CHANGED 12/16/98 by Jonathan Atkins. updated for imwheel.

Tue Nov 17 23:10:04 1998  Ian T Zimmerman  <itz@transbay.net>

	* configure.in (release): Step to 1.16.0

Mon Oct  5 22:00:19 1998  Ian T Zimmerman  <itz@rahul.net>

	* Makefile.in (append-diff): Add some flags to the diff command
	(most important, --unidirectional-new-file).

Wed Sep 30 11:08:59 1998  Ian T Zimmerman  <itz@rahul.net>

	* doc/manpager: Add handling of forced line breaks,
	cross-references, and boldface (used for cross-references).

	* doc/doc.gpm: Fixed manpage cross-references.
	(Command Line): Mention disable-paste.

Tue Sep 29 10:39:48 1998  Ian T Zimmerman  <itz@rahul.net>

	* gpn.c (xfer_options): Move the option structure type into
	gpmInt.h to prevent stupid future problems.

	* Makefile.in (install): Add disable-paste (and make it setuid).

	* disable-paste.c (main): Add a new program (not a large one :)

	* mice.c (M_brw): Apply eb patch for broken Fellowes Browser mice.

	* sample/Makefile.in (append-diff): Added rules to recursively
	append to the diff file being created in top directory.

	* Makefile.in (DIST): Remove exec.el from the distribution because
	it is not needed anymore.  Modern Emacs has en --eval command line
	option that allows me to do the same thing directly.

	* configure.in (ITZ_PATH_SITE_LISP): Use it.

	* aclocal.m4 (ITZ_PATH_SITE_LISP): Added.

Mon Sep 14 18:03:45 1998  Ian T Zimmerman  <itz@rahul.net>

	* doc/Makefile.in (gv): Replace "$(TARGET)" with "gpm".

	* Makefile.in (gpm-%-to-@release@.diff): Added rule to create a
	diff from snapshot of previous release.

	* mice.c: Reapply Edmund Grimley Evans' patch with new mouse types
	ms+ and ms-lr which get lost somewhere.

Sun Sep 13 00:08:54 1998  Ian T Zimmerman  <itz@rahul.net>

	* configure.in: Set CURSES_OBJS dependent on finding the curses header.

Sat Sep 12 10:53:54 1998  Ian T Zimmerman  <itz@rahul.net>

	* Makefile.in (LOBJ): Make the buliding of libcurses.o decided by
	configure again.

	* libcurses.c (Gpm_Wgetch): Remove the ELF assembler-dependent
	wgetch hack again, as it is not really needed as long as clients
	are linked statically.

	* gpm.c (processRequest): Removed GPM_REQ_CLEAR again, now that its
	functionality is available with the -A flag and opt_aged.

	* gpmInt.h:  Removed GPM_REQ_CLEAR again, now that its
 	functionality is available with the -A flag and opt_aged.

	* libxtra.c: Removed Gpm_ClearSelection() again, now that its
 	functionality is available with the -A flag and opt_aged.

	* gpm.h: Removed Gpm_ClearSelection() again, now that its
 	functionality is available with the -A flag and opt_aged.

	* mev.c (interact): Removed "clear" command again, now that its
	functionality is available with the -A flag and opt_aged.

	* gpmInt.h: Change signature of xfer_options().
	Remove it altogether, it's better static and split to gpn.c and
	gpm.c.
	Add opt_aged.

	* gpn.c (check_uniqueness): Completely rewritten.  This function
	was a messy mixture of different things.

Wed Sep  9 09:50:59 1998  Ian T Zimmerman  <itz@rahul.net>

	* libxtra.c (Gpm_GetServerVersion): Provide explicit path for
	popen of gpm daemon.

	* sample/Makefile.in (rmev.o): Added target because of usage of
	RMEV_RELEASE. 
	(prefix): Looks like this has to be done in _every_ Makefile,
	sigh.  Even if it is not used explicitly, configure subsitutes
	e.g. "@sysconfdir@" with "$(prefix)/etc".
	(sysconfdir): Added sysconfdir.

	* Makefile.in (gpn.o gpm-root.o twiddler.o): gpn.o added to this
	target because it uses RELEASE symbol substituted by configure.

Tue Sep  8 14:04:44 1998  Ian T Zimmerman  <itz@rahul.net>

	* mev.c (interact): Added case for "clear" command.

	* libxtra.c (Gpm_ClearSelection): Added.

	* gpmInt.h (GPM_REQ_CLEAR): Added (rubino's objection overruled;
	the gpm protocol is not really ugly, it's the code that is, in
	effect casting the data exchanged to Gpm_Conn and Gpm_Event.  It
	could be coded cleanly or even fixed from present state, but don't
	hold your breath, you could suffocate.)

	* gpm.h: Add Gpm_ClearSelection.

	* t-mouse.el (t-mouse-swap-alt-keys): Added.  This is what is
 	called "personal variable" :-) I swap the left and right Alt keys
 	in my Linux keytable, but I still want Emacs to handle left-Alt
 	events.

Wed Sep  2 00:09:42 1998  Ian T Zimmerman  <itz@rahul.net>

	* doc/Makefile.in (maintainer-clean): Ignore errors because the file
	list can be empty.

	* sample/Makefile.in (maintainer-clean): -rm -i for uniformity.

	* Makefile.in (maintainer-clean): Ignore errors because the file
	list can be empty.

	* mouse-test.c (makedev): Add O_NONBLOCK to open ttyS devices
	(otherwise the program hangs in this function, at least on my
	system).

	* doc/doc.gpm (gpm-root): Use @file at several places in
	preference to other markup.  Fix several other general
	documentation bugs, mostly referring to obsolete (non-)features. 

	* doc/Makefile.in (.SECONDARY): Need both %.ps and gpm.ps ---
	looks like a gmake bug.

	* sample/gpm-xterm.h: Applied patch by Jean-Daniel
 	<jd@univ-lyon1.fr> for looking back when scanning arrow key
 	sequences.

	* Makefile.in (datadir): Oops, need to define this if I define
	elispdir.  GNU standard document a little incomplete on this
	point. 

Fri Aug 28 00:06:05 1998  Ian T Zimmerman  <itz@rahul.net>

	* configure.in: Oops, autoconf hates newlines between filenames in
	AC_CHECK_HEADERS.

	* create_vcs: Create symlinks in /dev for the new (or old?) names
	mandated by kernel device list.

	* gpm-root.y (main): Check for both /dev/vcs and /dev/vcs0.  Also
	check for correct device number.

	* configure.in: Check for sys/sysmacros.h and linux/major.h.

Thu Aug 27 09:33:56 1998  Ian T Zimmerman  <itz@rahul.net>

	* sample/Makefile.in: Add TAGS rule.

	* Makefile.in (libgpm.so.@release@): Change sed script to extract
	the major release number, now that there are 2 dots in the
	complete number.

	* configure.in (release): Now includes the patchlevel number, for
	instance 1.15.5.

	* sample/configure.in (release): Define this as RMEV_RELEASE.

	* configure.in (release): Define this as both GPM_RELEASE and
	RMEV_RELEASE. 

	* gpmCfg.h: Delete RELEASE definition. (will be passed from
 	configure).

	* sample/rmev.c (RMEV_DATE): Define from RCS checkin date.
	Delete RELEASE definition. (will be passed from configure).

	* gpmCfg.h (GPM_DATE): Define from RCS checkin date.

	* sample/gpm-xterm.c: Delete RELEASE definition.

	* Makefile.in: For submakes, add -lgpm to LIBS.  Put do-all before
	@SHLIB@ in the rule for all; let's hope that means submakes will
	use the static library.
	Oops, that didn't work.  Another try: name the shared library
	libgpm.so.@release@ maybe that way linker won't find it.

	* configure.in (GPMXTERM): This is needed because we now configure
	under sample.

	* Makefile.in: Add sample to subdirs for recursion.

	* sample/gpm-xterm.c: Include whatever curses header we can find.

	* sample/rmev.c (main): Rewrite to use termios to set raw terminal
	mode, not n?curses.

	* Makefile.in (top_builddir): Add top_builddir because it is now
	mentioned in @LDFLAGS@.

	* configure.in (CPPFLAGS): Add top_srcdir to the include path.
	(LDFLAGS): Add top_builddir to library path.
	Add sample/Makefile to output files.

	* sample/configure.in: Simplified to use default autoconf behavior
	where possible.
	Only look for n?curses if we're building without gpm.
	
	* sample/Makefile.in: Rewrite to be able to act as slave for
 	toplevel.

	* doc/Makefile.in: Mostly rewritten to act as slave for toplevel
 	Makefile.
	(mostlyclean): Instead of moving gpm.ps to a funny filename and
 	then back just to delete all the gpm.?? files, let's just rename
 	it permanently to gpmdoc.ps.  Create a SECONDARY rule to avoid
 	re-TeXing after gpm.ps has been renamed.

	* configure.in (lispdir): Set as recommended by GNU standard.

	* Makefile.in (DIST): Fix typo (create-vcs => create_vcs).
	(.SUFFIXES): Clear this before setting it.  Magic suggested by the
 	Standards Info file.
	(install): Add install-strip.
	Added subdirectory recursion magic.

	* configure.in: Rewrite for new Makefile.

	* aclocal.m4: Delete useless stuff.

	* Makefile.in: Complete rewrite, now conforms to GNU standard.

	* libcurses.c: Remove the curses argument, it's quite useless.
	Check for the curses header in reasonable places.

	* configure.in: Remove the curses argument, it's quite useless.
	Check for the curses header in reasonable places.

	* README: Fix references to /usr/bin/gpm as gpm is now installed
	in /usr/sbin.
	Fix reference to /etc/gpm-root.conf as this file is now installed
	in SYSCONFDIR (== /usr/etc by default).

	* gpm-root.y (SYSTEM_CFG): Place this in SYSCONFDIR not hardcoded
	in /etc.

	* README: Fixed reference to debugging version.

Wed Aug 26 18:47:16 1998  Ian T Zimmerman  <itz@rahul.net>

	* aclocal.m4: Add ITZ_PROG_EMACS.

	* twiddler.h (TW_CUSTOM_FILE): These files now go to SYSCONFDIR
	which is the GNUsically correct way.

	* configure.in: Use LIBS (GNU standard) not LDLIBS.
	Add AC_REVISION.
	Make srcdir an absolute path.
	Figure out default prefix based on location of gcc.
	Use AC_CHECK_TYPE to look for __u32.
	Don't cache the values from command-line options.
	(REL_MAJ): This and RELEASE shouldn't be cached either, as
	Alessandro has discovered.

Thu Aug 20 00:07:42 1998  Ian T Zimmerman  <itz@rahul.net>

	* Makefile.in (all): Should now recurse into doc ...

	* doc/doc.gpm (Server Invocation): The server manpage should go to
 	section 8, not section 1.
	(gpm-root): Fixed all references in the documentation to
	/var/run/gpmctl to /dev/gpmctl.

	* doc/Makefile.in (install): Added this rule here instead of
 	installing directly from top make.  Since I care about consistency
 	so much, I must have a little mind ...
	(distclean): Don't directly rm doc/Makefile from top make,
 	instead recursively make distclean in doc subdirectory.

	* Makefile.in (distclean): Don't directly rm doc/Makefile here,
 	instead recursively make distclean in doc subdirectory.

Wed Aug 19 13:01:00 1998  Ian T Zimmerman  <itz@rahul.net>

	* gpm.c (processConn): Added code to bind the client side of the
	socket to the name of a temporary file.  This is a quick and dirty
	authentication scheme described in Stevens p. 504.

	* liblow.c (Gpm_Open): Added code to bind the client side of the
	socket to the name of a temporary file.  This is a quick and dirty
	authentication scheme described in Stevens p. 504.  Server side
	hack to follow shortly.

	* Makefile.in (distclean): Have to clean doc/Makefile as well.

	* doc/doc.gpm (gpm-root): Argh, haven't understood that the
	manpages are generated from the info source.  Added description of
	the options here.  TODO: cross-references in manpages should be in
	bold type.  Someone has to hack doc/manpager to do that.
	(Emacs Support): Deleted obsolete rubbish about t-mouse.

Thu Aug 13 13:39:23 1998  Ian T Zimmerman  <itz@rahul.net>

	* Makefile.in (install): gpm should go to /usr/sbin not /usr/bin.

	* configure.in: Enable libtinfo.

	* libcurses.c: Added weak definition of wgetch() that programs
 	which don't need the curses part of the library don't need to link
 	the whole libncurses (instead they can link with libtinfo).

	* gpm-root.y (main): restructured to use the new style logging exclusively.

	* gpm.c (main): moved file cleanup into an atexit hook.

	* gpn.c (cmdline): restructured to use the new style logging exclusively.

Thu Jul  2 00:05:16 1998  Ian T Zimmerman  <itz@rahul.net>

	* gpm-root.1: Add -D & -V to the options string. Add boldface to
 	man references.
	Update pathname for /dev/gpmctl (should this be
 	done automatically?)
	
	* gpm.1: Described new -D & -V options.  Added boldface to other
 	manpage references, to make them behave as hyperlinks where
 	software allows.

	* gpm-root.y (main): Add -D & -V to the options string.

	* configure.in: Fixed show-stopping problem of repeated calls to
 	AC_OUTPUT.

	* gpm-root.y: Replaced PERROR calls with gpm_debug_log.
	(usage): Document new flags -D and -V, in analogy to gpm flags.
	(cmdline): Implement the new options.

Wed Jul  1 09:56:30 1998  Ian T Zimmerman  <itz@rahul.net>

	* gpm-root.y: Replaced PDEBUGGGG calls with gpm_debug_log.
	(PERROR): Remove.  

	* gpm-root.y: Reindent.

	* debuglog.c (gpm_oops): Moved va_list declaration in front of
	initialization statements (I hate C!).

	* gpm.c (processRequest): Fixed debug statement to look inside
	passed Gpm_Cinfo structure.

	* tar.exclude: added sundry files created by configure.

	* Makefile.in (tar): changed tar flags to GNU style, otherwise
	options get confused with filenames.  If anybody tries to use this
	rule with a non-GNU tar, good luck.  Also added -9 flag to gzip
	to squeeze that last byte :-)

	* special.c: Replaced PDEBUG invocations.

	* mouse-test.c (_oops): Delete, it is in the library now.

	* mice.c: Replaced PDEBUG invocations with gpm_debug_log.
	Reindent.

	* libxtra.c: Reindent.
	(Gpm_GetSnapshot): Replaced PDEBUG invocations with gpm_debug_log.

	* liblow.c: Replaced PDEBUG invocations with gpm_debug_log.

	* liblow.c: Reindent.

	* gpn.c (cmdline): replaced old oops invocations.

	* wd.h (gpm_oops): Add declaration.

	* gpmInt.h: remove declaration of _oops (see above).

	* debuglog.c (gpm_oops): Added as a replacement for _oops in gpn.c.

	* configure.in: check for syslog() as well
	Added check for alloca().

	* gpn.c: removed _oops; see above.

	* gpmInt.h (_oops): Changed signature of _oops
	(oops): redefined accordingly

	* gpm.c: Removed a few leftover newlines in logging calls

	* gpn.c: Reindent, because it was so bad as to confuse my tools.
	(usage): Added description of new options -D and -V.  Used D
 	partly because it is mentioned in the string passed to getopt(),
 	but is actually unused.
	(cmdline): Added -D and -V to the getopt() loop.

	* gpm.c: Replaced ugggly macros by calls to gpm_debug_log.

	* debuglog.c (gpm_debug_log): Print a trailing newline if using
 	stdio.

	* configure.in: Added generation of doc/Makefile

	* gpn.c: Removed the Log function (now done through debuglog.c)

	* gpmInt.h: Removed debug/log related declarations (now in
 	wd.h/debuglog.c)

	* gpmCfg.h: Removed CONFIG_GPM_LOG line.

	* gpm.h: Twiddled various pathnames, included paths.h

	* Makefile.in (tar): exclude the files mentioned in tar.exclude.
	Create the archive in the current directory, not parent.

	* configure.in: Move doc/Makefile to doc/Makefile.in and make
	configure generate doc/Makefile.  Right now it doesn't substitute
	anything, but it might one day, the new way is better conformant
	with GNU distribution guidelines, and it makes it easier to create
	clean tarballs.  (There, a fine oxymoron.)

	* Makefile.in (LOBJS): Added debuglog.o
	(TARGET): Likewise.
	(depend dep): The preprocessor run needs $(CPPFLAGS).
	(DEPEND DEP): Likewise.
	(CPPFLAGS): Split from CFLAGS to correctly distribute flags to
	compilation and preprocessor rules.
	(.sh.o): needs CPPFLAGS.
	($(TARGET)): This needs LDFLAGS, not CFLAGS.
	($(TARGET).static): Likewise.
	(mouse-test): Likewise.

	* tar.exclude: create

	* wd.h: Rewritten to refer to debuglog.c.

	* debuglog.c (gpm_debug_log): Prefixed globals with gpm_ as this
	is going into the library.
	(gpm_debug_log): Reversed level test.
	(gpm_debug_level): 2 alternative definitions depending on DEBUG.

	* aclocal.m4 (DEFS): -DDEBUG should go into DEFS, not CFLAGS.
  	This would actually make a difference if we autogenerated header
 	dependencies.  Which we do, but the make rule for it is
	imprecise.  See Makefile.in entry above.
	Also, use AC_DEFINE instead of twiddling the flags by hand.
	
	* configure.in (DEFS): check for syslog.h and vsyslog().
	(DEFS): Moved definition of PREFIX here from Makefile.in.
	
	* debuglog.c (debug_log): created.

Sat Jul 25 21:08:46 1998 Edmund Grimley Evans (edmund@vocalis.com)

	* mice.c: added M_ms_plus and M_ms_plus_lr

CHANGED 9/8/98 for imwheel by Jonathan Atkins.  Diff in imwheel contains changes

Mon Jul  6 13:00:23 1998 Robin Houston (robin.houston@guardian.co.uk)

	* mev.c (main): "-p" doesn't take an argument

Sat Jul  4 14:40:09 1998  Dan Fandrich  (danf@npsnet.com)

	* mice.c (I_serial): Send only "*X" to mman

Mon Jun 15 21:31:35 1998  Pavel Machek  (pavel@ucw.cz)

	===> 1.14 released
	* mice.c: added support for Genius NetMouse

Sun May 10 11:12:46 1998  David Given (dg@tao.co.uk)

	* mice.c (M_js): joystick support

Sun May 10 10:41:58 1998  Ian Zimmermann (itz@lanminds.com)

	* t-mouse.el: partial rewrite to enhance its workings
	* mev.c: partial rewrite to fit t-mouse.el

Tue Apr 21 18:24:52 1998  Holger Jakob (jakob@ph-cip.Uni-Koeln.DE)

	* mice.c: changed all (char) to (signed char), to suit ppc Linux 

Tue Jan  6 00:31:25 1998  Alessandro Rubini  (rubini@amelia)

	* gpn.c (cmdline): -M forces -R (why didn't it?)
	* Makefile.in (YACC): Use autoconf to assign YACC

Mon Jan  5 14:48:23 1998  kw@dtek.chalmers.se (Kristian Wiklund)

	* mice.c (M_geni): new mouse type

Mon Jan  5 00:53:57 1998  Alessandro Rubini  (rubini@amelia)

	* twiddler.c: new file to support keyboard workings
	* README.twiddler: documentation about the twiddler keyboard
	* gpn.c (cmdline): opt_scale ("-r") is now limited to 100

Mon Dec 29 09:43:28 1997    (rubini@morgana)

	* mice.c (M_twid): new mouse type
	* mice.c: removed the PROTO field
	* kmouse.h: removed the file

Thu Sep 11 21:43:27 1997    Jan Daciuk (jandac@pg.gda.pl)

	===> 1.13 released
	* t-mouse.el (t-mouse-tty): another possibility for ps
	* gpm.h (GPM_XTERM_OFF): the string was buggy (James Troup)
	* gpm-root.c (get_winsize): use tty0 not console

Sun Sep  7 12:00:35 1997  Steve Bennett (s.bennett@lancaster.ac.uk)

	* mice.c: new type "-t ms3" for intellimouse devices.

Sun Sep  7 12:00:35 1997  Alessandro Rubini  (rubini@amelia)

	* gpn.c (cmdline): bugfix: accel allowed to be 1

Mon Sep  1 14:22:31 1997    (rubini@morgana)

	* mice.c: new type "-t pnp", supporting braindead ms devices.

Sun Aug 31 18:28:01 1997    (rubini@morgana)

	* gpm-root.y (f_mktty): use /sbin/mingetty instead of agetty

Fri Aug 29 08:11:32 1997    (rubini@morgana)

	* Makefile (configure): update ./configure when needed

Fri Apr 11 10:19:18 1997    (rubini@morgana)

	* mev.c (main): Ctrl-C terminates the program
	* doc/doc.gpm: revision of the docs. 

Tue Apr  8 17:06:06 1997    (Dave Flater: dave@universe.digex.net)

	* mice.c (M_ms): removed a typo in the disabled decoder
	* FAQ: Marked the change in the decoder for ms-middle

Tue Apr  8 16:57:17 1997    (Ian Zimmermann: itz@rahul.net)

	* gpm.c: fix to handle the whole stack of clients
	* liblow.c (Gpm_Open): support for SIGTSTP

Tue Apr  8 01:27:09 1997    (Thomas E. Dickey: dickey@clark.net)

	* aclocal.m4, configure.in: added support for autoconf

Tue Apr  8 00:48:46 1997    (rubini@morgana)

	* doc/doc.gpm (Mouse Types): fixed the list
	* gpmCfg.h (MAX_VC): raised from 32 to 64 (needed by some users)
	* README: Added a check for "window-system" in the .emacs clause

Tue Apr  8 00:35:33 1997    (Pavel Machek: pavel@Elf.mj.gts.cz)

	* mice.c (M_logimsc): new mode needed for some mice

Tue Apr  8 00:13:30 1997    (Matthias Grimrath: y0001032@rzserv8.rz.tu-bs.de)

	* gpm.h: check for c++ inclusion

Tue Apr  8 00:10:33 1997    (Stephen Tell: tell@cs.unc.edu)

	* mice.c (M_ps2): patch to support some strange devices

Tue Aug 27 15:54:10 1996    (rubini@morgana)

	* gpmCfg.h (SELECT_TIME): increased, to avoid possible hd spinup.

Fri Jul 26 12:33:39 1996    (rubini@morgana)

	===> 1.10 released
	* FAQ: new file
	* doc/doc.gpm: provided docs about special commands

Thu Jul 25 09:38:29 1996    (jrt@miel.demon.co.uk)

	* gpm-root.y (main): added a missing arg in a message

Wed Jul 10 15:59:33 1996    (rubini@morgana)

	* liblow.c (Gpm_Open): fixed a fault when stdin not a tty (John Davis)
	* mice.c (I_wacom): fixed initialization
	described the fields of Gpm_Type, in order to easily add new types
	* special.c: new file, based on an idea by Karsten Ballueder
	* gpm.c (processMouse): use a different "repeater" (Markus)

Sun May  5 22:01:48 1996    (rubini@morgana)

	===> 1.09 released
	* gpmCfg.h (DEF_GLIDEPOINT_TAP): changed to 0. The tapping code
	for ps2 mice must be disabled for non glidepoint protocols.

Tue Apr 23 14:05:43 1996    (rubini@morgana)

	===> 1.08 released
	* mice.c: support glidepoint ps2 mice (suggested by Don Schwarz)

Mon Apr 22 15:20:05 1996    (rubini@morgana)

	* mice.c: added the wacom protocol

Sun Apr 14 20:59:08 1996    (rubini@morgana)

	* t-mouse.el (t-mouse-tty): comply with procps-0.99

Sun Feb 11 01:05:07 1996    (rubini@morgana)

	* Makefile (MAN): libgpm.so.1 was installed. Fixed.

	* gpn.c: try to kill even if connection fails
	(usage): removed the line about "-P" (dropped time ago)
	(cmdline): added a missing "break": -o forced -p.

	* gpmInt.h: new globals: opt_kernel, opt_explicittype

	* gpn.c (cmdline): support for "-K" (kernel module)

Fri Jan 12 00:31:40 1996    (Michael_Plass.PARC@xerox.com)

	* mice.c (M_ms): allow motion and button change in a single event
	   however, some mice won't work, so the code is still "#if 0"

Tue Nov 14 14:18:26 1995    <rubini@godipmec.unipv.it>

	===> 1.06 released, unexpectedly
	* mouse-test.c: added missing variable to link with mice.c
	(main): don't use absolute pointing devices as possible mice.

Mon Nov 13 11:21:42 1995    (Marc Meis 100334.1426@compuserve.com)

	===> 1.05 released
	* gpmInt.h: added flag "absolute" to Gpm_Type to indicate
	absolute pointing device
	* mice.c: added NCR3125 pen support
	* gpm.c: modified processMouse to support absolute pointing device


Sat Nov 11 11:40:27 1995    (rubini@morgana)

	* t-mouse.el (t-mouse-run): fixed bug in linux-to-linux remote link.
	* doc/doc.gpm (gpm-root): documented the need for /dev/vcso
	* gpm-root.y (main): check for /dev/vcs0 before starting
	Only run with /dev/vcs -- disregard kernels 1.1.73-92
	(scr_dump, scr_restore): if open(/dev/vcs) fails, write on syslog

	* gpm.c (processMouse): only one margin is active at a time.

Fri Nov 10 22:37:45 1995    (rubini@morgana)

	* gpm.h (Gpm_FitEvent): bug fix: fit DRAG, not MOVE.
	* gpm-root.y (main): only ctrl-move gets the default behaviour

Fri Nov 09 20:33:26 1995    (lou@lounix4.conc.tdsnet.com)

	* mice.c (M_mman): added support for glidepoint mice (Lou Sortman)

Wed Sep 27 09:51:17 1995    (rubini@morgana)

	===> 1.04 released
	* Makefile: added a dirty check for __u32 

Wed Sep 26 01:31:13 1995    (rubini@morgana)

	* doc/doc.gpm: some fixes
	* doc/gpmdoc.txt: new file: ascii for "less" readers.

Sat Sep 23 08:44:01 1995    (rubini@merlino)

	* doc/doc.gpm: updated the lisp form to load t-mouse
	* README: added instructions on compiling and installing
	* Makefile: minor fixes 

Sun Sep 17 13:35:26 1995    (rubini@morgana)

	===> 1.03 released
	* doc/doc.gpm (Variables ): gpm_mx and gpm_my are documented.
	* gpm.h (Gpm_FitEvent): this too was buggy, when dealing with motion.
	* gpm.c (do_client): masked with GPM_BARE_EVENTS (long-standing bug)
	(processMouse): removed GPM_SINGLE from motion events
	* libxtra.c (Gpm_GetSnapshot): return 0 if the input queue is full
	* mev.c (main): another fflush, I forgot about it.
	* liblow.c (Gpm_Open): the fix of two days ago broke normal clients.

Sat Sep 16 22:47:13 1995    (rubini@morgana)

	* gpn.c (cmdline): added a missing "break;" (-L)

Fri Sep 15 09:40:59 1995    (rubini@morgana)

	* liblow.c (Gpm_Open): fixed a problem whit pty's (t-mouse broke)

Thu Sep 14 01:50:23 1995    (rubini@morgana)

	===> 1.02 released
	* liblow.c (Gpm_Open): fixed a bug with multiple opens
	* mev.c (user_handler): added fflush() 'cause didn't work with elf/lib5
	* gpmInt.h (inline): inhibit inlining when debugging
	* Makefile (install): some fixes to make it work (Mark Shadley)
	(static): new target (I've an old gdb :-( )
	* hltest.c (xhandler): fixed a problem with messages
	* libhigh.c (Gpm_HandleRoi): fixed a segfault when no roi matches
	* gpm.c (get_console_size): create a different resolution for x and y
	(processMouse): return tiny motions (0,0) if some time has elapsed
	* gpn.c (cmdline): introduced "-q". Obsoletes "-L".
	* gpmCfg.h (SELECT_TIME): decuplicated time-out (10 minutes).
	(MAX_VC): 32 virtual consoles

Tue Sep 12 00:12:47 1995    (rubini@morgana)

	===> 1.01 released
	* Makefile: supports elf shlibs (Mark Shadley)

Wed Sep  6 22:46:55 1995    (rubini@morgana)

	* mev.c (main): use current tty instead of /dev/console
	* liblow.c (Gpm_Open): use /dev/ttyxx instead
	  of /dev/console (Pavel Kankovsky)
	* gpn.c (cmdline): fixed "-o rts" 

Wed Jul 26 14:44:59 1995  Alessandro Rubini  <rubini@ipvvis>

	===> 1.00 released
	* doc/doc.gpm (Emacs Support): upgrade
	* t-mouse.el: upgrade by Ian T Zimmerman
	* gpm-root.y (cmdline): last bug fix: missing break in switch

Wed Jul 26 00:38:51 1995    (rubini@morgana)

	* t-mouse.el: fixes and enhancements by Ian T Zimmerman
	* Announce: new file. The announce for c.o.l.a.
	* liblow.c (Gpm_Open): open console O_WRONLY instead of O_RDONLY
	(Gpm_Getc): swapped two lines, to handle SIGWINCH in clients

	* doc/doc.gpm: final fix

Tue Jul 25 00:41:44 1995    (rubini@morgana)

	* mouse-test.c: a new utility to "detect" your mouse type

Mon Jul 24 15:52:11 1995    (rubini@morgana)

	* gpm.c (gpm_killed): removed spurious and annoying message

Fri Jul 21 10:57:07 1995    (rubini@morgana)

	* libhigh.c (Gpm_HandleRoi): fixed a problem with gpm_visiblepointer

Thu Jul 20 23:04:42 1995    (rubini@morgana)

	* gpn.c (cmdline): "-D" removed (use "-k" instead)

Wed Jul 19 09:12:29 1995    (rubini@morgana)

	* gpm.c (processRequest): fixed nasty bug introduced with "-R"
	* t-mouse.el (t-mouse-keymap): added double-drag and triple-drag
	* gpm.c (processMouse): the GPM_SINGLE,GPM_DOUBLE and GPM_TRIPLE
	flags are always delivered (as they should have always been).
	* gpn.c (cmdline): removed the "-p" option (useless).
	* doc/doc.gpm: Revised and updated.
	* liblow.c (Gpm_Open): use the foreseeable device /dev/gmctl
	* gpn.c (cmdline): no more symlink /tmp/gpmctl
	* gpm-root.y (main): recursive menus definitely fixed

Sat Jul 15 11:16:18 1995    (rubini@morgana)

	* various places: applied axp patches. 
	* libhigh.c: update, at last.
	* hltest.c: new file, to test the high level library

Mon Jul 10 20:03:46 1995    (rubini@morgana)

	* gpm-root.y (main): recursive menus
	* gpm.c (main): chmod 777 /var/run/gpmctl (1.2.9 and later broke it)
	* mice.c (M_mman): chord-middle detected

Wed Jul  5 09:29:52 1995    (rubini@morgana)

	===> 0.99 released
	* doc/Makefile: minor cleanup
	* doc/manpager: changed any "continue" to "next"

Fri Jun 30 09:14:13 1995    (rubini@morgana)

	* gpm.c (wait_text): fixed interaction with X when non repeater.

Wed May 31 09:17:55 1995    (rubini@morgana)

	===> 0.98.3 released
	* gpm.c (processMouse): fixed update of number of buttons
	* doc/doc.gpm: minor fixes
	* libxtra.c (Gpm_GetSnapshot): check if a connection is open

Sun May 28 22:02:20 1995    (rubini@morgana)

	* gpm-root.y (main): reset gpm_tried after failure

Tue May 16 10:28:01 1995    (rubini@morgana)

	* gpm.c: setuid(0)
	* gpm-root.y: setuid(0)

Tue Apr 25 18:02:03 1995    (rubini@morgana)

	* t-mouse.el (t-mouse-make-event): fixed problem with split windows

Sat Apr 22 13:19:50 1995    (rubini@morgana)

	* t-mouse.el: mouse-2 in Info-mode is handled
	mouse-2 in dired-mode is handled
	* gpm-root.y (main): added  sleep() if Gpm_Open() fails

Wed Apr 12 00:29:46 1995    (rubini@morgana)

	===> 0.98.2 available
	* mev.c (main): use the new library functions
	* gpm.c (processRequest): added information retrivial
	* liblow.c (Gpm_Open): try /tmp/gpmctl as well
	* libxtra.c (Gpm_GetLibVersion): new function
	(Gpm_GetServerVersion): new function
	(Gpm_GetSnapshot): new function
	* gpm-root.y (postmenu): use /dev/vcs and fine line-drawing chars
	(main): change of semantic: use button-down to choose item.
	(tableMod[]): use KG_SHIFT	
	(main): logging out is correctly managed (in /etc/utmp)

Tue Apr 11 15:48:27 1995    (rubini@morgana)

	* t-mouse.el (t-mouse-process-filter): catch errors in mev messages
	* mice.c (I_serial): mouseMan fix (Stephen Lee)
	(mice[]): added an information string and a synonym.
	(M_listTypes): new function to give some help
	* mev.c (user_handler): added "0x" in printf (ajohnson)
	(tableMod[]): use keyboard.h instead of constant values.
	* gpm-root.y: added missing ';'s (ajohnson@apanix.apana.org.au)
	(postmenu): fixed "unsigned char" bug (johnvl@einstein.et.tudelft.nl)
	* Makefile (install): fixed problems with info file

Thu Apr  6 00:33:03 1995    (rubini@morgana)

	===> 0.98.1 made available
	* gpm.c gpn.c: read two mice at a time
	* gpm-root.y (f_jptty): added ioctl(VT_WAITACTIVE)
	(main): console disallocation is provided

Wed Apr  5 08:03:37 1995    (rubini@morgana)

	* gpm.c (wait_text): only invoked if already graphics
	(processMouse): be a repeater if we're in graphic mode
	* gpn.c (check_uniqueness): fixed bug with "-k" and nobody to kill
	* gpm-root.y (scr_dump): fixed bug with clr_selection and dump/restore
	* wd.h: changed DEBUG(()) to PDEBUG(())

Tue Apr  4 08:56:33 1995    (rubini@morgana)

	* gpn.c (usage): changed semantics, to be more useful

Sun Apr  2 13:20:28 1995    (rubini@merlino)

	* gpn.c (cmdline): device is opened with O_NDELAY

Mon Mar  6 09:49:19 1995    (rubini@morgana)

	===> 0.98 released
	* Makefile (info): new target: no more info file from simple "make"
	* gpn.c (loadlut): use octal codes and latin-1 chars (Andries Brouwer)
	(cmdline): create a symlink /tmp/gpmctl to run precompiled executables

Sat Mar  4 18:44:54 1995    (rubini@morgana)

	* t-mouse.el (t-mouse-keymap): use M-mouse to enable the scrollbar

Tue Feb 28 00:20:35 1995    (rubini@morgana)

	* mev.c (main): added report of kdb modifiers for key events
	* Makefile (clean): remove *.1, since they're not sources
	* kernel/*: removed the directory as obsolete
	* root.y (scr_dump): added support for /dev/vcs

Sun Feb 26 14:25:31 1995    (rubini@morgana)

	* wd.h: when intactive use "" instead of "0" (to use "-Wall")
	* various fixes to face "-Wall" messages

Mon Feb 20 18:08:02 1995    (rubini@morgana)

	* liblow.c (Gpm_Open): bug fix, closed an open comment (John E. Davis)

Wed Feb 15 12:25:14 1995    (rubini@morgana)

	* mev.c (sigHandler): propagate SIGWINCH to the father
	* gpm.c (main): propagate SIGWINCH to the clients
	* gpmInt.h: modified Gpm_Type; one more field to allow extra bytes
	* mice.c (I_serial): changed prototype to autodetect mouseman
	(M_mman): new mouse type for damned MouseMan.

Tue Feb 14 00:33:34 1995    (rubini@morgana)

	* gpn.c (cmdline): modem-line toggle added

Thu Feb  9 00:42:51 1995    (rubini@morgana)

	* root.y (main): the "-B" option is no longer supported
	(scr_restore): be sensitive to SIGWINCH

Wed Feb  8 11:46:52 1995    (rubini@morgana)

	* gpm.c (main): added support for SIGWINCH

Thu Jan 26 12:11:59 1995    (rubini@morgana)

	===> 0.97 released
	* root.y (postmenu): get color to restore from the cursor cell instead
	of the previous one (it failed with `less')
	(yylex): allow double quotes in strings
	(main): go to background
	* gpn.c (cmdline): go to background and setsid() (Stefan)
	(cmdline): new switch -k (kill) for busmouse users
	* gpm.h (GPM_NODE_DIR): use "/var/run" instead of "/tmp" (Larry)
	* Makefile: use "-O3" and don't release binaries any more
	* mice.c (I_logi): don't jump to I_serial anymore (Larry)

Fri Jan 13 12:57:41 1995    (rubini@morgana)

	===> 0.97alpha releasd
	* gpm.c (gpm_killed): use signal SIGUSR1 to kill an existent gpm
	(main): removed the annoting warning in signal()

Thu Jan 12 11:45:08 1995    (rubini@morgana)

	* mev.c: added "any" to event list
	* t-mouse.el (t-mouse-suspend): fixed behaviour (thanks Larry)
	* mice.c (I_serial): flushed pending input (thanks Miguel)
	(I_logi): fixed "howmany" problem with serial vs. busmice
	* gpn.c (cmdline): m_type is passed to m_type->init
	* gpmInt.h: removed unused field "device" in Gpm_Type

Mon Dec 26 01:21:54 1994    (rubini@morgana)

	* gpn.c (cmdline): authentication moved here.

Wed Dec 21 11:28:41 1994    (rubini@morgana)

	* gpm.c (main): new (smarter) authentication

Sun Dec 18 11:18:40 1994    (rubini@morgana)

	===> gpm-0.96 released
	* root.y: allright and tested
	* gpn.c: new '-P' option
	* doc/manpager: new file to generate man pages
	* Makefile: fixed the installation stuff, use /usr as $(PREFIX).
	* gpn.c, mice.c: removed the last #ifndef linux
	* getopt.[ch]: removed from the distribution

Fri Dec 16 17:09:19 1994    (rubini@morgana)

	* doc/doc.gpm: general update, addition of gpm-root docs.

Thu Dec  1 01:03:51 1994    (rubini@morgana)

	* liblow.c (Gpm_Open): fixed console detection (thanks Miguel)
	and removed unused ttyp* and ttyq* recognition

Fri Nov 25 01:37:37 1994    (rubini@morgana)

	* root.c (main): new file: a tentative default handler
	* gpm.c (do_client): bug-fix: default handlers now can run
	* gpm.h: new struct and prototypes for libhigh.c
	* doc/doc.gpm: minor fixes
	* libhigh.c: new file with "Region-of-Interest" management

Fri Nov 25 01:24:45 1994    laurent@brasil.frmug.fr.net (Laurent Chemla)

	* gpm.c (processMouse): fix for button-up reporting

Tue Nov 22 19:54:54 1994    (rubini@morgana)

	* gpm.c (do_selection): draw a running pointer or none at all

Thu Nov 17 00:24:13 1994    (rubini@morgana)

	===> gpm-0.93 released
	* t-mouse.el (tm-scroll-jump): changed behaviour: doesn't jump
	* doc/doc.gpm (Emacs Support): documented new word and line facilities
	* liblow.c (gpm_convert_event): fix: use gpm_zerobased
	(gpm_convert_event): fix: report all buttons on button-up

Wed Nov 16 21:15:15 1994    (rubini@morgana)

	* t-mouse.el: tm-word and tm-line implemented

Thu Nov 10 15:27:07 1994    (rubini@godipmec)

	===> gpm-0.92 released
	* gpmInt.h (Gpm_Type): new field "howmany"
	* mice.c: fixed behaviour for "-t bm" (thanks to Reuben Sumner)
	* gpm.c (getMouseData): fixed data reading

Sun Nov  6 13:18:34 1994    (rubini@morgana)

	* removed all the outdated "#ifdef linux"

Sat Nov  5 10:23:30 1994    (rubini@morgana)

	* Makefile (all): executables are stripped

Tue Nov  1 13:10:48 1994    (rick@razorback.brisnet.org.au)

	* gpm.c (check_mode): fix in mouse re-openig

Sun Oct 30 01:49:14 1994  Alessandro Rubini  (rubini@coroide)

	===> gpm-0.9 released
	* doc/infofilter: fixed indentation of lisp code

Sat Oct 29 17:24:54 1994    (rubini@morgana)

	* sample/configure.in: shows how to setup for both gpm and xterm
	* sample/rmev.c: new file, to test xterm compatibility 
	* mev.c (usage): fixed ugly bug in printf()
	(main): added initscr() to work under sunos-hpux

Fri Oct 28 13:19:37 1994    (rubini@godipmec)

	* gpm-xterm.c, gpm-xterm.h: new files, for xterm applications
	* mev.c (main): fixed ascii value reporting

Thu Oct 27 01:56:13 1994    (rubini@morgana)

	* libcurses.c: xterm support is there as well

Wed Oct 26 21:07:11 1994    (rubini@morgana)

	* liblow.c: xterm support is there

Tue Oct 25 00:32:36 1994    (rubini@morgana)

	* libcurses.c (Gpm_Wgetch): new file, single function

	* gpm.h (Gpm_Getchar): removed a spurious "eject" string

Mon Oct 24 22:47:37 1994    (rubini@morgana)

	* liblow.c (Gpm_Repeat): new function
	(Gpm_Open): uses gpm_consolefd
	* doc/doc.gpm: updated info, created indexes

Sun Oct 16 11:36:07 1994    (rubini@morgana)

	==> gpm-0.17 released
	* gpm.c (processConn): give the client a motion event at connection
	* t-mouse.el (t-mouse-run): added interactive to restart mouse
	(t-mouse-tty): bug fix: the -.16 one won't work on most machines

Thu Oct 13 09:48:07 1994    (rubini@morgana)

        ===> gpm-0.16 released
	* gpm.c (main): a better check for uniqueness
	(gpm_killed): a signal handler to remove /tmp/gpm* on exit

Wed Oct 12 16:10:26 1994    (rubini@godipmec)

	* Makefile (all): the info file is generated
	* t-mouse.el (t-mouse-tty): procps 0.95 is recognized

Fri Sep 23 12:26:37 1994    (rubini@morgana)

        ===> gpm-0.15 released
	* t-mouse.el (t-mouse-keymap): the scroll-bar is there.
	* mev.c: new flag "-p" to show the pointer during drags

Thu Sep 22 13:15:22 1994  Alessandro Rubini  (rubini@ipvvis)

        ===> gpm-0.14 released
	* doc/doc.gpm: fixed the section about emacs

Fri Sep 23 12:26:37 1994    (rubini@morgana)

	* t-mouse.el (t-mouse-keymap): the scroll-bar is there.

Mon Sep 19 01:06:56 1994    (rubini@morgana)

	* mev.c: new flag "-f" to fit events.
	* Makefile (quickinst): new target
	gpm.info is installed as well
	(distrib): the precompiled binaries are distributed as well
	* t-mouse.el: it is really full-featured, now.

Sun Sep 18 11:44:36 1994    (rubini@morgana)

	* mev.c (emacs_handler): up events now return the press location
	* Makefile: binaries are now released with the package

Wed Sep 14 14:12:48 1994    (rubini@morgana)

	* t-mouse.el: removed free variables: clean compile
	* Makefile: t-mouse.elc is properly compiled
	* exec.el: used to compile elisp.
	* doc/*: texinfo is used, and online docs are available

Wed Sep  7 19:34:28 1994    (rubini@morgana)

        ===> gpm-0.13 released
	* gpmCfg.h (MAX_VC): manage 16 virtual consoles
	* t-mouse.el: avoid invoking awk on startup
	* mev.c (cmdline): accept hex numbers with "-C" 

Sat Aug 27 11:41:57 1994    (rubini@morgana)

	* t-mouse.el: added the "-s" option to 'mev', to prevent error msg.

Mon Aug 22 20:28:47 1994    (rubini@morgana)

	* liblow.c (Gpm_GetEvent): Closing connection correctly handled.
	(Gpm_Close): fix to help the above handling

Fri Aug 19 10:23:01 1994    (rubini@morgana)

	===> gpm-0.12 released
	* gpn.c (cmdline): new options: "-2" and "-3"
	* doc/gpm.tex: (Emacs Support) new section

Thu Aug 18 10:01:22 1994    (rubini@morgana.unipv.it)

	* gpm.c (do_selection): use right button to extend selection,
	if the middle one is present
	* t-mouse.el: recognizes the controlling terminal
	* doc/gpm.tex (Compiling and Installing) note about rc.local
	* gpm.c (processMouse): fix: first click is single

Wed Aug 17 09:20:54 1994    (Olav Woelfelschneider)

	===> gpm-0.11 released
	* gpmInt.h: changed "char proto[]" to unsigned

Tue Aug  2 05:26:55 1994    (rubini@morgana)

	===> gpm-0.10 released
	* liblow.c (Gpm_FitValuesM): bug fix

Mon Aug  1 19:22:50 1994    (rubini@morgana)

	* gpn.c (cmdline): fixed bug with opt_device (Miguel de Icaza)
	* liblow.c (Gpm_Close): fixed the gpm_tried bug
	* gpm.c (processMouse): clustering added
	(getMouseData): new function to ease readability

Fri Jul 29 13:54:38 1994    (rubini@ipvvis)

	* gpn.c: moved many functions to gpm.c to make them inline

Fri Jul 29 01:56:39 1994    (rubini@morgana)

	* gpm.h: Gpm_DrawPointer: new facility
	* liblow.c: gpm_zerobased added
	* gpn.c (do_client): added an option to map "drag" to "move" (ugly)

Thu Jul 28 23:55:18 1994    (rubini@coroide)

	===> gpm-0.06 released
	* doc/gpm.tex: updated

Thu Jul 28 11:02:48 1994    (rubini@morgana)

	* gpn.c (cmdline): configurable button order
	(cmdline): double click time, too.
	(cmdline): "-L" option.
	(do_selection): fixed bug about wrap-around.

Wed Jul 27 23:56:14 1994    (rubini@morgana)

	* liblow.c: no more public gpm_timeout
	* gpn.c (do_client): bug fix

Tue Jul 26 14:44:37 1994    (rubini@mimmo)

	* gpm.h: ctl file now in /tmp 

Sat Jul 23 13:22:29 1994    (rubini@morgana)

	* gpn.c (cmdline): /dev/mouse is the default
	* gpm.c (processMouse): button-up: report which one

Fri Jul 22 15:57:35 1994    (rubini@morgana)

	* mev.c: emacs syntax is there

Wed Jul 20 19:35:16 1994    (rubini@morgana)

	* gpm.h: added GPM_HARD
	* gpn.c (do_client): fixed default management.

Tue Jul 19 16:31:23 1994    (rubini@morgana)

	* gpm.h: changed Gpm_Event to fit ptys

Tue Jul 19 09:21:37 1994    (rubini@morgana)

	===> gpm-0.04 released
	* gpn.c (cmdline): use "/dev/mouse" as a second trial for the name
	* gpm.c (processRequest): can override info
	* mev.c (interact): new function to test stacking

Mon Jul 18 11:44:13 1994    (rubini@morgana)

	* liblow.c (Gpm_Open): added stacking and an ugly goto
	* gpn.c (_oops): fixed bug about errno (Steven S. Dick)
	* doc/gpm.tex: fixed the terrible language, hopefully.

Fri Jul 15 18:21:08 1994    (rubini@merlino)

	* mev.c: raw stdin added

Tue Jul 12 16:57:45 1994    (rubini@ipvvis)

        ===> gpm-0.03 released


/* Local Variables: */
/* tab-width:8      */
/* End:             */

 
