Gopher LOG 3.3  7/20/93
   Andy Wick - awick@csugrad.cs.vt.edu
   bug fixes:  John-David Childs - jdc@selway.umt.edu

   * Fixed up main() routine so that: errors on the command line (or -h
   as the first parameter) cause glog to print the help information and
   abort.
   * Fixed PrintHelp() routine so help message is more understandable.
   * changed FILETYPE character from ' ' to 'I' (What about Image?)
   
============================================================

Announcing Gopher LOG 3.2 
   Andy Wick - awick@csugrad.cs.vt.edu

   A day before vacation and a bug in glog.  Oh Well.
   John-David Childs jdc@selway.umt.edu pointed out a slight bug
   in my ProcessLine routine dealing with search.  Should
   be fixed now.

=============================================================
   Since I am going on vacation for a while, I thought I would
release a minor update.  I have placed glog31.c (ANSI), 
glog31u.c (NON ANSI, Generated using unprotoize).  In boombox's
gopher incoming directory.  Or it is on the csugrad.cs.vt.edu gopher
server (port 70) in Software/Gopher Log Analyzer.

Original version by
   Chuck Shotton - cshotton@oac.hsc.uth.tmc.edu

3.1 adds some features of the old glog that got removed by mistake.
   * Supports all known log entries now. (I was missing search, binary,
      and someothers)
   * Can now have a tIme report
   * The report presentation is a little better
   * Can specify a starting and ending month for the whole command
   with -s <num> and -e <num>.  These are the only options that
   are used for the WHOLE command.  The rest are from left
   to right.

=============================================================
3.0 is almost a total rewrite of the original, with old ideas
from the original, and some new ones of my own.  I tried to make it easier
to get the type of reports that I wanted.  About the only thing
missing that 2.3 could do, is printing reports without the hits
field sorted.

I have tested 3.0 with purify and can report there are no memory leaks/bugs
what so ever (at least when I use it :).  The program is very memory hungry
though, you need about 1.5 times the size of the log file of memory.  This
is because the log file is completely parsed and stored in memory 
at startup. You can lower that amount if you compile without the DETAIL 
option, see "glog30.c".

How to use:
 glog30 -h will display the options :).  But as a short example.  

 How to create a [H]ost report, a [D]ata report with detail of [H]osts and
 width of 132, and a plot of [T]ypes with basename of "test".  The log file
 is called gopher.log.  Don't forget, reports always go to stdout.

 % glog30 -H -w 132 -DH -f test -pT < gopher.log | more

Tested on:
 gcc 2.3.3 was used on all platforms.
 Amiga DOS (main development), SunOS 4.1.3 (purify 2.0.3), Ultrix 4.2A

New Features:
 * The command line is parsed as the reports/plots are generated, so that
   it is possible to change options for each separately.  But options
   are NOT reset for each new one.

 * It now can produce reports on the [D]ata, [H]osts, day of the [W]eek,
   data [T]ypes, and [M]onth/date sorted (ie. sorted by date).

 * It can now produce detail reports on any thing you can produce a normal
   a report for.  Yes, you can even produce a host detail report of a 
   hosts report. (But it doesn't do you much good :)

 * Change the width of the reports.

 * You can specify the filename for plots, it defaults to gopher. (The 
   program appends a .dat, and and .run to the name)

 * You can request as many plots as you want. (But remember to change the
   filename or you will overwrite the last plot)
   
 * The program is now a lot easier to add new types of reports to, you
   only need to add 3 short routines, a variable, and edit 2 functions.
   (Well, maybe it is simple for me. :)

TODO:
 * Fix up plots some more: add histograms, and more options.

 * option for how detail works (which way should it work?)

 * Someone will probably want to have the hits field unsorted, and I will
   need to hack a way to do that.

 * Make it so reports can go to files.

 * Make some documentation

 * Improve error checking

If anyone has any ideas or suggestions please send them to me.