wmud/FAQ

WARNING: THIS FAQ IS OUT OF DATE
         Please reference the FAQ in the doc/ directory of the distribution
         instead.  It is significantly more current.


Frequently Asked Questions (FAQ) for Circle DikuMud with Answers
Alex Fletcher, furry@circlemud.org

This file is intended to cover common questions related to the CircleMUD
distributon source by Jeremy Elson (http://www.circlemud.org/~jelson/) and
not general DikuMUD questions.  Any contributions and corrections are more
than welcome.  It is currently maintained by Alex Fletcher (aka Furry)
<furry@circlemud.org>.  Please direct corrections to this address.  The
original author was Ryan Watkins (aka VampLestat). More information about
CircleMUD, including up-to-date versions of this documentation in ASCII and
Postscript, can be found at the CircleMUD Home Page
(http://www.circlemud.org/) or FTP site
(ftp://ftp.circlemud.org/pub/CircleMUD/).

______________________________________________________________________

Table of Contents

1. Introduction

   1.1 I've never played a MUD before.  What should I do?

   1.2 I'm new to C and/or coding. What do I do?

   1.3 I want to build my own MUD.  Where do I start?

   1.4 What is CircleMUD?

   1.5 What is the history of CircleMUD?

   1.6 Where is the original CircleMUD so I can check it out?

   1.7 Will the CircleMUD creators visit my mud?

   1.8 What is UNIX?


2. Resources

   2.1 Where do I find the source code for CircleMUD?

   2.2 Where do I find areas, etc. for CircleMUD?

   2.3 I have questions about CircleMUD.  Where should I go?

   2.4 How do I subscribe/unsubscribe to the Mailing List?

   2.5 To what platforms has CircleMUD been ported?

   2.6 How can I submit code or areas for use with CircleMUD?

   2.7 How do I use a patch file and how can I make one?


3. Compiling CircleMUD

   3.1 Why do I get many syntax errors with Sun's ``cc'' compiler?

   3.2 Why do I get all sorts of errors with ``crypt'' functions and header files?

   3.3 When I try to compile, why do I get a lot of undefined symbols
       referenced in comm.o for functions like socket, accept, and bind?

   3.4 Every time I try to compile Circle (or any other piece of software)
       under Linux, it gives me errors and says it cannot find include header
       files in the linux/  and asm/  directories.  What can I do?

   3.5 I'm getting compilation errors from a header file, and I didn't even
       change it?

   3.6 I'm trying to compile the MUD on Windows '95 and am having problems,
       what can I do?

   3.7 How can I do a ``grep'' on Windows 95?

   3.8 While compiling the MUD, why do I get errors like ``foo.c:1231:
       Undefined symbol `_whereamI' referenced from text segment''

   3.9 What is a parse error and how do I fix it?

   3.10 I have this piece of code that calls bcopy(), bzero(), and bcmp()
        and it won't compile, so what can I do?

   3.11 My compiler doesn't have ``strdup()'', what can I do?

   3.12 I am having trouble with my ``makefile'', what could be the problem?

   3.13 How can I handle directories in C?


4. Running CircleMUD

   4.1 I typed ``autorun'' but then my terminal just froze.

   4.2 I typed ``bin/circle'' and got lots of boot messages, but then it
       said ``Entering game loop'' and froze.

   4.3 Okay, I think the MUD is running but why don't I get a login prompt?

   4.4 How come I get this error when running my MUD: ``Error reading board:
       No such file or directory''

   4.5 I just got this SIGPIPE, what is it and what Can I Do About It?

   4.6 When I run Circle under Linux, it tells me ``gethostbyaddr:
       connection refused'' when the MUD boots, and then dies.  Why?

   4.7 When I run Circle under Windows 95, it tells me ``Winsock error
       #10047'' when the MUD boots, and then dies.  Why?

   4.8 When I run Circle under Windows 95, players can't rent---their
       equipment is just dropped on the ground, syslogs don't work, so what
       is the problem?

   4.9 When someone logs on to my Windows 95 MUD, the console screen gives:
       ``gethostbyaddr: No such file or directory''

   4.10 My MUD crashed and my connection got closed.  What can I do?

   4.11 Ok, so how do I use ``gdb''?

   4.12 How can I hunt bugs more effectively?

   4.13 I just added n levels to my MUD (from the stock 34).  How do I set
        my imps up to level n without a pfile wipe?

   4.14 I decided to wipe my pfile away anyway.  What steps should I take to
        do this?

   4.15 I want to expand the ability to pk in my MUD, allowing ASSASSINS
        that'll be able to PK without getting flagged.  How can I do this?

   4.16 Why does it say ``Connection closed by foreign host.'' and not
        display the ``Byebye!'' message I'm trying to send before cutting
        someone off?

   4.17 I run my MUD on a unix system and the pfile/rent file works great,
        but on my home system it's all screwed up.  What gives?

   4.18 How do I get the CircleMUD to autoload when the Linux server is
        restarted?

   4.19 My server shuts down my MUD everytime I logoff. How do I keep the
        MUD running when I logoff?


5. Code Changes for CircleMUD 2.20

   5.1 How do I fix the bug where people can junk more coins than available?

   5.2 How do I fix the ``vstat'' bug that crashes the MUD?

   5.3 How do I fix the ``wizlock'' bug that lets lower immortals lock out
       higher immortals?

   5.4 How do I fix the ``mudlog'' bug that lets people see me log in, even
       if I'm wizinvis?


6. CircleMUD 3.0 Questions

   6.1 Are there any bugs in patch level 19?

   6.2 How do I access Online Creation?

   6.3 How does the new bitvector system work?

   6.4 When will the production release of Circle 3.0 be?

   6.5 If someone logs in and just sits at the password prompt, the MUD
       hangs (i.e., no one else can connect or do anything) until the person
       enters their password.

   6.6 Why does CircleMUD use BUF switches all through the code, what's
       happening here?

   6.7 How do I add a new class?  How do I add more levels?

   6.8 Is there a complete ``coding.doc''?

______________________________________________________________________

1.  Introduction


1.1.  I've never played a MUD before.  What should I do?

Don't try to use your own copy of CircleMUD!  There are two levels of MUD
users: players and administrators.  Administrators do what you're trying to
do now -- get a copy of a MUD's source code, compile it, and run it. 
Players use MUDs that are being administered by someone else. If you try to
actually run a MUD before you've ever played one, you'll get very confused
indeed! Your best bet for now is to play someone else's MUD first.  There
are a large number of excellent MUDs out there already, some of which are
based on CircleMUD code.  A good place to start looking is the MUD Connector
at:

http://www.mudconnect.com/

The CircleMUD web server also has a smaller CircleMUD Site List at:

http://www.circlemud.org/sites.html


1.2.  I'm new to C and/or coding.  What do I do?

First, a MUD is not a learning project. It has thousands of lines to it,
many of which are obscure and unclear to even moderately skilled
programmers.  Those little, ``Hello, world,'' programs are for learning,
maybe little math tests, etc.  A MUD is a pretty ambitous project to start
with.  That's like trying to run before you can walk, and while there's more
difficult things than a MUD to start with, there's a ton of easier things
you should start with.  Second, if you are persistent, get a good C
reference book and read the code, try to comprehend what everything is doing
(to some small extent).  You should probably avoid comm.c as it is home to
the socket functions which general C books don't cover and are usually
explained in other books on network programming.  Then try small projects,
something similar to what already exists.  This way, you can get away with a
cut-and-paste job and changing some of the code.  Adding a simple version of
races isn't all that difficult, just examine the class code in class.c and
the CON_QCLASS block in interpreter.c, cut, paste, and modify.  Also look at
structs.h, for "#define CLASS_" You'll begin understanding more and more of
the code as you copy and change it, eventually you'll be able to write a
whole function by yourself.  Spend time learning, going with haste will hurt
you more than it will help you.


1.3.  I want to build my own MUD.  Where do I start?

Many common questions arise from new MUD admins.  It is a good idea to pay
attention to them and take their answers to heart.  These include things
like the following:

I don't have any coding experience with MUDs, but do have a lot of ideas for
my own.  I have played many MUDs for a LONG time, though.

Read the FAQ.  MUD Experience doesn't help a huge amount.  Code experience
does.

I am interested in having a level system of 1-50 mortal and 51-60 imms.  I
am also interested in adding races and classes. How can I accomplish these
things?

By checking the FTP Site under contrib.
(ftp://ftp.circlemud.org/pub/CircleMUD/contrib).  Learn a lot from there.
Learn from the Snippets page (http://developer.circlemud.org/snippets/), and
*know* the FAQ.

Also, is there anything that I should know about CircleMUD being a "newbie"
to it?

See the above comment.


1.4.  What is CircleMUD?

CircleMUD is a DikuMUD derivitave, developed by Jeremy Elson
<jelson@circlemud.org> and is from the Gamma v0.0 of DikuMUD created by Hans
Henrik Staerfeldt, Katja Nyboe, Tom Madsen, Michael Seifert and Sebastian
Hammer at DIKU (Computer Science Instutute at Copenhagen University).  Note
that CircleMUD is a Diku derivative, so its users must follow the DIKU
license agreement---most notably it cannot be used to make money in ANY way,
the original developers' names must be in the login screen and that the
credits command always presents the same information, etc.

Quoting from CircleMUD's release.doc:

     "CircleMUD is highly developed from the programming side, but highly
     UNdeveloped on the game-playing side.  So, if you're looking for a huge
     MUD with billions of spells, skills, classes, races, and areas, Circle
     will probably disappoint you severely.  Circle still has only the 4
     original Diku classes, the original spells, the original skills, and
     about a couple dozen areas. On the other hand, if you're looking for a
     highly stable, well-developed, well-organized "blank slate" MUD on
     which you can put your OWN ideas for spells, skills, classes, and
     areas, then Circle might be just what you're looking for."


The latest full release of Circle is 2.20, released on November 17, 1993. 
Currently 3.0 is in beta, at patch level 19, released on August 14, 2001.


1.5.  What is the history of CircleMUD?

o  Version 2.00: July 16, 1993
o  Version 2.01: July 20, 1993
o  Version 2.02: Early August
o  Version 2.10: September 1, 1993
o  Version 2.11: September 19, 1993
o  Version 2.20: November 17, 1993

Version 3.00 is currently in beta right now and is up to patchlevel 19.  The
final release is due out Real Soon Now(tm).  Don't bother posting requests
for it.  It will be out when Jeremy gets it done, and it will be announced
on the mailing list and the newsgroup rec.games.mud.diku.


1.6.  Where is the original CircleMUD so I can check it out?

CircleMUD is a public code base, freely distributable, but the author of
Circle doesn't actually run one personally.  There used to be CircleMUD, and
while Jeremy Elson continues to develop it, there is no original CircleMUD
any more.  To see other MUDs that are using the CircleMUD code base, check
out this Site List: http://www.circlemud.org/sites.html


1.7   Will the CircleMUD creators visit my mud?

While there is a possibility that one (or more) of the CircleMUD creators
will drop by your MUD for a visit, to play, or to simply look around, there
is a slim chance that they will even identify themselves.

Would you like to see us on your MUD?  You won't.  We don't want free wizzes
or favours for our work here.  In fact, we will state categorically that if
anyone comes on your MUD claiming to be associated with us, they aren't.

If you want to check whether or not someone on your MUD really is one of us,
we can all be contacted by email for verification:
  Jeremy Elson  <jelson@circlemud.org>
  Chris Epler   <cepler@circlemud.org>
  Alex Fletcher <furry@circlemud.org>
  George Greer  <greerga@circlemud.org>
  Daniel Koepke <dkoepke@circlemud.org>
  Tony Robbins  <tonyr@circlemud.org>

As an interesting side note, there are two interesting things about the
order of the addresses mentioned above:
* They are in alphabetical order by last name.
* They are in chronological order of when each person joined the
  CircleMUD team.


1.8.  What is UNIX?

UNIX is not an operating system of itself, it's a type (flavour, if you
will) of operating systems.  Many different kinds of UNIXes exist.  Some of
them are free, some of them are not.  How to tell if you have a UNIX
operating system? Well, UNIXes have the `ps' command, tend to have a `%' or
`#' prompt, give you a home directory, `who' will show who else is on the
system, etc.  Many UNIX systems (such as Linux) strive to be POSIX
compatible, so you'll probably see POSIX mentioned, too.  POSIX is, roughly,
the standards which UNIX operating systems go by.  It says what makes an
operating system part of the UNIX family and so forth.  Some UNIX operating
systems are not 100% POSIX compatible, actually, most aren't.  The following
are types of UNIX (but not all the existing flavours): Linux, FreeBSD, BSD,
BSDi, Solaris.  There are others.  UNIX operating systems are command-based
and Microsoft does not make a variant.



2.  Resources

2.1.  Where do I find the source code for CircleMUD?

Circle's complete source code and areas are available for anonymous FTP at
the CircleMUD FTP site:

ftp://ftp.circlemud.org/pub/CircleMUD/

There is also a CircleMUD Home Page:

http://www.circlemud.org/

for additional CircleMUD links and information.


2.2.  Where do I find areas, etc. for CircleMUD?

A number of CircleMUD based Implementors have submitted areas to the public
and they are archived at:

ftp://ftp.circlemud.org/pub/CircleMUD/contrib/areas/  (California, USA)

Or you can use one of the CircleMUD mirrors:

ftp://ftp2.circlemud.org/pub/CircleMUD/  (Ohio, USA)
ftp://ftp.stormhaven.org/pub/CircleMUD/  (Virginia, USA)
ftp://ftp.mono.org/pub/mud/CircleMUD/  (London, UK)

If you cannot use FTP, you can contact one of the site maintainers and
request a file.  The maintainers are:

Alex Fletcher <furry@circlemud.org>
Chris Epler <cepler@circlemud.org>
George Greer <greerga@circlemud.org>

If the FTP site is down, please try one of the mirrors and then contact Alex
at his alternate address <fletchra@qsilver.queensu.ca> since he will be
able to find out the situation.  There is also a code snippets site at:

http://developer.circlemud.org/snippets/

Where a number of code suggestions and hints are located.


2.3.  I have questions about CircleMUD.  Where should I go?

If you have general questions about the MUD such as how to get it running,
how to add new spells, how to add new skills, etc., the first place you
should look is the documentation. `coding.doc' will have information about
how to add new spells, skills, commands, etc. `building.doc' has information
about how to create new worlds, how to read the database files, etc. There
are many other documents in the doc directory with useful information.

There is also a new project, started in June of 1996, called the CircleMUD
Documentation Project (http://www.circlemud.org/~jelson/circle/cdp/) which
will eventually be a repository for all Circle-related information.  It's
still being built as of this writing, but hopefully will become a valuable
resource as more documentation is added.

If you still have questions after reading the doucmentation, you can try
asking on the CircleMUD mailing list (see next section).

If you have a question you think is too "newbie" for the mailing list, try
the help database <help@circlemud.org>.  Previous questions are accessible
though the bugs database (http://bugs.circlemud.org/). That is also,
incidently, where you should go to report bugs.

You can also contact the author, Jeremy Elson <jelson@circlemud.org>. 
George Greer <greerga@circlemud.org> can also be contacted for most
problems.


2.4.  How do I subscribe/unsubscribe to the Mailing List?

There is a CircleMUD mailing list for coders, builders, and administrators.
To subscribe, send a message to the list server <listserv@post.queensu.ca>
with a message body of subscribe circle <first name> <last name>.  To
unsubscribe from the list send a message to <listserv@post.queensu.ca> with
the words unsubscribe circle as the message body.  DO NOT send subscription
or unsubscription requests to the list in general.  There are hundreds of
people on the list, and it will only irritate a ton of people who have no
power to remove you from the list.  Read the Mailing List FAQ
(http://qsilver.queensu.ca/~fletchra/Circle/list_faq.html) for more
information.


2.5.  To what platforms has CircleMUD been ported?

Version 3.0, although still officially in beta-testing, is very portable
because it uses the GNU autoconf system, meaning you only need to type
``configure'' to have it automatically determine various features of your
system and configure the code accordingly.  3.0 compiles without changes
under most BSD and SVR4 systems, including SunOS, Solaris, Ultrix, IRIX,
AIX, Linux, BSD/OS, HP/UX, and others.

Version 3.0 is also being ported to various non-UNIX platforms.  As of
patchlevel 14, you can compile Circle under OS/2 2.x and 3.x with the OS/2
port of gcc, Windows 95/NT using Microsoft Visual C++ version 4.0 or 5.0,
Borland (now Inprise) C++ 4.5, Watcom v.11, Cygnus GNU-WIN32, LCC, Macintosh
with CodeWarrior, Amiga, and Acorn RiscOS.

The older version of the code, Version 2.20, compiles mainly on BSD UNIX
systems but has trouble under SVR4-based systems such as Solaris. The author
has personally compiled and tested v2.20 under Ultrix 4.0, IRIX 4.0.1, 4.0.4
and 4.0.5, SunOS 4.1.1 and 4.1.3, AIX 3.2, Linux 0.99.x and 1.0.x, and
ConvexOS V10.2.  Users have reported that v2.20 compiles with relatively
minor changes under NeXTStep 2.1 and 3.0, and HP/UX 9.0.

Jean-Jack Riethoven <J.J.M.Riethoven@ab.agro.nl> ported CircleMUD version
2.20 to the Amiga and has contributed his code for version 3.0 of CircleMUD. 
Questions about the Amiga source should be directed to Jean-Jack Riethoven,
not Jeremy Elson or George Greer.


2.6.  How can I submit code or areas for use with CircleMUD?

There is a special uploads area:

ftp://upload.circlemud.org/pub/CircleMUD/incoming/

..in the CircleMUD domain for submissions of code, areas, utilities,
scripts, and anything else that might be of use to Circle users.  Please
make sure only to upload to upload.circlemud.org and not any of the
CircleMUD mirrors. These portions of code or areas will probably not be
added to the full release of CircleMUD unless you make specific arrangements
with Jeremy for code items, or with Alex for area/lib files.


2.7.  How do I use a patch file and how can I make one?

Patch files are created and used using the ``diff'' and ``patch'' utilities,
respectively.  They can both be downloaded from the GNU FTP Site
(ftp://ftp.gnu.org/pub/gnu/) under the name ``diffutils- xxx.tar.gz''.  If
you happen to be using a DOS or Windows machine, you should get patch for
DOS (ftp://204.119.24.14/pub/patch/patch12.zip). This does not support long
filenames (but can be used with short filenames).  There is also a set of
Windows 95 Utilities now available from the Cygnus Gnu-Win32 Project
(http://www.cygnus.com/misc/gnu-win32/).

These are the various parameters to use with diff (all work in general on
unix based systems, but check out the help entries to be certain.

diff -uN [original_src_directory] [altered_src_directory] > Patch

-u is the unified output. ie. it tells diff to output the text what is
called ``patch'' style.  On some systems, you will have to use -c but it
generates much larger and harder to follow patches.

-N  Tells diff to treat files that are in one directory and not there in the
other as being empty in the one they are not there.  It allows entire files
to be included into the patch.

-r  recursive, add r to the uN above if you want it to recursively add in
any subdirectories.  (be careful with this one)

-p  Tells diff to indicate what function is being ``patched'' in each
section. This may not be supported by all versions of ``diff.''

If you download a patch file and would like to add it to your code, first
make sure to read any instructions that the patch author might have written. 
The command used to add the patch may vary depending on how the patch was
created. This should given in the first line of the patch or in the
instructions.  Normally, if using GNU patch with a unified diff, the command
should be:

        patch -u < [patchfile]

If the patch was created with a SYSV patcher (i.e. not a unified diff), the
patch should be added with:

        patch -c < [patchfile]

Of course, if the instructions state otherwise, ignore any instruc- tions
given here and follow the instructions given with the patchfile instead.

Finally, in modern patches, there are three characters of interest to note:

o  ! :: The line changes between new and old.
o  + :: This line is added to the old to make the new.
o  - :: This line is removed from the old to make the new.
o  The rest of the lines are just there to give you an idea of where
   to change.


3.  Compiling CircleMUD

3.1.  Why do I get many syntax errors with Sun's ``cc'' compiler?

Because Circle is written in ANSI C, and Sun's standard cc compiler isn't
capable of compiling ANSI C code.  You can try acc, Sun's ANSI C compiler,
but it costs extra money to get it from Sun so your sysadmin may not have
installed it.  Most don't.  The best solution is to get the GCC compiler
from the GNU FTP site (ftp://ftp.gnu.org/pub/gnu) and install it, if you
have enough time and space.


3.2.  Why do I get all sorts of errors with ``crypt'' functions and
      header files?

(This information applies ONLY to Version 3.0 of the code.) CircleMUD
normally uses the UNIX crypt() function to enrypt players' passwords.
Because of export restrictions imposed by the U.S., some systems do not have
the crypt() function. ``configure'' will usually be able to figure out
whether or not your system has crypt(), but if it guesses incorrectly and
you see problems with the crypt() function or headers, you can manually
disable password encryption by going into the sysdep.h source file and
uncommenting the line that reads:

        #define NOCRYPT

Be warned, however, that doing this causes the MUD to store players'
passwords in plaintext rather than as encrypted strings.  Also, if you move
from a system which has crypt to one that doesn't, players won't be able to
log in with their old passwords!


3.3.  When I try to compile, why do I get a lot of undefined symbols
      referenced in comm.o for functions like socket, accept, and bind?

SVR4 systems require the socket and nsl libraries for network programs.  You
shouldn't see this error any more with version 3.0 because ``configure''
should automatically use those libraries for you; however, if you still have
problems, try adding ``-lsocket -lnsl'' to the line in the Makefile that
links all the object files together into the 'circle' binary.

If you're using V2.20 and you have this error, the best thing to do is
simply to use V3.0 instead.  If you insist on using 2.20, go into the
Makefile and search for the comment next to ``SVR4''.

3.4.  Every time I try to compile Circle (or any other piece of soft- ware)
      under Linux, it gives me errors and says it cannot find include header
      files in the linux/ and asm/ directories.  What can I do?

Under Linux, you cannot compile any program unless you install the kernel
source code because the kernel source includes the ANSI C header files.  You
need the files in /usr/include/linux, which are distributed separately from
the rest of /usr/include.

The easiest way to do this if you're using the Slackware Linux distribution
is simply to install the 'K' series of disks which is the kernel source.

Otherwise, you'll have to set up your include files manually. The easiest
way to get these is to download kernel source from:

http://www.kernel.org/mirrors/

Get the kernel source that matches the kernel you're running (type `uname
-a' to find your kernel version).  Then unpack the kernel into the usr/src
/directory.  It's about 13 megabytes compressed and 54 megabytes
uncompressed (For the 2.2.9 kernel).

Read the README file that comes with the kernel, and make the symbolic links
you need for /usr/include/asm and /usr/include/linux.

Now compile the MUD.  This will take care of most of the errors.  You may
have to do `make config' and `make dep' in /usr/src/linux as well, in order
to make linux/config.h and other files that get generated by these steps.

You can remove the whole kernel source tree except for include/ at
this point and get most of your 48.5 megs back.

(Thanks to Michael Chastain for providing this answer.)


3.5.  I'm getting compilation errors from a header file, and I didn't even
      change it?

Okay, if you really didn't change ``structs.h'' then the error isn't in
``structs.h''.  I think I've seen 2 cases where this has happened, the
first, is that the header file you included right before the header file
messing has an error in it.  I can't really say much beyond that, but look
for a missing semicolon, are any other errors you can find.

If you include files out of order, it can mess things up.  For example, B.h
has stuff in it that is defined in A.h, and if you include B.h before A.h,
you can get errors, your best bet here is to mess with the order of the
headers, making sure you put ``conf.h'' and ``sysdep.h'' at the top,
followed by ``structs.h'', ``utils.h'', etc. Any file specific headers
should be the last one included just for coding style.


3.6.  I'm trying to compile the mud on Windows '95 and am having prob- lems,
      what can I do?

The first thing to do is to make sure you are compiling a recent version of
the source code.  Patch Level 11 and onwards all support Windows '95 winsock
sockets now.  Second, you should ensure that you have carefully read the
README.WIN file for instructions on what to include.  Next, ensure that you
are using a C compiler that supports long filenames (for example, MSVC 4.0
does, MSVC 1.0 does not).  If you happen to be trying to patch something
into your code, you should use patch for DOS
(ftp://204.119.24.14/pub/patch/patch12.zip).  This does not support long
filenames (but can be used with short filenames).


3.7.  How can I do a ``grep'' on Windows 95?

1. Select ``start menu''->``find''->``files or folders''
2. Enter the files/dirs to search in.
3. Select ``Advanced''
4. In the ``Containing Text'' input box, type in the text you want.
5. Double click on a match to bring up the file that matched.
   Even better is to use MSVC's find command (if you have it).


3.8.  While compiling the mud, why do I get errors like ``foo.c:1231:
      Undefined symbol `_whereamI' referenced from text segment''

You forgot to include a source file into the make.  Go edit your Makefile
and make sure all the necessary *.c files are in there, in particular,
whichever C file defines the function that the compiler is complaining is
undefined.  If all else fails, try deleting all the *.o files and
recompiling from scratch.


3.9.  What is a parse error and how do I fix it?

A parsing error is often a missing or extra semicolon, parenthesis, or
bracket ({).

If the parse error is before a semicolon at the end of a line of code, it is
something on that line.

If it is at the beginning of a line within a function, it is usually a
missing semicolon on the previous line.

If it is at the beginning of a function, count your brackets (especially the
{} ones) in the previous function.

I can't think of any other parse errors.  These are the ones I commonly see. 
With a bit of practice, they are very easy to locate and fix.  For a more
detailed explanation, check out the C Language FAQ.


3.10.  I have this piece of code that calls bcopy(), bzero(), and bcmp() and
       it won't compile, so what can I do?

All three of these functions are fairly standard on BSD systems. However,
they are not considered to be very portable, and thus should be redefined. 
For example, the equivalents for SYSV are:


#define bcopy(from,to,len)      memmove(to,from,len)
#define bzero(mem,len)          memset(mem,0,len)
#define bcmp(a,b,len)           memcmp(a,b,len)


3.11.  My compiler doesn't have ``strdup()'', what can I do?

Use Circle's built-in str_dup() function instead.


3.12.  I am having trouble with my ``makefile'', what could be the problem?

If you used cut and paste to insert items into your makefile, it is likely
that you accidentally put spaces at the beginning of lines where tabs are
needed.  This is how the makefile must be constructed:

foo.o: foo.c conf.h sysdep.h structs.h utils.h interpreter.h \
  handler.h db.h
{TAB}$(CC) -c $(CFLAGS)

To add these lines properly, you can use gcc to assist you with the
following shell script (from Daniel Koepke):

#!/bin/sh
gcc -MM $1 >> Makefile
echo "{TAB}\$(CC) -c \$(CFLAGS) $1" >> Makefile

To use this script, replace {TAB} with a tab, and then run the script
like: add_file foo.c


3.13.  How can I handle directories in C?

Note that this seems only to be valid for UNIX OSes.  Handling of
directories is accomplished through the dirent.h and sys/types.h files.  The
function opendir() returns a ``DIR*'' pointer (it's like but not the same as
the ``FILE *'' pointer) when you pass it the name of a directory to open or
NULL if it can't open the dir.  After the directory has been opened, you can
step through the files or search for particular files, etc. using readdir(),
seekdir(), and scandir(). When you reach the end of the directory list, you
can either go back to the start with rewinddir() or close the directory with
closedir(). The following code (which has not been tested) should open a
directory and go through it one by one and prints the filenames:

  struct dirent * ffile;
  DIR * my_dir;

  if (!(my_dir = opendir("foo")))
    return;

  while (1) {
    if (!(dirent = readdir(my_dir)))
      break;
    printf("%s\n", dirent->d_name);
  }

  closedir(my_dir);

The dirent structure contains only two useful elements, the file's name
(d_name) and the files length (d_reclen).

Thanks to Daniel Koepke for the above.


4.  Running CircleMUD

4.1.  I typed ``autorun'' but then my terminal just froze.

autorun is a script which automatically runs, logs, and reboots the game for
long-term runs.  You should run autorun in the background by typing
``./autorun &'' -- the MUD will start running in the background and you'll
get the normal UNIX prompt back immediately (see section 4.3). The game will
then run unattended until you explicitly shut it down.

On some systems, you may need to prepend ``nohup'' to the autorun command
since some systems will kill off any processes left running when you leave
the shell.


4.2.  I typed ``bin/circle'' and got lots of boot messages, but then it said
      ``Entering game loop'' and froze.

It is not frozen, it is just waiting for people to connect.  You have to run
the MUD in the background by typing ``bin/circle &'' and then use telnet to
connect to the game (see next section).


4.3.  Okay, I think the MUD is running but why don't I get a login prompt?

In order to play the MUD, you must connect to it using the telnet command,
i.e. ``telnet localhost 4000''.


4.4.  How come I get this error when running my mud: ``Error reading board:
      No such file or directory''

This is not a bad thing, all it means is that you have some boards on the
mud and that it can't find the file for them.  Since it can't find the file,
the mud will just create the file on the fly and use that, so the next time
something is posted to the board, the files will exist.  However, if you did
have files for the boards and you are suddenly getting this error, it means
that the board files have been deleted or something similar.


4.5.  I just got this SIGPIPE, what is it and what Can I Do About It?

Often it appears that other people send your system SIGPIPEs when their
connection is closed, in fact, it is not the person sending the SIGPIPE, it
is your system.  The SIGPIPE is generated when your program attempts to
write to descriptor which has no one listening to it.  This occurs if the
character is sent a message by the mud after connecting, but before the
socket is flagged with an exception or reads 0 bytes.  By default, CircleMUD
ignores these SIGPIPEs, with the line my_signal(SIGPIPE, SIG_IGN) in
signal_setup().  Where most people see the problems with SIGPIPE is while
debugging with GDB.  By default, GDB responds to a SIGPIPE by stoping the
program, printing that a SIGPIPE was received, and passing it to the
program.  You can change the action taken by GDB by using the `handle'
command. To stop the program from stoping at SIGPIPE, you would give GDB the
command `handle SIGPIPE nostop'


4.6.  When I run Circle under Linux, it tells me ``gethostbyaddr: con-
      nection refused'' when the MUD boots, and then dies.  Why?

You need to make sure you have Networking and TCP/IP support compiled into
your Linux kernel, even if you aren't actually connected to the Internet. 
The easiest way to do this if you're using Slackware is to install one of
Slackware's precompiled networking kernels.  Also, make sure to install
Slackware's `N' series of disks which contains other networking support
files.

If Slackware's precompiled kernel isn't available you'll have to compile the
kernel yourself.  First make sure the kernel source is installed in
/usr/src/linux (see section 3.1.4) and follow the instructions in
/usr/src/linux/README. When you do `make config' it will ask you a series of
questions about which kernel features you want; make sure to answer ``Y'' to
``Networking support'' and ``TCP/IP support''.


4.7.  When I run Circle under Windows 95, it tells me ``Winsock error
      #10047'' when the MUD boots, and then dies.  Why?

You need to configure TCP/IP networking from the Network Control Panel, even
if you are not connected to the Internet.  From the Network Control Panel,
select ``Add Protocol'', and under the vendor ``Microsoft'', choose
``TCP/IP''.  It may ask you to insert the Windows 95 CDROM in order to copy
the drivers onto your hard drive.


4.8.  When I run Circle under Windows 95, players can't rent---their
      equipment is just dropped on the ground, syslogs don't work, so what
      is the problem?

The reason that objects aren't saved when your players quit is that certain
unzip programs are buggy and don't completely recreate the MUD's directory
structure (in particular, it doesn't create directories which have no files
in them.) This is fixed in Circle 3.0 patchlevel 12 and above.  Before
patchlevel 12, you can fix it simply by manually creating the needed
directories:

        CD \Circle30bpl11
        cd lib\plrobjs
        mkdir A-E
        mkdir F-J
        mkdir K-O
        mkdir P-T
        mkdir U-Z
        mkdir ZZZ

Object saving should then work.  The syslogs are a different story; no data
is written to the system logs because the code currently is configured
simply to write all errors to the standard error file descriptor (stderr),
and Windows doesn't seem to let you redirect stderr to a file the same way
UNIX does.  pl12 allows you to direct logs to a specific file instead.


4.9.  When someone logs on to my Windows 95 MUD, the console screen
gives: ``gethostbyaddr: No such file or directory''

This means the MUD can't resolve the IP address of the connecting player's
source site innto a hostname.  You probably don't have DNS correctly
configured in the Windows Network Control Panel menu (under configuration of
the TCP protocol).  Make sure you have the IP address of your ISP's DNS
server listed.


4.10.  My MUD crashed and my connection got closed.  What can I do?

Just because your connection got closed from the MUD (for example, if you
get too much information sent to you and the telnet session gets closed),
this doesn't always mean that the game itself crashed. Before reporting
something as a crash bug, make sure that the game itself crashed, and above
all, try to duplicate the circumstances before reporting it as a crash bug. 
You can also try using gdb to find out why the MUD is crashing if it gives
you a core dump.


4.11.  Ok, so how do I use ``gdb''?

GDB has some online help, though it is not the best.  It does at least give
a summary of commands and what they're supposed to do.  What follows is
Sammy's short intro to gdb with some bughunting notes following it:

If you've got a core file, go to your top circle directory and type:

> gdb bin/circle lib/core

If you want to hunt bugs in real time (causing bugs to find the cause as
opposed to checking a core to see why the MUD crashed earlier) use:

> gdb bin/circle

If you're working with a core, gdb should show you where the crash occurred. 
If you get an actual line that failed, you've got it made. If not, the
included message should help.  If you're working in real time, now's the
time to crash the MUD so you can see what gdb catches.

When you've got the crash info, you can type ``where'' to see which function
called the crash function, which function called that one, and so on all the
way up to ``main()''.

I should explain about ``context''  You may type ``print ch'' which you
would expect to show you the ch variable, but if you're in a function that
doesn't get a ch passed to it (real_mobile, etc), you can't see ch because
it's not in that context.  To change contexts (the function levels you saw
with where) type ``up'' to go up.  You start at the bottom, but once you go
up, and up, and up, you can always go back ``down''.  You may be able to go
up a couple functions to see a function with ch in it, if finding out who
caused the crash is useful (it normally isn't).

The ``print'' command is probably the single most useful command, and lets
you print any variable, and arithmetic expressions (makes a nice calculator
if you know C math).  Any of the following are valid and sometimes useful:

print ch (fast way to see if ch is a valid pointer, 0 if it's not)
print *ch (prints the contents of ch, rather than the pointer address)
print ch->player.name (same as GET_NAME(ch))
print world[ch->in_room].number (vnum of the room the char is in)

etc..

Note that you can't use macros (all those handy psuedo functions like
GET_NAME and GET_MAX_HIT), so you'll have to look up the full structure path
of variables you need.

Type ``list'' to see the source before and after the line you're currently
looking at.  There are other list options but I'm unfamiliar with them.

(From Sammy <Samedi@cris.com>)

For more information, you can try checking out the GDB Debugger
(http://www.dgii.com/people/robertl/skunk/gdb/gdb_toc.html)


4.12.  How can I hunt bugs more effectively?

There are only a couple of commands to use in gdb, though with some patience
they can be very powerful.  The only commands I've ever used are:

run                     well, duh.
print [variable]        also duh, though it does more than you might think
list                    shows you the source code in context
break [function]        set a breakpoint at a function
clear [function]        remove a breakpoint
step                    execute one line of code
cont                    continue running after a break or ctrl-c

I've run into nasty problems quite a few times.  The cause is often a memory
problem, usually with pointers, pointers to nonexistent memory. If you free
a structure, or a string or something, the pointer isn't always set to NULL,
so you may have code that checks for a NULL pointer that thinks the pointer
is ok since it's not NULL.  You should make sure you always set pointers to
NULL after freeing them.

Ok, now for the hard part.  If you know where the problem is, you should be
able to duplicate it with a specific sequence of actions. That makes things
much easier.  What you'll have to do is pick a function to ``break'' at. 
The ideal place to break is immediately before the crash.  For example, if
the crash occurred when you tried to save a mob with medit, you might be
able to ``break mobs_to_file''.  Try that one first.

When you `medit save', the MUD will hang.  GDB will either give you segfault
info, or it will be stopped at the beginning of mobs_to_file.  If it
segfaulted, pick an earlier function, like copy_mobile, or even do_medit.

When you hit a breakpoint, print the variables that are passed to the
function to make sure they look ok.  Note that printing the contents of
pointers is possible with a little playing around.  For example, if you
print ch, you get a hex number that shows you the memory location where ch
is at.  It's a little helpful, but try print *ch and you'll notice that it
prints the contents of the ch structure, which is usually more useful. 
print ch->player will give you the name of the person who entered the
command you're looking at, and some other info. If you get a no ch in this
context it is because the ch variable wasn't passed to the function you're
currently looking at.

Ok, so now you're ready to start stepping.  When GDB hit your breakpoint, it
showed you the first line of executable code in your function, which will
sometimes be in your variable declarations if you initialized any variables
(ex: int i = 0).  As you're stepping through lines of code, you'll see one
line at a time.  Note that the line you see hasn't been run yet.  It's
actually the _next_ line to be executed.  So if the line is a = b + c;,
printing a will show you what a was before this line, not the sum of b and
c. If you have an idea of where the crash is occurring, you can keep
stepping till you get to that part of the code (tip: pressing return will
repeat the last GDB command, so you can type step once, then keep pressing
return to step quickly).  If you have no idea where the problem is, the
quick and dirty way to find your crash is to keep pressing return rapidly
(don't hold the eturn key or you'll probably miss it).  When you get the seg
fault, you can't step any more, so it should be obvious when that happens.

Now that you've found the exact line where you get the crash, you should
start the MUD over and step more slowly this time. What I've found that
works really well to save time is to create a dummy function.  This one will
work just fine:

void dummy(void){}

Put that somewhere in the file you're working on.  Then, right before the
crash, put a call to dummy in the code (ex: dummy();).  Then set your
breakpoint at dummy, andwhen you hit the breakpoint, step once to get back
to the crashing code.

Now you're in total control.  You should be looking at the exact line that
gave you the crash last time.  Print *every* variable on this line.  Chances
are one of them will be a pointer to an unaccessable memory location.  For
example, printing ch->player.name may give you an error.  If it does, work
your way back and print ch->player to make sure that one's valid, and if it
isn't, try printing ch.

Somewhere in there you're going to have an invalid pointer.  Once you know
which one it is, it's up to you to figure out why it's invalid. You may have
to move dummy() up higher in the code and step slowly, checking your pointer
all the way to see where it changes from valid to invalid.  You may just
need to NULL a free'd pointer, or you may have to add a check for a NULL
pointer, or you may have screwed up a loop.  I've done all that and more :)

Well, that's it in a nutshell.  There's a lot more to GDB that I haven't
even begun to learn, but if you get comfortable with print and stepping you
can fix just about any bug.  I spent hours on the above procedure trying to
get my ascii object and mail saving working right, but it could have taken
weeks without gdb.  The only other suggestion I have is to check out the
online gdb help.  It's not very helpful for learning, but you can see what
commands are available and play around with them to see if you can find any
new tools.

(From Sammy <samedi@cris.com>)


4.13.  I just added n levels to my MUD (from the stock 34).  How do I set my
       imps up to level n without a pfile wipe?

You can write a quick and nasty function that will advance your imp (and imp
only) to the max level (LVL_IMPL).  This can be done with a quick idnum
check so that only the original player (the first imp, he with id 1) can use
the command to get to maximum level.

ACMD(do_upme)
{
  if (GET_IDNUM(ch) != 1) {
    send_to_char("You think IMP positions are that easy to come by?  Go Figure...\r\n", ch);
  } else {
    GET_LEVEL(ch) = LVL_IMPL;
    send_to_char("Advanced.\r\n", ch);
  }
}


4.14.  I decided to wipe my pfile away anyway.  What steps should I take to
       do this?

In order:
1. Shutdown the MUD with ``shutdown die'' so that it won't restart.
2. Remove the player file, lib/etc/players
3. Restart the MUD and login to recreate your imp character.

You should probably also remove files in plrobjs, unless you want the
recreated characters to come back with the same equipment they had when they
were deleted.


4.15.  I want to expand the ability to pk in my MUD, allowing ASSASSINS
       that'll be able to PK without getting flagged.  How can I do this?

The simple way to do this is to find all the ``pk_allowed'' checks and
replace them with a can_murder(ch, vict) function call.  Prototype the
function in utils.h.  Then, in utils.c, create a can_murder() function
something like this:

int can_murder(struct char_data *ch, struct char_data *victim) {
  if (pk_allowed == TRUE)
    return TRUE;
  if (IS_NPC(ch) || IS_NPC(victim))
    return TRUE;   /* you can always kill these */
  if (PLR_FLAGGED(ch, PLR_ASSASSIN))
    return TRUE;   /* they can kill anyone */
  /* Add further checks here */
}


4.16.  Why does it say ``Connection closed by foreign host.'' and not
       display the ``Byebye!'' message I'm trying to send before cutting
       someone off?

This usually happens if you are doing something like this:

  send_to_char("Bye bye.  Come back soon, ya hear?", ch);
  close_socket(ch->desc);

The close_socket immediately dispatches/closes the connection, while
send_to_char puts the message on the output queue to be dispatched next
game_loop cycle. Therefore, the socket is gone.  On some systems (ie old
linux), this can even cause a infinite loop attempting to write to a closed
socket. The proper way of doing this and other ``Byebye'' messages is to set
the CON state of the player to CLOSE, like this:

  send_to_char("Bye bye.  Come back soon, ya hear?", ch);
  STATE(ch->desc) = CON_CLOSED;

This will then cycle to the next game_loop, dispatch the output queues
(therefore sending the byebye message) and then close the socket. Further
note, in some bizarre cases, this only seems to send about 40 characters and
no escape codes.  Sending more than 40 characters or escape codes (like the
clear screen sequence) will crash the process reporting a problem similar to
writing to a closed socket.


4.17.  I run my MUD on a unix system and the pfile/rent file works great,
       but on my home system it's all screwed up.  What gives?

Some operating systems write binary data least-signifigant-digit- first,
while others write it most-significant-first (big endian vs. little endian).
Moving player files, rent files, mail files, and board files from one of
these systems to the other will result in corrupted files.  The solutions to
this problem include:

o  Don't bother trying to move those files.
o  Edit your code to always write in one format.
o  Develop a binary to ascii conversion tool (and ascii to binary) for
   all binary files.
o  Develop an ascii read/write system to allow portability.


4.18.  How do I get the CircleMUD to autoload when the Linux server is
       restarted?

In /etc/rc.d/rc.local find where things like sendmail and (maybe) gpm
are started.  Add something like:

cd /home/mudlogin/circlebpl19/
su mudlogin -c ./autorun &
cd

Of course, change the "mudlogin" to whatever the name of the account is that
normally has control of the MUD, and change the path in the first cd to
whereever the MUD is run from.

For more info: man su


4.19.  My server shuts down my MUD everytime I logoff. How do I keep the MUD
       running when I logoff?

Instead of typing "autorun &" to start the autorun script (which starts and
keeps the MUD running), type "nohup autorun &".  Running the autorun via
nohup will keep the script and the MUD running when you logoff of the
server. For more information type "man nohup" at the prompt on your server.


5.  Code Changes for CircleMUD 2.20

5.1.  How do I fix the bug where people can junk more coins than available?

Apprently in Circle 2.2, you can drop any amount of coins, and then be
rewarded with more coins than you had in the first place.  Here is the fix
from Jeremy Elson:

Around line 480 of act.obj1.c, you will find the code:

            if (!str_cmp("coins", arg) || !str_cmp("coin", arg))
              perform_drop_gold(ch, amount, mode, RDR);
            else {
              /* code to drop multiple items.  anyone want to write it? -je */
              send_to_char("Sorry, you can't do that (yet)...\n\r", ch);
-->           return;
            }
          } else {
          ....

It should be changed to:

            if (!str_cmp("coins", arg) || !str_cmp("coin", arg))
              perform_drop_gold(ch, amount, mode, RDR);
            else {
              /* code to drop multiple items.  anyone want to write it? -je */
              send_to_char("Sorry, you can't do that (yet)...\n\r", ch);
            }
-->         return;
          } else {
          ....


5.2.  How do I fix the ``vstat'' bug that crashes the MUD?

To the fix for the vstat bug, from Jeremy Elson:

In the file act.wizard.c, in the function do_vstat, in the mobile section of
the switch (around line 1150), you'll find the code:

        mob = read_mobile(r_num, REAL);
        do_stat_character(ch, mob);
        extract_char(mob);

Add the line char_to_room(mob, 0) before extract_char(), like this:

        mob = read_mobile(r_num, REAL);
        do_stat_character(ch, mob);
        char_to_room(mob, 0);
        extract_char(mob);


5.3.  How do I fix the ``wizlock'' bug that lets lower immortals lock
      out higher immortals?

Simple, insert this code into 'do_wizlock()' in 'act.wizard.c':

        if (value < 0 || value > LEVEL_IMPL) {
          send_to_char("Invalid wizlock value.\n\r", ch);
          return;
        }

+        /*   Do not allow people to wizlock above their level.
+        *   This bug came with Circle 2.20 source -- VampLestat
+        */
+        if (value > GET_LEVEL(ch)) {
+          send_to_char("You may only wizlock below your level.\n\r", ch);
+          return;
+        }
        restrict = value;


5.4.  How do I fix the ``mudlog'' bug that lets people see me log in,
even if I'm wizinvis?

For all the mudlog calls for entering the game, quitting, and so
forth, you must change

        mudlog(MAX(LEVEL_IMMORT, ...

to

        mudlog(MAX(GET_LEVEL(i), ...

where ``i'' is either ``ch'' or ``d->character'' depending on the
call.


6.  CircleMUD 3.0 Questions

6.1.  Are there any bugs in patch level 19?

There are no bugs.  Only features.  Seriously though, if perchance you find
a bug, please mail it (along with a possible fix) to the bugs database
<bugs@circlemud.org> and the CircleMUD list.  Once in the bug database, it
will be routed to the correct person.  Note that no confirmation is sent
back automatically.  You can view the status of the bug at the bugs database
(http://bugs.circlemud.org/).


6.2.  How do I access Online Creation?

Online Creation is not yet part of the Circle beta release. When it does
become part of the release, it'll be accessed through a command called olc.
Update: OLC will probably be in release 3.10 now, instead of 3.0. In the
mean time, check out the CircleMUD FTP Server:

ftp://ftp.circlemud.org/pub/CircleMUD/contrib/olc/


6.3.  How does the new bitvector system work?

The new bitvector system in CircleMUD 3.0 is an ascii based one. The old
numeric values can still be used, but the new system should make sorting
flags out substantially easier.  The system uses an ``a'' as the value for
1, ``b'' for 2, ``c'' for 4, ``d'' for 8, etc.  Once ``z'' is reached, the
next letter in sequence is ``A''.  Detailed information about how to use
bitvectors with CircleMUD can be found in the CircleMUD Builder's Manual
(http://www.circlemud.org/~jelson/cdp/building/).


6.4.  When will the production release of Circle 3.0 be?

I don't know.  Don't ask again.  Same for the next patch level.  If you must
have bleeding edge stuff, check out the CVS Snapshot
(ftp://ftp.circlemud.org/pub/CircleMUD/cvs/) until we get an online
anonymous CVS.  And don't ask when the next CVS snapshot will be, we'll do
it when we have some changes worth it.


6.5.  If someone logs in and just sits at the password prompt, the MUD hangs
      (i.e., no one else can connect or do anything) until the person enters
      their password.

Your system's POSIX non-blocking I/O might be broken.  Look in the source
file sysdep.h at the comment above the line that says ``#define
POSIX_NONBLOCK_BROKEN" for a possible fix. Once this is done, recompile the
MUD and try again.  If you use the POSIX_NONBLOCK_BROKEN constant and it
fixes your problem, please send mail to the bugs database
<bugs@circlemud.org> and let us know exactly what kind of system you are
using and what you had to do to fix it.


6.6.  Why does CircleMUD use BUF switches all through the code, what's
      happening here?

From Jeremy:

This code is the new output buffering system that I wrote for Circle in the
early (non-released) beta versions of 3.0.  The old Diku code for queueing
output (which stayed with Circle until version 2.20) was memory- and
time-inefficient in many cases (and, in my opinion, was inefficient for the
normal behavior of most MUDs).

First, I should explain what output queueing is and why it is necessary.  On
each pass through the game_loop(), the MUD performs a number of steps: check
to see if there are any new players connecting, kick out people with bad
links, read input over the network for all players, then process the input
for each player that has sent a complete line over the net.  The processing
step is usually where output is generated because it is where MUD commands
are processed (e.g., ``kill'' might generate output of ``Kill who?'') When
output is generated, it is not immediately sent out to the player, but
instead queued for output in a buffer.  After all players' commands are
processed (and each command generates the appropriate output for various
players), the next step of the game_loop() is to send all the queued output
out over the network.

The new output system that Circle now uses allocates a small, fixed size
buffer (1024 bytes) for each descriptor in which output can be queued.  When
output is generated (via such functions as send_to_char(), act(), etc.), it
is written to the fixed size buffer until the buffer fills.  When the buffer
fills, we switch over to a larger (12K) buffer instead.  A ``buffer
switch'', therefore, is when the 1024-byte fixed buffer overflows.

When a large (12K) buffer is needed, it is taken from a pool of 12K buffers
that already been created.  It is used for the duration of that pass through
the game_loop() and then returned to the pool immediately afterwards, when
the output is sent to the descriptor.  If a large buffer is needed but none
are in the pool, one is created (thereby increasing the size of the pool);
the ``buf_largecount'' variable records the current pool size.

If a player has already gone from their small to large buffer, and so much
output is generated that it fills even the large buffer, the descriptor is
changed to the overflow state, meaning that all future output for the
duration of the current pass through the game loop is discarded.  This is a
buffer overflow, and the only state in which output is lost.

Now that I've described how the system works, I'll describe the rationale. 
The main purpose for the two-tiered buffer system is to save memory and
reduce CPU usage.  From a memory standpoint: Allocating a fixed 12K buffer
for each socket is a simple scheme (and very easy to code), but on a large
MUD, 100 12K buffers can add up to a lot of wasted memory.  (1.2 megs of
memory used for buffering on a 100-player MUD may not seem like very much,
but keep in mind that one of Circle's big selling points several years ago,
when memory was expensive, was that it had a very small memory footprint (3
or 4 megs total!) And from a CPU standpoint: the original Diku used a
dynamically allocated buffer scheme to queue output, which unfortunately
meant that for each player, on each pass through the game loop, dozens of
tiny buffers (often one for every line of output, depending on the code to
execute the command) were allocated with malloc(), individually written to
the system using individual calls to write(), and then free()'d.  My system
saves hundreds or thousands of calls per second to malloc() and free(), and
reduces the number of system calls *drastically* (to at most one per player
per pass through the game loop).

The trick is to choose the size of the small and large buffers correctly in
order to find the optimal behavior.  I consider ``optimal'' to mean that 90%
of the time, most players stay within the limits of their small buffer (for
example, when wandering through town or mindlessly killing some monster
while watching damage messages go by).  Hopefully, a large buffer switch is
only necessary when a player executes a special command that generates an
unusually large amount of output, such as ``who'', ``read board'', or
``where sword''.  This critically depends on the fact that not everyone will
be executing such a special large-output command at the same instant.

For example, imagine you have 10 players on your MUD.  They are all
wandering around town, and every once in a while one of them types ``who'',
or reads the board, meaning that they are seeing more than 1024 bytes of
output at a time.  On such a MUD, I would hope that there would only be a
single 12K buffer allocated which gets passed around among all the 10
players as needed.  Now, all players think they can queue up to 12K of
output per command without getting truncated even though only one 12K buffer
actually exists -- they are all sharing it.

But - there's a problem with this.  There are certain cases when many
players have to see a lot of output at the same instant (i.e. on the same
pass through the game_loop()), all of them will need a large buffer at the
same time and the pool will get very big.  For example, if an evil god types
``force all who''; or if the MUD lags for several seconds, then suddenly
gets unlagged causing many commands to be processed at the same moment; or
if 20 people are all trying to kill the same MOB and are all seeing 20
damage messages (more than 1024 bytes) on the same pass through the
game_loop().

Unfortunately, the current patchlevel of Circle has no way to destroy large
buffers so such cases are pathological and cause wasted memory.
Unfortunately since I don't run a MUD I can't actually tell how often this
happens on a real MUD.  (If there are any IMPs out there who run large MUDs
(say, >= 30-50 players on regularly), and you've read this far, please send
me the output of ``show stats'' after your MUD has been played for at least
several hours.)

Hopefully this clears up the way buffers work.


6.7.  How do I add a new class?  How do I add more levels?

Adding a new class is fairly easy in 3.0, in fact, someone has taken the
time to put together a fairly complete document on adding a class, in this
case a Knight class. The class.doc is available on the FTP site:

ftp://ftp.circlemud.org/pub/CircleMUD/contrib/docs/3.x/class.txt

The same applies for levels, in the file levels.doc
(ftp://ftp.circlemud.org/pub/CircleMUD/contrib/docs/3.x/levels.txt)
which can be found in the same place.


6.8.  Is there a complete ``coding.doc''?

No.  Look at the README.NOW in the docs directory.  It isn't complete, nor
does anyone have the entire thing.  It is being worked upon.

RTFM