Also
	note, that there is no guarantee that the code will or will not run if
	$* is empty.

	-n will place the current requests at the head of the queue.  This is
	useful when many requests are waiting to be sent and it is necessary to
	have this one replied to quickly.  Note that if the request isn't
	actually sent by the time the next -n is used, the newer request will
	always get the higher priority.

	-s will "kick" the ison system back into action in cases where
	something has gone wrong and it has become necessary to use the -f flag
	for eg.  This is more of a debugging tool.  It will not actually cause
	more than $serverctl(get $servernum() maxison) requests to be sent.

*** News 10/01/2004 -- New status format, %{3}W
	This is a compromise between %W and %{2}W.  %W only shows in the
	input window when there are split windows, and %{2}W shows in all
	split windows.  So %{3}W shows in the input window, even if it is
	the only visible window (ie, there are no split windows).

*** News 10/01/2004 -- New window option, "toplines"
	You may now reserve 0 to 9 lines at the top of every windows to be
	removed from the scrollable portion of the window, creating a place
	for you to put things like a channel topic, or channel users, or
	whatever.

	/WINDOW TOPLINES <N>    	
		Reserves and displays <N> lines at the top of the window, 
		which will not be part of the window's scrollable display.
		By default, toplines are blank until you set them with...
	/WINDOW TOPLINE <N> "<string>"	
		Sets the window's <N>th topline to <string>.  <N> must be 
		1 to 9.  You should put <string> in double quotes.  You can
		change toplines even if they aren't visible.

	$windowctl(GET <refnum> TOPLINES)
		Returns the number of toplines reserved at the top of the
		window.
	$windowctl(GET <refnum> TOPLINE <n>)
		Returns the <n>th topline for the window.

*** News 09/14/2004 -- Added Howls shebang script support.
	It is now possible to write epic scripts that run from the shell
	command line.  Yaaay..!  The nature of the epic binary itself made this
	a little difficult at the interface level, so a little hackery was
	required.  The form of the shebang line is this:

		#!/path/to/epic -S [command line options] -l

	Note that the -l switch must be the last on the line, and -S must be
	the first.  Also note that at this point in time, -S will only work as
	the first part of the first argument.  The results of the use of this
	switch anywhere else is currently undefined.

*** News 09/14/2004 -- Added some features to the commandqueues script.
	The first argument to /1cmd may now have a second number attached, with
	a coma.  This number if given, will cause any recurrences of the same
	command within that number of seconds to reset the last-executed time
	_without_ executing the command.

	Use of the 0 or 1 argument form of /qcmd will now cause the timer to be
	reset to 5 seconds, and if called as a function, the command will be
	returned without executing it.

*** News 08/25/2004 -- New target syntax, -<serverdesc>/<target>
	You can now send messages to the special message type:
		-<serverdesc>/<target>
	where <serverdesc> is a server description (see below) and <target>
	is obviously a nick or channel on that server.  This allows you to
	send a message to a nickname on a server other than your current
	window's server.  For example:
		@serverctl(SET 0 ALTNAME booya)		(set an altname)
		/msg -booya/nick hi there!
	Will send the message to "nick" on server 0.

*** News 08/25/2004 -- "Server descriptions"
	Anywhere EPIC expects you to give it a server, it now expects a 
	"server description" which is one of the following (in this order)

		1) A number, which is taken as a server refnum

		2) An "ourname" of an open server
		3) An "itsname" of an open server
		4) A "group" of an open server
		5) An "alternate name" of an open server

		6) An "ourname" of a closed server
		7) An "itsname" of a closed server
		8) A "group" of a closed server
		9) An "alternate name" of a closed server

	The server description may be a wildcard!  The first server 
	(starting with server 0) that matches is used.

	This means you can do something like this:
		@serverctl(SET 0 ALTNAME booya)		(set an altname)
		/server -0				(disconnect from 0)
		/server +booya				(reconnect to "booya")
	and it will connect to server 0, because server 0 has the alternate
	name of "booya"!

*** News 08/25/2004 -- Alternate server names
	You may now give a server a list of "alternate names".  There is no
	limit.  You add a new alternate name with:
		$serverctl(SET <refnum> ALTNAME <name>)
	You can totaly replace the list with:
		$serverctl(SET <refnum> ALTNAMES <names>)
	<name> and <names> should be space-separated list of words.

	You can get the list of alternate names with
		$serverctl(GET <refnum> ALTNAMES)

*** News 08/25/2004 -- Aliases now shown with their argument lists
	When you do /alias, you now see the argument list along with all
	of the other stuff.

*** News 08/25/2004 -- Mangling now supports "ALT_CHAR"
	How did this ever get missed?

*** News 08/25/2004 -- New function, $mktime()
	Usage: $mktime(year month day hour minute second DST)
	The first six arguments are required.  Returns -1 on error.
	Returns whatever mktime(3) on your system would return, usually
	the number of seconds since the epoch represented by the arguments.

*** News 08/25/2004 -- Support for ircnet's "unique id" nicknames
	You can now always use your "unique id" as your nickname, and also
	the "0" shortcut nickname, on ircnet.

*** News 08/25/2004 -- New status expando, %{1}F
	This is just like %F, it displays all of the "notified" windows,
	except it uses the window's "notify_name" instead of its refnum.

*** News 08/25/2004 -- New window command, /WINDOW NOTIFY_NAME
	If you change the /WINDOW NOTIFY_NAME, and use the %{1}F status
	expando, then the notify_name, and not the window's refnum, will
	show up.  Howl wanted to colorize the refnum up, and this is how
	you should do it.

*** News 08/25/2004 -- Support for +I channel mode (ratbox)
	Because adm asked me to

*** News 08/25/2004 -- User-created /SETs, /SET -CREATE   -- WARNING
			*** DEPRECATED ***
	You can create your own /set's for now with 
		/SET -CREATE <name> <type> {<code>}
	where <type> is one of BOOL, STR, or INT.  <code> is code that 
	will be run any time the user does /SET <name> <newval>.  You can
	/SET the value within <code> to override the user's value.  

	*** WARNING *** This interface is temporary and will go away in 
	the future.  It will be replaced by $symbolctl() which has not
	yet been written, so stay tuned for more information!
			*** DEPRECATED ***
	This feature only existed in EPIC5-0.0.3 and was/will be removed
	in EPIC5-0.0.4.  Do not use this feature.
			*** DEPRECATED ***

*** News 08/25/2004 -- Unification of symbol namespaces
	There is now one big table that holds all of the symbol names for
	aliases, assigns, built in commands, built in functions, sets, and
	inline expandos.  You should not notice any changes at all, except
	maybe epic runs faster.  This was done to pave the path towards
	plugins, which will need to be able to add their own commands and
	functions on the fly!

*** News 08/25/2004 -- The /IRCNAME and /REALNAME commands removed here...
	because they are duplicates of /SET REALNAME.  Use the /SET now.

*** News 08/25/2004 -- $stripcrap(ALL) no longer strips "ALL_OFF"...
	because the crap-mangler makes liberal use of ALL_OFFs and it is of
	no harm to leave them in there, and it does great harm to take them
	out. ;-)  If you want to remove them, do $stripcrap(ALL,ALL_OFF)

*** News 08/25/2004 -- New script 'builtins' loaded from global
	Some things are starting to migrate from hardcoded builtins to 
	script features.  They are not "being removed", their implementation
	is just changing.  This script will contain backwards compatability
	stuff for epic4.  You really do need this script!  Do 'make install'!

*** News 08/25/2004 -- Automatic command completion removed
	You can no longer do /whoi as a replacement for /whois.  You'll have
	to spell out the command name in full now.

*** News 08/25/2004 -- Using your nickname as a command removed here
	You can no longer do /<mynick> as an alias for /me.  Just use /me.

*** News 08/25/2004 -- The COMMAND_COMPLETION keybinding removed here
	A new script replacement is forthcoming -- stand by!

*** News 08/25/2004 -- New serverctl, $serverctl(GET <refnum> STATUS)
	This returns the server's current status, which is described below
	in the /on server_status stuff.

*** News 08/25/2004 -- New script, 'slowcat'
	This script cats a file to your current target, 2 lines per second
	to avoid triggering flood control.  
		/load slowcat
		/slowcat filename

*** News 08/25/2004 -- New status expando, %{2}W
	This acts just like %W, but it shows in every window, and not just 
	the current window.

*** News 08/25/2004 -- New /SET, /SET OLD_SERVER_LASTLOG_LEVEL
	When you /WINDOW SERVER to move a window to a server that is already
	connected, the window's level will be set to this /set value.  This
	is important, because of the following situation:
		Window 1, server 0, level ALL
		Window 2, server 1, level ALL
	If you do /window 1 server 1  then you have two servers connected
	to server 1 with level "ALL".  Who wins?  Not you. ;-)  This defaults
	to NONE, which is probably the only sensible value.

*** News 08/25/2004 -- /WINDOW KILL_ALL_HIDDEN kills your hidden windows.
	After you run this, you will be left with only your visible windows.

*** News 08/25/2004 -- /ON TYPE !"PATTERN" acts as an exception.
	If you use this syntax, then the default action will occurs whenever
	the pattern is matched.  This is how ircII's /ON TYPE ^PATTERN works.
	For example:
			on ^msg * echo msg from $0: $1-
			on ^msg !"nick" 
	In this case, if anyone but nick sends you a msg, it is echoed as
	in the first /on.  But if nick sends you a message, it will be
	displayed in the "default" way by epic, as though you did not have
	an on at all.

*** News 08/25/2004 -- It is now always safe to delete ONs from within ONs
	Up until now, you needed to /defer the removal of any /ONs from
	within other /ONs, or you risked crashing epic.  This meant you 
	could not safely stop an /ON with a higher serial number from
	running by deleting it.  All of this has been fixed now.  You can
	delete /ONs without restriction and the change takes place 
	immediately.

*** News 08/25/2004 -- ONs no longer compile patterns to regexes
	This was fraught with peril, so ONs no longer compile their
	patterns to regexes, and now we do things like we have always
	done with ONs

*** News 08/25/2004 -- New built in function, $regcomp_cs()
	This is just like $regcomp(), but it's case sensitive.

*** News 08/25/2004 -- In /ON DCC_RAW "* E *", $3 is the port number
	Previously it held the "othername", which wasn't terribly useful.

*** News 08/25/2004 -- /WAIT =<fd> waits for a dcc connection to complete
	If you do $connect(), it is nonblocking and returns before the
	connection is ready to be used.  If you need to wait until the
	connection completes, like it did in epic4, do this:
			@fd = connect(host port)
			wait =$fd
	and it's pretty much the same.  This wait is of course recursive
	(and does not block the client)

EPIC5-0.0.2

*** News 08/25/2004 -- Level names are always plural, except for CRAP
	The levels have these names now, and they're gonna stay this way:

		CRAP	PUBLICS	MSGS	NOTICES	WALLS	WALLOPS
		OPNOTES	SNOTES	ACTIONS	DCCS	CTCPS
		INVITES	JOINS	NICKS	TOPICS	PARTS	QUITS
		KICKS	MODES	USER1	USER2	USER3	USER4
		USER5	USER6	USER7	USER8	USER9	USER10
		ALL (special)	NONE (special)

*** News 08/18/2004 -- Overloadable function aliases.
	When an alias name "collides" with a built in function, the built in
	function has traditionally been called.  This update changes this
	behaviour to calling the alias first.

	NOTE:  This will cause recursive loops in scripts that rely on this
	behaviour.  To fix this the aliases in question need to be renamed or
	rewritten to use the following ::function() feature.

	It is possible to call the built in with the :: notation used for
	global variables, as in $::function().  $:function() will explicitly
	call the alias.

	NOTE:  Do not use the $::function() and $:function() features just yet,
	as they will crash epic if the alias or the built in doesn't exist.  It
	is safe to use if you are sure they do though.

*** News 08/02/2004 -- New commands for $dccctl().
	$dccctl(readables) will return the refnums of DCCs that have data
	waiting to be read and $dccctl(get [refnum] readable) will return a 1
	or 0 depending on whether the given refnum is readable or not.

	Since epic will automatically read data from all unheld DCCs, this
	feature is expected to be useful only for DCCs in the "held" state.

*** News 08/02/2004 -- General improvements to the flood detection system.
	Flood detection now works for channel PARTS and for ctcp replies, which
	is bound to the NOTICES flood.

	Also, the first argument from each flood returned from $floodinfo()
	when flood_maskuser is set to 1 or 2 is now a valid user@host mask for
	the flooder in question, suitable for putting in a ban or kline.

*** News 08/02/2004 -- New argument for /on flood.
	The fourth argument ($3) in the flood hook is now the number of repeats
	of the flood in question.  This makes it easy to deal with particular
	kinds of floods in different ways, as they occur.  For example:

	/on flood "% parts % 5" mode $2 +b *!*@$after(@ $userhost())
	/on flood "% ctcps % 5" mode $2 +b *!*@$after(@ $userhost())
	/on flood "% % % 50" mode $2 +b *!*@$after(@ $userhost())

*** News 08/02/2004 -- Improvements to $floodinfo().
	$floodinfo() will now accept as input the same list of lists it
	outputs.  The lists themselves don't have to be complete.  Any
	unspecified arguments will match all records.  For example:

	$floodinfo("% #chan joins")  # Return all join records for #chan.
	$floodinfo(*)                # This still works.
	$floodinfo($floodinfo(*))    # Same output as above.

	Feeding $floodinfo() output back into its input is useful for tuning
	the flood /sets by seeing which non-flooders are being caught long term
	in the system.

	The fields are these:

	u@h mask that matches the flooder.  Defaults to "*".
	channel mask.  Defaults to "*".
	flood type mask.  Defaults to "*".
	Server number.  Defaults to -1, which matches all.
	Numeric minimum number of flood hits.
	Numeric minimum duration of flood.
	Numeric minimum flood rate.

	The last three numeric arguments may be negative, in which case, they
	specify the _maximum_ values.  These fields make it possible to deal
	with different kinds of floods in different ways _after_ they occur.
	For example, a join flood may be falsely triggered by a net join, but
	it is reasonable to expect that if you have join and part flood records
	for the same u@h, then it is participating in a join/part flood.

	/on flood "% parts % 5" {
		if (floodinfo("$userhost() $2 joins $servernum() 5")) {
			mode $2 +b *!*@$after(@ $userhost())
		}
	}

	Or alternately:

	/on flood "% joins % 5" {
		if (floodinfo("$userhost() $2 parts $servernum() 5")) {
			mode $2 +b *!*@$after(@ $userhost())
		}
	}

*** News 08/02/2004 -- Changes to /userhost and notify.
	The changes to the ison back end mentioned in the previous entry now
	also apply to the userhost back end, along with the caveats relating to
	/wait and flush.  This applies to the /userhost, /userip and /usrip
	commands.

	A new option has been added to these commands.  /userhost -count
	[number] will change the number of items that epic will put into each
	USERHOST request.  This isn't a particularly significant change since
	this number is already tuned to a number that works on all servers.

	To change the number of USERHOST requests sent at one time, use
	$serverctl(set [servernum] maxuserhost [number]).  Set it to 0 to turn
	the new behaviour off.

*** News 07/08/2004 -- Changes to /ison and notify.
	The back end of the notify system and the /ison command has been
	changed to permit only a certain number of ISONs to be sent to a server
	at one time.  The benefit of this is that it will typically prevent a
	large notify list flooding the client off the server.

	The down side is that it will cause scripts like $is_on() in script/guh
	that use "/wait for ison .." to fail until they have been fixed.  The
	fix is to put "@ serverctl(set $servernum() maxison 0)" at the top of
	any alias that uses it.  This will turn the new behaviour off.

	The notify system itself will not queue an ISON to be sent if there are
	any ISONs waiting to be sent, but the /ison command will.

	One final note is that the "waiting to be sent" queue won't be flushed
	when the client reconnects to the server or when "/ison -f" is run.
	This won't cause any particular damage, but it's not nice and will
	probably change soon.

*** News 07/08/2004 -- Userhost updating in NICK changes.
	This is relevant to those who use the $serverctl() maxcache feature,
	which, if in effect, will prevent a /who message being sent to a
	server, and thereby make $userhost() fail for every nick that joined
	the channel before the client did.

	This patch will grab the userhost information for these users from the
	NICK message itself, and help to rebuild the $userhost() database
	faster.  It also makes it possible to manually get it into the database
	by /pretend'ing a NICK message with a userhost obtained from other
	sources such as the /userhost command.

*** News 03/19/2004 -- New built in function, $tobase(<base> <num>) [howl]
	This function converts <num>, a number in base 10, to base <base>.
	For example, $tobase(16 65536) returns "10000"

*** News 03/19/2004 -- New built in function, $strtol(<base> <num>) [howl]
	This function converts <num>, a number in base <base> to base 10.
	For example, $strtol(16 10000) returns "65536"

*** News 03/19/2004 -- Changes to /WINDOW NOTIFY, /WINDOW NOTIFIED
	These two /WINDOW operations now take ON, OFF, or TOGGLE arguments,
	instead of taking no arguments and behaving as toggle switches.
	If you do not provide an argument, they show you their current
	values as the other /WINDOW boolean values do.

*** News 03/19/2004 -- Addition and changes to $windowctl()
	$windowctl(GET <refnum> MISCFLAGS) always returns 0, and
	$windowctl(GET <refnum> NOTIFY) returns 1 if /window notify is on
	$windowctl(GET <refnum> NOTIFIED) returns 1 if /window notified is on.
	NOTIFY and NOTIFIED replace MISCFLAGS.

*** News 03/19/2004 -- New key binding, SWITCH_QUERY
	Whenever a window has multple nicknames in its nickname list,
	and one of those nicknames is active as the window's query, it
	is possible to use this binding to switch between all of the 
	nicks in the nick list, just in the same way you can switch
	between channels using SWITCH_CHANNELS.  If the window does not
	have an active query, this key binding will have no effect, even
	if the window has nicks in its nicklist!

*** News 03/19/2004 -- Unification of /WINDOW QUERY and /WINDOW ADD
	Historically, when you /WINDOW QUERY (or just /QUERY) <NICK>, then
	it would add <NICK> to the window's "NICK LIST".  The "NICK LIST" 
	is a list of nicknames for which output goes to that window, just
	like output to channels go to windows.  Output to or from a nick
	that is not on any window's "nick list" goes to the LEVEL_MSG level.
	When you used /WINDOW QUERY <NICK2> to change the query, it would
	remove the old query from the nick list, and messages to and from
	the original query went back to LEVEL_MSG.

	Well, this has been unified somewhat.  Now the following rules apply:
	1) When you /WINDOW QUERY <NICK>, then <NICK> is added to the 
	   window's nick list.
	2) If the window already has a query, then the old query nickname
	   is NO LONGER REMOVED from the window's nick list.  Output to 
	   that nick will continue to go to the window as it had before.
	3) When you use /WINDOW QUERY to cancel a query, then the current
	   query IS STILL REMOVED from the window's nick list, and output
	   to or from that nick will go to LEVEL_MSGS.

	It is no longer possible to have a window query that is not on the
	window's nicklist, because the query is selected from the members of
	the window's nicklist, rather than being a separate thing.

*** News 03/17/2004 -- Change to how /SET INDENT behaves
	Historically, if you have /SET INDENT ON, and the first word of the
	first line of output is wider than 1/3 of your screen, then the
	second (and subsequent) line(s) of output are NOT INDENTED.  This 
	has been changed so subsequent lines are indented 1/3 of the window's
	width.  To understand this change, think about how /set indent usually
	works, and if it would indent more than 1/3 of your screen, then it 
	will indent 1/3 instead of not at all.

*** News 03/17/2004 -- New flag to /XECHO , /XECHO -F
	If you use the /XECHO -F flag, "hidden window output notification"
	will not occur for any hidden windows that receive the output. 
	
*** News 03/16/2004 -- You can now bind the 255 character ()
	There has been a problem with the new key binding system that
	made it difficult for Russian language speakers to bind the 255
	character which is in their alphabet.  This should be fixed now.

*** News 03/16/2004 -- Can now join channels simultaneously per window
	Previously, if you attempted to join multiple channels in the same
	window simultaneously, you were not assured that all of the channels
	would go to that window.  Now you can be assured of this.  This should
	make reconnection/rejoin scripts much more sane.

*** News 03/16/2004 -- New built in function, $startupfile()
	This expands to the file that the client loaded at startup as your
	"startup file".  Usually this is ~/.ircrc or ~/.epicrc or whatever
	you specified as the IRCRC environment variable or the argument to
	-l or -L on the command line.

*** News 03/16/2004 -- Unknown CTCP requests offered via /ON CTCP_REQUEST
	It was pointed out that unknown/unhandled CTCP requests were only
	being hooked through /on ctcp, so it wasn't possible to use 
	/on ctcp_request to handle EVERY request.  Well, now unhandled CTCPs
	are hooked through both /on's just like handled CTCPs are.

*** News 03/16/2004 -- Semantic changes to $connect()
	You used to be able to depend on /ON DCC_RAW "% % E %" or 
	/ON DCC_RAW "% % C" hooking before $connect() returned.  Now that
	$connect() is nonblocking, YOU CAN NO LONGER DEPEND ON THIS.  You
	must set up your script to assume that /ON DCC_RAW will be hooked
	asynchronously, after at least the next sequence point.  Think of 
	it as being like not being able to depend on /WHOIS returning the
	numerics.  I'll probably add a way to /wait for a connection in the
	future.  Stay tuned.

*** News 03/16/2004 -- DCC connections are now nonblocking
	All connect()ions for DCC, including /DCC GET, /DCC CHAT, /DCC RESUME
	and $connect() are all fully nonblocking.  This means all connects
	in EPIC are now fully nonblocking! HUZZAH!

*** News 03/15/2004 -- /HELP command now handled by script
	The built in /HELP command has been replaced by a script that
	was written by howl for our use.  Much thanks to him!

*** News 03/14/2004 -- Six new USER lastlog levels
	You may now use USER5, USER6, USER7, USER8, USER9, and USER10
	as levels with your window, lastlog, flood, and ignore.  Just 
	use /xecho -l USER5 for example to send to your USER5 window.

*** News 01/20/2004 -- kqueue() support
	You can uncomment #define USE_FREEBSD_KQUEUE in newio.h if you
	want to play around with this experimental feature.

*** News 01/15/2004 -- /WINDOW DISCON and /WINDOW NOSERV now the same
	There was a subtle semantic difference between /WINDOW DISCON 
	and /WINDOW NOSERV that had to do with the window's "last server"
	that was used for reconnects.  Because the client no longer does
	reconnections, this difference is moot.  These two commands now 
	always do the same thing, which is to disassociate the window
	with any server.  The window becomes "server-less".

*** News 01/15/2004 -- Changes to /SERVER command
	/SERVER
	     Show the server list.
	/SERVER -DELETE <refnum|desc>
	     Remove server <refnum> (or <desc>) from server list.
	     Fails if you do not give it a refnum or desc.
	     Fails if server does not exist.
	     Fails if server is open.
	/SERVER -ADD <desc>
	     Add server <desc> to server list.
	     Fails if you do not give it a <desc>
	/SERVER +<refnum|desc>
	     If the server's state is "CLOSED", change it to "RECONNECT".  
	     This allows the server to reconnect if windows are pointed to it.
	     Note: server reconnection is asynchronous
	/SERVER -<refnum|desc>
	     Unconditionally close a server connection
	     Note: server disconnection is synchronous!
	/SERVER +
	     Switch windows from current server to next server in same group
	/SERVER -
	     Switch windows from current server to previous server in same group
	/SERVER <refnum|desc>
	     Switch windows from current server to another server.

*** News 01/08/2004 -- /ON WIDELIST went away here
	This /ON hasn't been hooked in many a year, and here it officially
	passed into the void.

*** News 01/07/2004 -- Removal of WINDOW BIND feature
	As part of the larger project to decouple windows from channels,
	the "window bind" feature has been removed.  This means you can 
	no longer /WINDOW BIND, /WINDOW REBIND, /WINDOW UNBIND, and you 
	cannot use $windowctl(* BIND_CHANNEL *) or $winbound().  It is 
	expected that eventually scripts will take over the job of routing 
	channels to the appropriate windows and EPIC will stay entirely 
	out of the way.

*** News 01/07/2004 -- New /ON, /ON SERVER_STATUS
	This /ON is thrown every time a server changes its "state".
	The states are listed below in "Server States" and I won't
	go into that again here.
		$0 - The server changing state
		$1 - The old status (a string, not a number)
		$2 - The new status
	If you find that you do something particularly onerous in 
	this /ON and EPIC panics or crashes, try /DEFERing it, and
	if that doesn't work either, let me know.

*** News 01/07/2004 -- Removal of NOTE support
	I doubt anyone will notice this, and if you do, bummer.

*** News 01/07/2004 -- Server states
	Servers now exist in one of several "states" each time it connects
	to the server.  It moves through each of the states from start to
	end, and stays at the end until manually reset by the user (or script)

	RECONNECT	As soon as a window is attached to the server, the 
			server should be connected to.
	CONNECTING	A connection to the server is in progress.  The server 
			is not ready to be used.
	REGISTERING	We are attempting protocol registration (NICK/USER) 
			with the server.  The server is open, but we cannot 
			really use it yet.
	SYNCING		Our registration has been accepted and we're doing 
			whatever meta-tasks are needed to get the connection 
			fully active
	ACTIVE		The connection is fully ready for all use.
	EOF		An End Of File (EOF) has been detected from the server.
			The connection was closed by the server and cannot be 
			used any longer.
	CLOSING		The connection to the server is being shut down.  If 
			the previous state was "ACTIVE" then you can still 
			send something to the server.  If the previous state
			was "EOF" then it's too late.  You cannot stop the
			closing of a server.
	CLOSED		The server is disconnected and cannot be used.  The
			server (and any windows connected to this server) stay
			in this state until the user resets the state to 
			RECONNECT.

*** News 01/07/2004 -- Channels are not tracked across disconnects
	When you are disconnected from a server for *any* reason, EPIC
	will not retain knowledge of the channels for the next connection
	and will not rejoin them.  It is expected that scripts will use
	this to their advantage to fully control the semantics you will
	have governing "auto-rejoin-on-reconnect".

*** News 01/07/2004 -- Server connections are now brought up asynchronously
	When you do /WINDOW SERVER or /SERVER or otherwise change the
	server of a window, the server is not immediately connected or
	disconnected, and the change will not take effect until the
	next time through the event looper.  This means that all server
	connections are "asychronous" (they don't interrupt the current
	flow of the script).  This means you most definitely cannot
	do /WINDOW SERVER <host> CHANNEL <channel> any more.  So please
	stop doing that. ;-)  Use /ON SERVER_STATUS to join channels.

*** News 01/07/2004 -- /XDEBUG SERVER_CONNECT a lot more interesting
	If you want to watch epic work its gory nonblocking connects,
	you can turn on this /xdebug and see everything in its glory.

*** News 01/07/2004 -- Nonblocking server connects
	EPIC now does all server connections using asynchronous, 
	nonrecursive, nonblocking connections.  And yes, it still 
	supports multiple protocols and multiple addresses (ie, 
	"us.undernet.org"), and *yes*, it will try another address if 
	a server refuses us registration ("You do not have access to 
	this server").

*** News 01/07/2004 -- EPIC no longer tracks server "dialect" per se
	The $version() string now always returns "2.8" since all 
	servers are nominally 2.8 class (rfc1459) servers, and epic
	does not attempt to determine if it's an undernet, ircnet,
	efnet, or dalnet server, etc.  This is mostly because scripts
	can hook /on 004 if they care, and the 005 numeric (ISUPPORT)
	is making dependance on the server's version much less important.

*** News 01/07/2004 -- EPIC loads ~/.ircrc (or ~/.epicrc) on 001 now
	Traditionally, ircII has loaded your ircrc when it received 
	the 002 numeric, and traditionally, epic has done it when it
	received the 004 numeric.  Due to some refactoring in epic,
	it is now possible for epic to load your ircrc when it receives
	the 001 numeric *and before it hooks /on 001*  This means you 
	shouldn't have to suffer the default epic output for any of
	the numerics from the server.

*** News 01/07/2004 -- Usermodes now tracked as strings instead of bits
	Before this change, ircII clients had always tracked your user
	and channel modes as bits, and the valid (supported) modes were
	hardcoded into the client at compile time.  With this change, 
	EPIC will no longer track your modes using bits, but instead 
	using strings.  This means that epic won't need source code
	changes to support new modes from your server.  You can't do
	$serverctl(SET|GET <refnum> UMODES) any more (but the old
	"UMODE" still works)

*** News 01/07/2004 -- /ON wildcard patterns now compiled into regexes
			*** OBSOLETE ***
	At or around this date, EPIC started converting wildcard patterns
	used by /ON into extended POSIX regexes and compiling them, and
	using the regexes instead of the pattern matcher.  In the future,
	epic will allow you to specify your own regexes.  "Flexible"
	/on hooks are still wildcard pattern matched (for now) because
	recompiling the pattern every time the /on is thrown is senseless.
			*** OBSOLETE ***
	This feature was removed (see note above for 08/25/2004)
			*** OBSOLETE ***

EPIC5-0.0.1

*** News -- 12/16/2003 -- New levels, KICK, QUIT, and MODE
	So just for a canonical list, here are all of the levels supported
	by flood, ignore, lastlog, and windows:
		CRAP	PUBLICS	MSGS	NOTICES	WALLS	WALLOPS
		OPNOTES	SNOTES	ACTIONS	DCCS	CTCPS
		INVITES	JOINS	NICKS	TOPICS	PARTS	QUITS
		KICKS	MODES	USER1	USER2	USER3	USER4
		USER5	USER6	USER7	USER8	USER9	USER10
		ALL (special)	NONE (special)

*** News -- 12/16/2003 -- Unification of ignore, flood, and lastlog levels
	Previously, the ignore, flood, and lastlog levels used the same
	names, but they had different meanings in each subsystem (ie, CRAP 
	in flood was different from CRAP in ignore, and CRAP in lastlog).
	Now all three subsystems use the same levels, all named the same,
	and (more or less) all defined the same.  There are some holes in
	this conversion cause I didn't check every possible combination.
	Report any odd behavior to me so I can fix it.

*** News -- 12/16/2003 -- New noise type for /ON, /ON %TYPE
	The /ON %TYPE modifier acts just like /ON ^TYPE, because it
	suppresses the default action, but it is unlike /ON ^TYPE
	because it does not turn off the display (what /ON ^TYPE
	does is it prefixes all the commands in the ON body with 
	the ^ modifier, which turns off output for that command.)
	This new modifer does not prefix each command with ^, so any
	commands not so prefixed will generate their normal output.	
	The idea is you can use this for /on set's
		/ON %SET "HOLD_MODE *" {WINDOW HOLD_MODE $*}

*** News -- 12/16/2003 -- Removed /SET BEEP_WHEN_AWAY
	This feature can be re-implemented in one line of script:
		/ON #MSG 617 * {IF (A) {BEEP}}

*** News -- 12/16/2003 -- Removed /SET BEEP_ON_MSG
	The /SET BEEP_ON_MSG feature has been removed because it was only
	half-implemented, and even that half didn't work right.  Keep an
	eye out for a scripted re-implementation of this in the future.

*** News -- 12/16/2003 -- Runtime auto-append-of-$* removed
	Historically, the ircII language has allowed you to auto-append
	$* onto the end of an alias at runtime by creating an alias that
	does not refer to any of the command line arguments.  For example,
		/alias m msg
	behaves at runtime as
		/alias m msg $*
	but with a performance penalty.  This behavior has now been removed
	and if you wish to have $* appended to your aliases, you need to 
	change them.  This change would be backwards compatable with epic4.

# End of file