Sophie

Sophie

distrib > Mandriva > 8.1 > i586 > by-pkgid > 26d880fac21e1bb21975e5b57d57795b > files > 3

ku-1.1BETA5-4mdk.i586.rpm

CHANGELOG for KU
================

VERSION 1.0
===========

-- Time variable (%T) added to the kill message.
-- Fixed an option (-m) that did not allow the default kill message
   to be displayed yet required an option.
-- The look of the kill message has changed, now displays the hostname
   and tty to the user of who is doing the killing.
-- Removed some misc non-needed code and cleaned up a few things.


VERSION 1.1BETA1
================

-- Removed gethostname(), now uses utsname.h
-- Fixed a possible security hole.


VERSION 1.1BETA2
================

-- Fixed a bug in --message that I created.
-- License changed to GPL version 2 and below (your choice)
-- A sleep option (-S) (--sleep) had been added.  This will make ku
   sleep in the background for a given amount of time before proceeding.
   The sleep suffix may be (s) for seconds, (m) for minutes, (h) for
   hours, or (d) for days.  An example: ku -tmvS 1h.
-- Release untested.


VERSION 1.1BETA3
================

-- Bug fixes


VERSION 1.1BETA4
================

-- Added a enviromental variable for the default KILL message, no longer
   needs a recompile to be changed, look in /etc/ku.msg to edit it after
   installation.
-- Added a %L (login time) Variable to use with the kill message.
-- Changed --sleep to --wait.
-- Other misc bug fixes and code added including documentation changes.


VERSION 1.1.BETA5
=================

* Done by Nathan Benson
  (i've probably added 10 times more bugs than i fixed, but i hope not :)


-- * Added an autoconf configure script 
-- * Added a TODO file
-- * Added a true enviromental variable (KU_MESSAGE) for the kill message

     e.g. - # export KU_MESSAGE="%l sucks and %Y rules"

-- * Cleaned up the code so it would compile clean with -Wall

     - * wait.c: Changed sleepTime() prototype to 'int' rather than 'const char'
         which cleaned up a compiler warning.

     - * wait.c: Fixed an incorrect escape sequence (\'), which cleaned up
         a compiler warning.

     - * main.c: Changed 'args.S_arg=optarg;' to int 'args.S_arg=atoi(optarg);'
         which cleared up another compiler warning.

-- * Did some commenting, code clean up, and modified how -K (--kill) is
     done (just added a check for the enviromental variable KU_MESSAGE,
     and added in support for reading /etc/ku.msg for the kill message.
     see source/main.c).

-- * Changed terminate.c kill signal to SIGHUP first, rather than SIGKILL.
     This was done because some shells didn't clean up the utmp entry when
     they got a KILL signal.  So we give them a hangup (SIGHUP) first, and
     hope they clean up, and exit cleanly.
     
-- * Fixed a bug in message.c, that would cause ku to segfault if you used
     the -m (--message) flag, and did NOT specify one on the command line.
     i.e. - ku was trying to use the default kill message, yet trying to
     get the length of the command line kill message (which didn't exist).

-- * Fixed another bug in message.c, that would try and open /etc/ku.msg
     to read the kill message from. if it couldn't open it, the code would 
     acknowledge that, and set the kill message to the default compiled
     in one (DKILL).  BUT, it would then try and close the file pointer
     that was not open, causing segfaults.  fclose(fp2) was just moved
     inside the else statement.

-- * Added the use of conf.h, rather than dynamically creating the
     include/ku.h every time, the way the old configure script did.

    - * options now include:
	--disable-header
	--with-message='%l is a loser so says %Y'

-- * Updated documentation:
	- * CHANGELOG
	- * README
	- * TODO
	- * doc/CHANGELOG
	- * doc/README
	- * doc/TODO
	- * doc/ku.1
	- * doc/ku.1.gz
	- * doc/ku.lsm