--------------------------------------------------------------------------
 C h a n g l o g
--------------------------------------------------------------------------

 Fri Jul  9 17:58:54 CEST 1999

 + ChangeLog started
 + Moved the documentation into separate files
 * Fixing the logexist () with code from Hallvar Helleseth

 Sun Jul 11 19:27:45 CEST 1999
 
 + We now format the logs authenticate () outputs with awk
 + Satisfying the dull. adding support for xterm.
 + Using Hallvar Helleseth's default color and nexus as the default font.
 + Also added from Hallvar that the cursor color is the same as the text color

 Tue Jul 13 06:52:35 CEST 1999

 * Logexist () fixed so it outputs all files that are missing simultaneously
 + Hallvar used eesh (which is E specific) Therefore I made a getreso ()
   using xdpyinfo (to find out the current screen resolution) which 
   ensures compatibility with other window managers. 

 Tue Jul 13 10:58:32 CEST 1999

 * Whoops! fixing the bug with getreso () and a 4 digit horizontal digit 

 Thu Jul 15 02:06:50 CEST 1999

 + Made a getfontsize ()
 * Cleaned up the docs a bit
 + Added some of Outcasts great fonts and made a README

 Thu Jul 15 10:28:14 CEST 1999
 * Discovered a bug in () authenticate which had been there ALL along. 
 * Put stuff from authenticate into access ()
 + authenticate () now has 3 retries only
 + Added verbose option

 Fri Jul 16 11:16:43 CEST 1999
 + Adding support for rxvt
 * Got around a problem that had me _really_ freaked in tailer ()
   stupid me forgot to export TRML in whichterm ()

 Sat Jul 17 13:36:16 CEST 1999
 * Wow! I now see there is actually a hope for me to
   get the geometry code finished..
 * Sheez, am I on crack? I had screwed up earlier and got the 
   horizontal and vertical in getreso () backwards. Fixed now

 Sat Jul 17 20:37:46 CEST 1999
 + Finished with the fcsking tedious and boring $PLACEMENT routine

 Mon Jul 19 13:36:21 CEST 1999
 * Now we are getting somewhere :-) 
 + Finished with the geometry stuff, now doing some cleanup in the script

 Thu Jul 29 20:23:27 CEST 1999
 + Using awk in getfontsize and getreso

 Tue Aug  3 03:28:47 CEST 1999
 * Just fixing some docs, version now 0.79

 Tue Sep  7 18:35:08 CEST 1999
 * Paralogger version 0.80 - first public release

 Thu Sep  9 00:54:15 CEST 1999
 * Fixed a Missing $.
 * _Hopefully_ fixed a bug in the authenticate function where the tailer 
   function would not be passed on if the root user didn't have /bin/sh
   as the default shell. I investigated this further to find out there was
   a bug the other way around too, e.g. if the user running paralogger didn't 
   have /bin/sh as the default shell

 Thu Sep  9 23:23:50 CEST 1999
 - Removed the -p in authenticate since it was not necessary

 Fri Sep 10 16:24:41 CEST 1999
 * Changed the variables HOFFSET to VOFFSET and vice versa

 Sun Sep 12 21:08:54 CEST 1999
 + Added a new error message if no logs are specified
 + Added kates font
 + Made debug messages more informative 

 Mon Sep 13 22:23:53 CEST 1999
 + Added the reverse () function

 Tue Sep 14 01:22:09 CEST 1999
 * Now we start cascading from bottom to top automatically in horizontal
   cascade mode if the base placement mode is somewhere at the bottom of 
   the screen
 * Also in vertical cascade mode we now start cascading from right to left
   if the base placement mode is somewhere to the right of the screen
 + Added some values to revert to if getfontsize () or getreso () fails 

 Sat Sep 18 23:05:45 CEST 1999
 * Workaround for buggy xlsfonts (?) in getfontsize (). It seems that the 
   xlsfonts that come with some versions of Xfree86 likes to output the values
   for PIXEL_SIZE and QUAD_WIDTH twice, making paralogger puke like a drunken
   sailor

 Sun Sep 19 04:48:01 CEST 1999
 + Added a parser and a help option. We now support all the placement modes
   as command line switches. eg: paralogger -tl and a -v for verbose mode along
   with -single, -vcascade and -hcascade
 * Updating the docs a little
 * paralogger version 0.87 released

 Wed Sep 22 23:12:52 CEST 1999
 * Removed some more fluff and unneeded stuff
 * Deciding to only accept input from the parser, eliminating the need to edit
   The config in the script directly. I think this is the way Hallvar wanted it
   to behave. I like this new behavior better too

 Sun Oct 31 06:18:30 CET 1999
 + Made a man page
 * Fixes to the html docs
 * use --follow=name instead of -f for tailing logs since -f is no good if
   the logs are rotated while tailing

 Sat Nov  6 04:43:05 CET 1999
 + Fonts back in again, since we have a way to get it reliable

 Wed Nov 17 22:25:22 CET 1999
 * Small fix in tailer () and Makefile

 Fri Nov 26 01:18:37 CET 1999
 * Moved redundant stuff in tailer () to runterm () to make it easier to
   maintain and made a generate stand alone shell script mode.
 * Long options -hcascade and -vcascade shortened to -hcas and -vcas 

 Fri Dec  3 04:17:24 CET 1999
 * Changed the background color on the html docs.

 Sat Dec  4 05:53:37 CET 1999
 + We are now trapping and waiting for the terminals so they are killed
   when the script is killed

 Mon Dec  6 00:38:12 CET 1999
 + New verbose function so we don't have to have to use an if sentence when 
   verbose

 Mon Dec  6 21:45:02 CET 1999
 * Finally getting some mixed up variable names right.  

 Thu Dec 16 05:09:21 CET 1999
 * Fix-ups in testsuite

 Fri Dec 17 11:16:40 CET 1999
 * Documentation fixes, uninstall in Makefile, made configure script to find 
   dependencies, made paralogger.spec so we can build RPMS

 Wed Dec 22 22:26:58 CET 1999
 + Simple configure script to check some dependencies
 * Finally got the winsize () dynamic which means we can use all combinations
   of fonts and sizes on the terminals
 * Split cascade () into plcment () and cascade ()
 * Rewrote cascade () to get rid of some redundancies. Also starts a new row if 
   there is no more space left. Probably isn't perfect but...
 * Split reverse () into cntlogs () and reverse ()
 - Now use tac in reverse () routine which should be faster
 * Split runterm () into trapids () and runterm ()
 + Added new -nowait option

 Mon Jan  3 17:35:58 CET 2000
 + Should work with FreeBSD, added a chktail () that select tail based on OS
 + Back in with the old reverse () since tac can't be found on FreeBSD
 * Fixed some problems with unexported variables
 + Added a check for $DISPLAY in case ppl try to run paralogger while not in X
 
 Wed Jan 26 18:30:46 CET 2000
 + Added some meta information in index.html
 * RPM spec file is now noarch as it should

 Tue Feb  1 20:09:23 CET 2000
 * Paralogger version 0.88 released 

 Sun Jan  9 23:31:00 CET 2000
 * Rewrote verbose () to get the "verbose && echo foo" out of my face
 + Made a err () to make it easier to redirect errors to standard error
 * Using more local variables, misc cleanups
 
 Tue Feb 15 22:33:09 CET 2000
 * Fixing the trap signals
 + Added code to detect if tail is from textutils 2.x or 1.x. For some odd
   reason, slackware 7.0 is shipping with 1.x
 + Now checking if the Eterm started for calibrating the windowsize is still
   alive in winsize () and also added that it bails out after 20 loops of
   trying to get any values from xwinsize
 * Updated paralogger.spec file with the new dependencies
 * Made an icon for the rpm
 * Paralogger 0.89 released

 Tue Feb 29 16:38:42 CET 2000
 * Made pintchk () more robust

 Sat Mar  4 03:05:35 CET 2000
 * Moving redundant cruft from useterm () into runterm ()

 Fri Mar 17 01:29:03 CET 2000
 - Removing old chktail () to use colortail instead
 + Mapping different colortail configurations to different files with the
   file /etc/paralogger/colortail.map

 Fri Apr 14 01:52:41 CEST 2000
 * Made the rpm build again

 Apr 14 21:46:58 CEST 2000
 + Made two separate levels of verbose output
 * fixes to configure
 * updating --help output alot

 Mon May 22 01:11:18 CEST 2000
  Incorporated these changes by Tuomas Jormola <tj@sgic.fi>:
 * $H is $HEIGHT and $W is $WIDTH
 * Minor sysinfo() changes
 * Use sleep(1) instead of usleep
 * Automatically construct the $ETERM variable based on the Eterm version
   information. Users don't have to worry about it at all
 * Direct all possible error messages (stuff they print to STDERR) of
   the external programs, such as xdpyinfo(1), kill(1) and ps(1), to /dev/null
 * Hard-coded default options go to default() function
 + Did a Debian package, well two of them. Fonts come in a separate package
 + Use getopt(1) of util-linux to parse the command-line switches
 + Display help if no command-line switches weren't detected
 + Implemented support for separate config files, namely
   /etc/paralogger.conf and ~/.paraloggerrc. They both override the hard-coded
   options and ~/.paraloggerrc overrides /etc/paralogger.conf
 + Every configurable option can also be defined at the command-line
   Command-line options override all the other ways the define
   configuration options
 + Display help if no command-line switches weren't detected
 * New function fndfont(). It checks whether the specified $FONT is valid
   and if not, it sets it to "fixed", which should found on every X11 setup
   It uses xlsfonts(1) to determine the validness of $FONT

 Tue May 30 19:55:07 CEST 2000
 * Changed 0.8.0 in termver () to 0.8.*
 * Changed sanity () name to amisane ()

 Fri Jun  2 11:13:39 CEST 2000
 * Fixed mismatch in size of winsize of terminal because of missing $FONT
 * Fixed authenticate again by exporting $RTAIL and $PCONFPATH

 Sun Jun  4 12:35:56 CEST 2000
 * Regular tail is the default (was not in devel version) have to use --color
 * Use "uname -n" instead of "hostname -s"
 * Updated paralogger.spec

 Thu Jun  8 14:22:37 CEST 2000
 * Default install paralogger in /usr/bin instead of /usr/local/bin
 * Use usleep in paracal if found, exit if not called from paralogger
 * Use tac in reverse () if found

 Mon Jun 12 04:06:30 CEST 2000
 + Command line parser is finished (thanks to Tuomas Jormola <tg@sgic.fi>)
 * All Eterm related command line options are the same as the Eterm command
   line options. Would be foolish to use different options so people had to
   re-learn them if they already knew the command line options of Eterm :-)

 Tue Jun 20 03:41:48 CEST 2000
 * Fixed a bug that made paralogger only tail one log when single mode was
   specified, even if more than one log was specified at the command line

 Wed Jun 21 03:56:15 CEST 2000
 * Updated the debian package stuff

 Thu Jun 29 13:37:24 CEST 2000
 + Did a *HUGE* manpage update
 * Change pintchk () name to isitint () and make 0 a valid integer
 + Doing isitint on all arguments that expect an integer

 Mon Jul  3 06:10:47 CEST 2000
 + Add --font long option and update manpage
 * Change fndfont () name to fntxist ()
 * Updated the Debian package stuff with misc fixes and suggests colortail 
 + Added alternative winsize method 2 - people have reported that winsize ()
   fails on some systems
 + Checking if the values xwininfo returns are integers

 Wed Jul  5 05:10:44 CEST 2000
 + Does not run winsize () any more when loading a configuration file if the
   saved values for height, width or font are not overided from the commandline
   Uses the values for WINV and WINH from the configuration file
 + Discard files based on certain criterias (stoplist) in logxist ()
 + Add a third verbose level -vvv and add option to not use stoplist -nostop
 * Updated the html docs and added new screenshots and examples

 Tue Jul 11 20:24:29 CEST 2000
 * Do not sleep 1 in cascade () if run with -g

 Fri Jul 14 22:00:24 CEST 2000
 * Paralogger 0.90 released

 Wed Aug 30 18:06:27 CEST 2000
 + Use the -q option with cvs Eterm 0.9.1 so the Eterm(s) refuses to take
   keyboard input
 * Paralogger 0.91 released

 Fri Dec 22 02:35:18 CET 2000
 * Fixed a bug with the -single and -color option combined which would send the
   colortail configuration files in the wrong order to colortail.

   Fixed with a patch from Ren Gallati <draxinusom@gallati.net> which wrote: 
   "Messing a bit around in paralogger I quickly found that paralogger sends "-k 
   <colormap1> logfile1 -k <colormap2> logfile2" although colortail needs "-k
   <colormap1>,<colormap2>,<cmap3>... logfile1 logfile2 logfileN

 Wed Jan 31 02:02:40 CET 2001
 * Paralogger 1.0 released

 Sun Mar 17 17:12:59 CET 2002
 * Maybe this will fix the problems people report about being unable to 
   calibrate. Now, when calibrating and launching Eterm, paralogger forces 
   name AND title to "paracal-$$". Patch by LeAg <leag@altern.org>.
 + The Makefile now installs the docs
 * Cleanups to the .spec file
 * Paralogger 1.0.1 released

