2번째 디스켓 

odin32bin-20031120-release.wpi

odin32bin-20040616-debug-uo.zip

odin32bin-20040616-release-uo.zip 

출처 :  ftp://testcase.software.ibm.com/ps/fromibm/os2

**********

Tivoli Management Agent 

클라이언트팩 2번째 시디에있던것입니다.

---------

lan distance입니다.. WSeB 시디 2번째 클라이언트팩 시디에있던것입니다..

                O D I N  Public Alpha 5 - 1999/11/16

                ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

WANTED: We are still looking for talented developers that want to

        contribute! Please take a look at the last section!

NOTE:   LICENSE.TXT describes the Project Odin Software License.

Project Odin Information

========================

1.0 Introduction

----------------

This readme is intended for end users that have downloaded the

latest public alpha release and want to give it a try. For more

information, be sure to regularly check http://www.netlabs.org/odin.

1.1 Purpose

-----------

Odin is a free Open Source project with the goal to eventually

provide a fully working Win32 (Win 95/98/NT/2000) runtime environment,

making it possible to run your favorite Windows programs under OS/2,

as if they were native OS/2 programs. The second goal of Odin is to

provide the full Win32 application programming interface (API) so

that it is possible to build OS/2 programs using Windows source

code. This is also refered to as Odin32 whereas Odin is the name

of the Win32 binary emulator.

1.2 Current Status

------------------

Odin is still in the early alpha state so that it is very far from

being widely usable. Most programs will not work or show many

problems. Even though it is not usable for most programs, we have

decided to do a public release in order to show how much progress

the project has been made. A second goal is to attract more developers

in order to speedup development. Our job includes rewriting most

parts of Windows 95/NT without having the original source code so

the more developers we have the more progess will be seen. The

last public release (Alpha 0.004) has been in 1997 and the project

has changed much since then. Everything is free and Open Source now

and the base architecture has been improved quite a lot. We are now

in a position where we can expect the first bigger applications to

run well and to make much more visible progress.

1.3 Changes and additions in alpha 5

------------------------------------

- Removed Open32 dependacy in User32 code: windows, dialogs, buttons and

  other controls creation and manipulation is now implemented fully in Odin32.

  Applications have Win32 look: Win32 specific controls are also supported

  (Open32 is still a requirement as other Odin32 functions use it!)

- New and improved PE2LX converter

- New Portable Executable Loader (PE) that overcomes some differences

  between OS/2 and Win32 memory models (this does not include 512MB

  per-process memory limitation in OS/2, you still need the Aurora-level

  kernel for that)

- New WIN32K32k.SYS driver that automates loading of Win32 binaries (no

  conversion or invocation through PE.EXE loader is needed), now

  compatible with WSeB (Aurora) and no longer dependant of code under

  GNU license

- Wine ports of common controls (comctl32), shell extensions (shell32) and

  OLE

- Full implementation of memory mapped files

- Joystick support (winmm)

- Direct Input (DirectX keyboard & mouse) added (Wine Port)

- Rewrite of DirectDraw. More complete implementation.

- Couple of (not core Win32 API, but often used) new DLLs (winasp32, lz32,

  avifil32...), so users don't have to have Windows license at all to run

  Windows programs

- Many new API functions implemented or ported from Wine, allows

  developers to port Win32 apps directly to OS/2 with little to no rewriting

- Many bugfixes and enhancements

- Many new programs that load and/or work with Odin

- Full source code available

2.0 Installing Odin (UPDATED)

-----------------------------

Daily builds of Odin are available at ftp://ftp.os2.org/odin/daily.

For people that do not have a suitable development setup for ODIN,

we provide the latest binary code that way.

Odin daily builds are now also distributed as WarpIn archives.

WarpIn is a generic OS/2 installation application that can be

downloaded from warpin.netlabs.org.

Requirements for installing Odin:

- You must install Odin on a partition that supports long filenames (e.g.

  HPFS/JFS). Odin will not work on FAT!

- OS/2 Warp 4, eComStation or Warp Server for e-Business (WSeB)

  When using Warp 4, we recommend you apply fixpack 13 or higher.

  * Warp 3 is not officially supported, but is reported to work by using

    Warp 4 open32 dlls and patching them.

Manual installation without WarpIn is not recommended. If you don't use WarpIn,

then you *must* execute odininst.exe in the Odin system32 directory. (the first

time Odin is installed or when the ChangeLog mentions fixes/changes in the

installer)

NOTE: You must install Odin on a partition that supports long filenames (e.g.

      HPFS/JFS). Odin will not work on FAT!

To install Odin, run warpin.exe with the name of the daily build archive

as parameter.

Follow the instructions to complete the installation.

The Odin installation consists of:

  - creating an Odin directory and copying the included files to

    x:\odin and x:\odin\system32

  - creating empty directories in x:\odin (to create a similar directory

    structure as found in Windows)

  - creating a basic registry for executing win32 applications in OS/2

  - adds x:\odin\system32 to the path and libpath settings in your config.sys

NOTE: You MUST install Odin at least once using WarpIn (unless you know

      what you're doing). Afterwards you can use the daily build zipfiles

      to upgrade to a newer version.

To run DirectDraw games or Quake 2 in Voodoo mode, you must install Holger

Veit's XFree86/2 support driver (XF86SUP.SYS).

Visit http://ais.gmd.de/~veit/os2/xf86sup.html to download it.

NOTE: The Glide and 3Dfx Opengl dll are no longer part of the basic Odin

      installation package. It's available as opengl3dfx.zip on ftp.os2.org.

2.0.1 Log files

---------------

The alpha 5 binaries and daily build zipfiles can generate logfiles to show

what a win32 application is doing. This can be very useful to determine

why certain applications don't run correctly.

The major disadvantage of loggging is the overhead. Therefor it has been

disabled by default in the alpha 5 release and daily builds.

To enable logging set the environment variable WIN32LOG_ENABLED:

   SET WIN32LOG_ENABLED=1

To disable logging again, you must clear this variable:

   SET WIN32LOG_ENABLED=

2.1 Using Odin to run your Windows applications

-----------------------------------------------

Odin still has several ways to execute Win32 programs.

2.2 The Ring3 Win32 Loader (PE.EXE)

----------------------------------

This is the new and also recommended method for loading programs.

Unlike with the previous alphas, it is not required to convert or

anyhow change the original files. Installing a device driver is

also not necessary. Be sure to put the Odin binary directory into

both the PATH and LIBPATH (alternatively you can use a batch file

with SET PATH=C:\Odin;%PATH% and SET BEGINLIBPATH=C:\Odin;)

and then call pe.exe winprog.exe param1 param2 ... If you wanted

to start the Windows calculator, you would type

pe calc

The PE (Portable Executable, the Win32 executable format) loader

will now try to load the application into memory and execute it.

This will either work or fail with a loader message. If you get

such a loader message, it is very likely that a required DLL is

missing. To correct that problem, take a look at pe_0.log (or

pe_1.log for the second program running at this time and so on).

At the bottom of the file you will see the DLLs and the functions

that the loader imports. It will tell you if it couldn't find a

DLL and you can correct the problem and try it again.

2.2.1 PE loader issues

----------------------

As the PE loader will create the process and load the program,

a process utility such as pstat will always show PE.EXE as the

executable file.

It is also possible that a certain Win32 application becomes

a zombie (that is when it does certain things the OS/2 kernel

cannot recover such as die when waiting on a semaphore). Then

it will not be possible to kill the application - only a system

restart will help.

Certain programs do use fixed memory addresses where they need

to be loaded (i.e. they won't use fixups). You can notice this

at the top of the loader log file (pe_0.log). If the fixed address

is above the 512MB barrier present in Warp 4 and Warp 3, the

program can not be run under these system. On Aurora (Warp Server

for e-business) this limitation has been removed. Programs known

to fall under this category are Microsoft Office (also the

Office viewers) and other mainly Microsoft programs.

2.3 Win32k.sys - native OS/2 PE-loader

--------------------------------------

The win32k.sys driver makes PE (Win32) executables equal to native OS/2

executables by implementing a Ring 0 on-the-fly loader / converter.

You may wonder how you may run a Win32 program using a device driver. You

don't have to worry about that, simply treat the Win32 program just as an

another OS/2 program. OS/2 can't tell the difference!

No conversion of dlls or running programs thru Pe.EXE.

For example, if you fancy playing windows solitaire.

    1) Make sure the necessary DLLs (MSVCRT.DLL and CARDS.DLL) are in you

       PATH, current dir, SOL.EXE dir, Odin32 System32 dir, Odin32 System dir,

       Odin32 dir, or LIBPATHs; and that you have SOL.EXE somewhere.

    2) Then start SOL.EXE by simply typing SOL.EXE (with path if necessary).

       It's also possible to start SOL.EXE from the desktop if you like.

       Simply find the SOL.EXE icon (or create it) and double click on it!

2.3.1 WARNING

-------------

The Win32k.sys is quite alpha software and any bugs which haven't been found

yet may crash and/or corrupt you entire system! Make sure you have backed up

important stuff first!

We are not responsible for ANYTHING AT ALL! See Odin32 disclaimer...

Before you install Win32k.sys make sure that you have a backdoor to your system

which works. For example try to boot your system using the Alt+F1 during bootup

and choose Commandline boot (F2). Make sure you have an textmode editor which

works; for example TEDIT.EXE.

2.3.2 Requirements

------------------

Win32k should at least work on OS/2 Warp 3 and above, but I don't get too

surprised if it worked on 2.11 too. DBCS are not supported yet AFAIK.

It no longer requires a kernel file for retail kernels (if you don't know what

kernel you have, you have a retail kernel). Win32k does now contain a

"database" for the most recent retail kernels. Currently this is:

    Warp 4 fp#1 -> fp#14

    Aurora GA, FP1, preFP2

    Warp 3 fp#32-> fp#42

If the running kernel is not found in the "database" or is a debug kernel

wink32.sys will look for symbolfiles as it allways did.

If you use the -s:<symfile> argument the given symbol file will be

investigated before the database is searched.

NOTE! It hasn't been fully verified that all the different national fixpaks

have the same symbols as the us-english ones. Feedback on this issue is

appreciated.

NOTE! A piece of advice for those playing with debug-kernels: Make sure that

c:\OS2KRNL.SYM is matching the running kernel.

2.3.2.1 Warp 3

--------------

Win32k is not tested with Warp 3. It is _supposed_ to work with Warp 3

provided you are running fixpack 32 or later.

(Fixpack 32 + y2kclock.exe is required for Y2K readyness anyway so...)

2.3.3 Installation

------------------

To use Win32k.sys you have to add the following line into your config.sys:

    device=<d:\path\system32>win32k.sys

Where <d:\path\system32> is the path to the win32k.sys, i.e. usually the odin32

system32 directory.

After doing this, you'll have to reboot your machine to load the driver. During

the boot process you will on a successful installation see some lines of

technical info and finally a line saying:

Win32k - Odin32 support driver.

    Build:            14046 - v20.45

    Found symbolfile: Win32k Symbol Database

Win32k.sys succesfully initiated!

If you don't see the last line something is wrong, and some error info will be

shown. In the case of error the 'rc=<value>' is important to remember.

2.3.4 Win32k parameters

-----------------------

The Capitalized letters are the one which is tested for when identifying the

argument.

There are some useful parameters for Win32k:

-Quited         Quiet driver initiation.

                Default: disabled (ie. -V)

-Verbose        Verbose driver initiation.

                Default: enabled

-S:<filename>   Full path to the current OS/2 kernel symbol file.

                When this is specified the given

                Default: symbol "database" and default locations are searched.

Loader parameters:

-SCript:<Y|N>   Enable / disable UNIX script support.

                Default: Enabled.

-REXX:<Y|N>     Enable / disable pure REXX script support.

                Default: Enabled.

-Java:<Y|N>     Enable / disable JAVA loading.

-Pe:<MIXED|PE2LX|PE|NOT>

                Enabled / disable and set mode of the PE loading.

                    MIXED:  Mix of Pe2Lx and Pe.exe. It uses Pe2Lx when

                            possible and backs out to pe.exe if the executables

                            aren't supported by the converter.

                    PE2LX:  Pe2Lx loader only.

                    PE:     Pe.exe only. The program is started using Pe.exe,

                            and the process name will be pe.exe.

                    NOT:    Disable PE loading.

Logging parameters:

-L<:E|:D>       Enables/Disables logging.

                Default: disabled.

-Com:<1|2|3|4>

or

-C<1|2|3|4>     Which com-port to log to.

                Default: -C2

-Warning:<1|2|3|4>

or

-W<1|2|3|4>     Warning/info level for PE loader. (This is warnings and info

                which is published thru the com-port as logging information.)

                Default: -W0

Heap parameters:

-HEAPMax:       Maximum swappable heap size in bytes.

                Default: 512KB

-Heapinit:      Initial heapsize in bytes.

                Default: 16MB

-RESHEAPMax:    Maximum resident heap size in bytes.

                Default: 256KB

-Resheapinit:   Initial resident in bytes.

                Default: 16MB

OS/2 Loader parameter:

-Dllfixes<:E|:D>

                Enables or Disables the long DLL name and non .DLL

                extention fixes. OS/2 DLLs have been limited to 8 chars

                names since version 2.<something>. This is a very bad

                limitation. Also OS/2 forces all DLL loaded to loadtime

                to have the extention '.DLL'.

                Both these retritions are removed if this option is

                enabled.

                For DLL with non .DLL extention the internal name must

                have to the extention included. (If you the experiences

                problems with IMPLIB.EXE, enclose the internal name

                in double-quotes.)

                You should not make a DLL with a name starting with

                DOSCALLS, DOSCALLS1, IBMVGA32, IBMDEV32 or IBMXGA32.

                Current state for this feature is EXPERIMENTAL.

                Default: Enabled

Obsolete:

-K:<filename> Ignored. Win32k scans the kernel image in memory!

2.3.5 PE Loading

----------------

Win32k.sys are able to do PE loading in two different ways.

2.3.5.1 PE2LX

-------------

The first one is called PE2LX and hence uses the original Win32k.sys method

which is similar to Pe2Lx. This only requires that KERNEL32.DLL (ie. Odin32

System32 directory) is present somewhere in the PATH or LIBPATHs. The DLL

search order of PE2LX is now conforming with the Win32 order:

    1. The directory of the executable (.EXE).

    2. The current directory.

    3. The Odin32 System32 directory. (Kernel32.dll directory)

    4. The Odin32 System directory. (sibling of system32)

    5. The Odin32 directory. (parent of system32)

    6. The PATH environment variable. (TODO: quotes are not currently supported)

    7. The BEGINLIBPATH.

    8. The LIBPATH.

    9. The ENDLIBPATH.

(Note that the OS/2 DLLs in the _Odin32_ process are also found using this

search order. This shouldn't matter much as nearly all of the OS/2 DLL are

allready loaded. But if this proves a problem it could be changed.)

2.3.5.2 PE

----------

The second one is called PE and invokes the Pe.EXE to execute the PE (Win32)

programs. This of course requies that PE.EXE is found in the PATH.

2.3.5.3 Modes

--------------

Win32k.sys have three (or four) different ways of combining these two methods,

called modes:

    PE2LX only

    PE only

    PE2LX with fallback on PE.

    (fourth: NOT load any PE executables at all)

The last one is the default one and is probably the best one currently.

2.3.6 UNIX Shell Script Loader

------------------------------

The UNIX Shell Script Loader makes OS/2 able to run UNIX shell scripts as

native executables, just as UNIX does. A unix shell script is a text file

which has a starting line which specifies the interpreter for the script.

For example a perl script:

#!/bin/perl5 -Sx

Or the most common one:

#! /bin/sh

Since Most OS/2 systems don't have a /bin directory or a /usr/bin directory

the executable specified are looked for in these ways:

    1. The exact filename specified.

    2. The filename specified with .EXE extention.

    3. Search the PATH for the filename (with out path of course).

    4. Search the PATH for the filename with .EXE extention (and with out path of course).

So, you could now execute UNIX shell script in OS/2 without even modify the

path of the interpreter!

The UNIX Shell Script Loader is enabled by default.

2.3.6 Pure REXX Loader

----------------------

The Pure REXX Loader is a win32k.sys feature which enables REXX script with

the extentions .RX and .REX to run as native OS/2 executables. These REXX

scripts are executed using kRX.EXE which is found in the Odin32 System32

directory.

When we say "Pure REXX" scripts, that means no use of command line should

occure in the script.

The Pure REXX Loader is enabled by default.

2.3.7 Java Loader/Invoker

-------------------------

Just for fun, I've added some code which detects JAVA class binaries and

executes the class using JAVA.EXE. I first thougth this should be easy, but

it turned out to have several problems.

So please note the following:

    - The classname is case sensitive, you have to write the exact class

      name (with .class extention or what ever extention the file have)

      when trying to execute a class. This could cause some throuble making

      WPS program objects for the classes....

    - To support executing of .class-files which are not in the current

      directory win32k.sys send the directory of the .class-file in to

      java.exe as a '-classpath' option. This option I soon found out had the

      sideeffect of making JAVA.exe ignore the other CLASSPATH and not use

      it's default class path. So now the contents of the CLASSPATH env. var.

      or a generated default classpath is also sent in with the '-classpath'

      option.

    - The CLASSPATH environment variable should not be longer than 639 chars.

When you have successfully installed Win32k.sys try type this at the

commandline:

c:\java11\demo\hello\hello.class

This is then printed at my screen:

Hello, Warped world!

The JAVA Loader/Invoker is enabled by default.

2.3.8 Rings - Ring 0 / Ring 3

-----------------------------

FYI. Ring 0 is the most priveleged level of execution in OS/2. The OS/2 kernel

and device drivers execute at this level, while user applications execute in

Ring 3, the least priveleged executing level.

2.3.9 Win32k Control Center

---------------------------

In your Odin32 System32 directory there is a program called Win32kCC.exe,

this is the Win32k Control Center (Win32kCC). Win32kCC is as the name says

a program controlling Win32k.sys.

Win32kCC allows you to enable, disable and change behaviour of different

parts in Win32k.sys.

For example: You could disable PE loading, or change it from MIXED mode (see

2.3.4, -PE) to PE2LX mode. Or change it from PE2LX mode to PE mode to see

if PE.EXE does a better job loading an executable.

Note that the changes would not effect allready loaded executables.

Win32kCC also displays status info on the running Win32k.sys and OS/2 kernel:

    -Win32k.sys version number

    -Win32k build time/date

    -Path of the symbol file which was used (or Win32k Symbol Database).

    -OS/2 kernel build number fixpack (guess)

    -OS/2 kernel type as win32k has detected it

    -Number of Pe2Lx currently loaded modules

    -Win32k resident and swappable heap stats.

PS. Beside the "PE Loader" area there is an "Elf Loader" area this is for a

future Elf2Lx loader.

2.4 The Ring3 conversion utility (Xx2Lx.exe)

--------------------------------------------

Xx2Lx.exe was the first method of running Win32 programs under

OS/2. It is still provided and also functional but it has been

deprecated. Using it is only sensible for debugging certain

problems as it requires all DLLs and EXEs to be converted into

another file format thus changing the original Win32 files.

(It's only future use will be to convert Win32 executable with

debuginfo to LX with HLL (IPMD/ICSDEBUG) debuginfo for

sourcelevel debugging of Win32 executables under Odin32. This

is not yet implemented but is being worked on.)

Note. The name changed from Pe2Lx to Xx2Lx. It is in prepare for Elf2Lx.

3.0 Applications known to be working

------------------------------------

- Windows 95/NT Help Viewer (winhlp32.exe)

- Windows 95/NT Registry Editor

- Windows NT Solitaire (not the one from Windows 95)

- Windows NT Freecell (not the one from Windows 95)

- Windows 95/NT Calculator

- Windows NT Pinball

- Windows NT Minesweeper

- Quake II

- Red Alert

- Professional Minesweeper

- Windiff (part of Win32 SDK)

- RealPlayer 7/8

- Lotus Notes 5.0 (partly anyway)

- Acrobat Distiller 3.0

Note that many other applications load and/or partly work.

4.0 Reporting Problems and Successes

------------------------------------

Bug reports can be sent to the Odin users mailinglist (odinusers@egroups.com;

visit www.egroups.com for information about joining).

The file ReportingBugs.txt explains how to send in a proper bug report.

An application compatibility database has been created at Netlabs

where you can share your experiences with Odin.

The application compatibility database can be found at

http://www.netlabs.org/odin/appdatabase

NOTE: The application database is not meant for bug reports.

5.0 Project Odin Source Code Notes

----------------------------------

5.1.0 Introduction

------------------

This chapter is intended for people more familiar with ODIN. It describes

the various aspects of the project such as accessing the latest sources via

cvs, building the binaries, running win32 applications, how to identify and

report problems and bugs, as well as how to participate in this very

promising project.

5.1.1 Prerequisites:

--------------------

Required compilers/tools:

        - IBM VisualAge 3.0 (CTx8 fixpack applied)

          (version 3.6 might also work; version 4 is not recommended)

        - OS/2 Warp 4 Toolkit (project apparently doesn't compile with

          VAC's OS/2 headers)

          Might also work with EMX headers. (haven't tried this!)

        - ALP 4.0 (IBM Assembly Language Processor)

          Download the tools zipfile from http://service.boulder.ibm.com/ddk/

          (you need to register, but it's free)

        - Unix-like rm.exe (nmake clean)

Changes from last source code release:

        - Uses Wine headers and a stripped down version of os2win.h (Open32

          header)

        - All Open32 apis have the 'O32_' prefix.

        - All Open32 declarations are appended with '_O32' (mostly due to

          calling convention differences)

        - nmake DEBUG=1 builds the debug version

          nmake         builds the release version

        - Several dlls (i.e. USER32, KERNEL32, ADVAPI32) contain extra .def

          files that are used to create libraries that contain the api names

          as required by the linker when you use them. (_apiname@nrparameterbytes)

Wine header notes:

        - Several header files have been changed to make them usable for VAC

        - Several header files contain bugfixes (mostly parameter type changes)

        - Several header files include additional declarations

        - At this time we don't intend to have our changes included in the

          Wine codebase as header files are not likely to change often.

Coding conventions:

        - Do NOT directly call the Open32 apis unless you're writing

          a wrapper for it.

        - Do NOT include os2.h directly. Use os2wrap.h instead.

          os2wrap.h contains all the OS/2 apis with a wrapper around it to save

          and restore the selector in FS. (some OS/2 apis set it back to the

          default value (150b))

        - Use the correct Win32 api when there are two version (ascii & unicode)

          Calling the standard one (SendMessage) causes a error during

          compilation.

          Use SendMessageA instead.

        - Source code style will be decided on in the near future.

5.2.0 Accessing ODIN source code via CVS

----------------------------------------

Please check out http://www.netlabs.org/ and http://www.netlabs.org/odin.

5.3.0 Building the binaries

---------------------------

The first time you want to build Odin, you must first run configure.cmd.

This REXX script creates the file makefile.inc, which contains the absolute

paths of important build directories.

If your system setup meets above stated requirements, everything you

need to do next for a binary build is:

  nmake DEBUG=1

or alternatively

  nmake release (not working right now!)

As the current state of the project is still known to be alpha level, one should

expect some functions not to work properly. Therefore, we recommend the use of

the debug-version which creates long logfiles for problem determination purpose

and to allow active debugging with debuggers such as IPMD, ICAT, KDB, etc.

Please note if you do not want to suffer from the performance loss of writing

logfiles, you can disable logging even with the debug binaries of ODIN by

setting:

  SET NOWIN32LOG=YES

To enable logging again, clear this environment variable by typing:

  SET NOWIN32LOG=

Logging is enabled by default in the debug build, but disabled by default

in the nodebuginfo build and not present in the release build.

5.3.1 Build problems

--------------------

ODIN is moving at a fast pace. Sometimes side effects of source dependencies may

require you to do a full rebuild of the binaries instead of an incremental build.

However, rarely it also happens the sources contained in the CVS repository do

not build completely due to erroneous source code. Usually, this is fixed within

two days at maximum.

In such event, one can normally stick with the previous version of the particular

module to further test win32 applications on ODIN.

Please report such build problems through the appropriate channel.

5.3.2 Downloading daily built binaries

--------------------------------------

http://www.os2.org/ has setup an ftp server that hosts the automatically

built binaries on daily basis. For people that do not have a suitable

development setup for ODIN, we provide the latest binary code that way.

5.4.0 Project participation

---------------------------

As ODIN became an open source project, everybody is kindly invited to

contribute his/her share to the progress of the project. May it be

active coding, fixing bugs or just providing detailed information about

examined problems.

We suggest you subscribe to win32os2-wai and the corresponding mailing lists

on http://www.egroups.com.

In case you are interested in participating, every member of the project will

be happy to give you direction to the right places and to give a personal

introduction to further development of the particular modules.

6.0 Warranty

------------

EXCEPT AS OTHERWISE RESTRICTED BY LAW, THIS WORK IS PROVIDED

WITHOUT ANY EXPRESSED OR IMPLIED WARRANTIES OF ANY KIND, INCLUDING

BUT NOT LIMITED TO, ANY IMPLIED WARRANTIES OF FITNESS FOR A

PARTICULAR PURPOSE, MERCHANTABILITY OR TITLE.  EXCEPT AS

OTHERWISE PROVIDED BY LAW, NO AUTHOR, COPYRIGHT HOLDER OR

LICENSOR SHALL BE LIABLE TO YOU FOR DAMAGES OF ANY KIND, EVEN IF

THEY HAVE BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

스크립트 파일 

도스용 네트워크 드라이버도 세팅되어 있지만 모든 pc에 적용되지는 않으리라 봅니다. 설정이 되어지고 나면 고스트에서 ip주소로 상태 pc를 복구하거나 할수도 있습니다. 물론 같은 형태의 2개의 부팅이미지를 통해서 한쪽은 서버가 되고 한쪽은 클라이언트가 되면 됩니다.

작성자: 마드모아 

IBM Networks Coordinated Logon Client for Windows NT

인데..아마 엔티4.0 윈2000까지 가능할꺼 같습니다..

XP이후는 않될거같습니다..

즉 엔티를 OS/2에서 관리하는것이죠..벌써 2000나오기전에 

액티브디렉토리를지원하는것이죠...

-------------------------------------------------------

    IBM Networks Coordinated Logon Client for Windows NT        Page 1 of 5

    ________________________________________________________________________

                              Table of Contents

                        _______________________________

    1.0 Introduction .................................................Page 1

    2.0 Prerequisites ................................................Page 1

    3.0 Installation .................................................Page 2

        3.1 Installing from Diskette .................................Page 2

        3.2 Installing Windows NT Service Pack .......................Page 3

    4.0 Up and Running ...............................................Page 3

        4.1 Coordinating the User Accounts ...........................Page 3

        4.2 Logging On ...............................................Page 3

        4.3 Starting the Graphical User Interface ....................Page 4

        4.4 Displaying the Users Guide ...............................Page 4

        4.5 Keeping Your Password Synchronized .......................Page 4

    5.0 Uninstall ....................................................Page 5

    ----------------

    1.0 Introduction

    ----------------

        Welcome to the IBM Networks Coordinated Logon Client for Windows NT

        (hereinafter called IBM Networks Client). IBM Networks Client is an

        add on to the Microsoft Windows NT Network function.  It provides

        your Windows NT system with all of the IBM OS/2 Warp Server

        functions, such as:

          *  Connecting to home directories

          *  Setting up and connecting to logon assignments

          *  Running shared applications

          *  Connecting to aliases

          *  Checking DASD limits

        IBM Networks Client performs a secondary logon validation to an

        IBM OS/2 Warp Server domain after Windows NT successfully logs on.

        Your home directory and logon assignments are automatically connected.

    -----------------

    2.0 Prerequisites

    -----------------

        You must have the following prerequisite components installed on

        your system before you can install IBM Networks Client:

          *  The Microsoft Workstation service

          *  An appropriate network adapter device driver

          *  A supported protocol (either NetBEUI or TCP/IP)

    IBM Networks Coordinated Logon Client for Windows NT        Page 2 of 5

    ________________________________________________________________________

        To verify that you have the prerequisite components installed:

          1.  Click Start button, point to Settings, and then click

              Control Panel.

          2.  Double-click Network.

          3.  From the Services and Protocols tabs you can view the

              components that are currently installed.

    -----------------

    3.0  Installation

    -----------------

        3.1  Installing from Diskette

            To install from diskette:

              1.  Log on to the system as a user with Administrator privilege.

              2.  Click Start button, point to Settings, and then click

                  Control Panel.

              3.  Double-click Network.  The Network notebook is displayed.

              4.  Click the Services tab.

              5.  Click Add... button.  The Select Network Service window is

                  displayed.

              6.  Click Have Disk... button.  The Insert Disk window is

                  displayed.

              7.  Insert the IBM Networks Coordinated Logon Client for Windows

                  NT diskette into the A: drive, and then click OK.  The OEM

                  Option window is displayed.

              8.  Select IBM Networks Coordinated Logon Client for Windows NT,

                  and then click OK.  The IBM Networks Client Properties

                  notebook is displayed, after the files are copied.

              9.  Click the General tab.

              10.  In the Domain name box, type the name of the default logon

                  IBM OS/2 Warp Server domain.

              11.  In the Alternate domain name box, if you want to use an

                  alternate domain, type the name of the alternate domain.

              12.  Click the Advanced tab.

              13.  If you select the button to allow the domain to determine

                  which server will validate your logon, you may change the

                  Logon Broadcast Attempts and Logon Broadcast Wait settings.

              14.  If you select the button to specify the name of your logon

                  server, you must enter the name of the logon server.

              15.  Click OK.  The Network notebook is displayed.

              16.  Click Close.  The Network Settings Change window is

                  displayed.

              17.  Click Yes to restart your system.

    IBM Networks Coordinated Logon Client for Windows NT        Page 3 of 5

    ________________________________________________________________________

        3.2  Installing Windows NT Service Pack

            Before you install a Windows NT service pack on top of a

            workstation that has the IBM Networks Client currently

            installed, do the following:

              1.  Uninstall the IBM Networks Client.

              2.  Install the service pack on Windows NT.

              3.  Install the IBM Networks Client.

    ------------------

    4.0 Up and Running

    ------------------

        4.1  Coordinating the User Accounts

            Your Windows NT user account and your IBM OS/2 Warp Server domain

            user account must be synchronized.  You must have a Windows NT

            user account whose username and password also exist in a IBM OS/2

            Warp Server domain.  To coordinate the user accounts do one of

            the following:

              o  If you have a Windows NT user account with a username that is

                the same as your IBM Warp Server domain account username, then

                you must change the password on your Windows NT account to be

                the same as your password on the IBM Warp Server domain user

                account.

              o  If you DO NOT have a Windows NT user account with a username

                that is the same as your IBM Warp Server domain account

                username, then you must create one.  Use the Windows NT User

                Manager utility to create your Windows NT account.

        4.2  Logging On

            To logon:

              1.  After completing the installation process, restart your

                  Windows NT session.  The logon window is displayed.

              2.  Type your network password.

              3.  Click OK to begin validation of your IBM Networks Client

                  logon.

    IBM Networks Coordinated Logon Client for Windows NT        Page 4 of 5

    ________________________________________________________________________

        4.3  Starting the Graphical User Interface

            To start the graphical user interface:

              1.  Click Start button, and then click Programs.

              2.  Click IBM Networks Client.

              3.  Click IBM Networks Client.

            Note: To view and connect network resources, use the IBM

                  Neighborhood Browser Enabler for OS/2 Warp Server.

        4.4 Displaying the Users Guide

            To view the Users Guide:

              1.  Click Start button, point to Programs, and then click

                  IBM Networks Client.

              2.  Click IBM Networks Client Users Guide.

        4.5 Keeping Your Password Synchronized

            To change your password:

              1.  Log on to the Windows NT workstation.

              2.  From the Windows NT desktop, press CTRL+ALT+DEL.  The

                  Windows NT Security window is displayed.

              3.  Click Change Password.  The Change Password window is

                  displayed.

              4.  Type your old password, new password and confirm new

                  password.

              5.  Click OK.

              6.  Your password will be updated on Windows NT and then on the

                  IBM OS/2 Warp Server domain.

    IBM Networks Coordinated Logon Client for Windows NT        Page 5 of 5

    ________________________________________________________________________

    -------------

    5.0 Uninstall

    -------------

        To uninstall:

          1.  Log on to the system as a user with administrator privilege.

          2.  Click Start button, point to Settings, and then click

              Control Panel.

          3.  Double-click Network.  The Network notebook is displayed.

          4.  Click Services tab.

          5.  From the Network Services list, click IBM Networks Coordinated

              Logon Client.

          6.  Click Remove.

          7.  Click Yes.

          8.  Click Close.

          9.  Restart your system.

          10.  Log on to the system as a user with administrator privilege to

              complete the update of Registry entries and the removal of the

              IBM Networks Client files.

END OF DOCUMENT. 

  IBM Networks Client for Windows 95     

작성일: 2004/01/22 17:46:23

작성자: 마드모아 

IBM Networks Client for Windows 95

PS-윈9X를 OS/2로 네트워크 로그인하게해주는... ^^

-------------------------------------------------------

    IBM Networks Client for Windows 95                          Page 1 of 5

    컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴

                              Table of Contents

                        컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴?

    1.0 Introduction .................................................Page 1

    2.0 Prerequisites ................................................Page 2

    3.0 Installing from Diskette .....................................Page 2

    4.0 Installing from OS/2 Warp Server .............................Page 3

    5.0 Up and Running ...............................................Page 4

        5.1 How to Logon .............................................Page 4

        5.2 Starting the Graphical User Interface ....................Page 4

        5.3 Displaying the Users Guide ...............................Page 4

    6.0 How to remove IBM Networks Client ............................Page 5

    Note:  If you installed the Beta version of the IBM Networks Client for

          Windows 95, you must remove those files before you can install

          this version.  To remove the files, follow the instructions in

          section 6.0 How to remove IBM Networks Client on page 5.

    ----------------

    1.0 Introduction

    ----------------

    IBM Networks Client for Windows 95 is an add on to the Microsoft Windows

    95 Network function.  It provides your Windows 95 system with all of the

    IBM OS/2 Warp Server functions, such as:

        *  Connecting to home directories

        *  Setting up and connecting to logon assignments

        *  Running shared applications

        *  Connecting to aliases

        *  Checking DASD limits

    When you use IBM Networks Client as the primary network logon and you

    start your Windows 95 system, your network logon is validated and your

    home directory and logon assignments are automatically connected.

    IBM Networks Client for Windows 95                          Page 2 of 5

    컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴

    -----------------

    2.0 Prerequisites

    -----------------

        You must have the following prerequisite components installed on

        your system before you can install IBM Networks Client:

            *  The Client for Microsoft Networks

            *  An appropriate network adapter device driver

            *  A supported protocol (either NetBEUI or TCP/IP)

        To verify that you have the prerequisite components installed:

          1.  Click Start button, point to Settings, and then click

              Control Panel.

          2.  Double-click Network.

          3.  From the Configuration tab you can view the components that

              are currently installed.

    -----------------------------

    3.0  Installing from Diskette

    -----------------------------

        To install from diskette:

          1.  Click Start button, point to Settings, and then click

              Control Panel.

          2.  Double-click Network.  The Network notebook is displayed.

          3.  Click the Configuration tab.

          4.  Click Add... button.  The Select Network Component Type window

              is displayed.

          5.  Click Client, and then click Add... button.  The Select Network

              Client window is displayed.

          6.  Click Have Disk... button.  The Install From Disk window is

              displayed.

          7.  Insert the IBM Networks Client for Windows 95 diskette into the

              A: drive, and then click OK.

          8.  From the Select Network Client window, select IBM Networks

              Client for Windows 95 and then click OK.

          9.  Depending upon the service level of Windows 95 on your system,

              you may be prompted to insert your Windows 95 disk.  If

              prompted, insert your Windows 95 disk and then click OK.

          10.  After the files are copied, the Network window is displayed.

          11.  From the Primary Network Logon list, click IBM Networks Client.

          12.  From The following network components are installed list,

              click IBM Networks Client.

          13.  Click Properties... button.  The IBM Networks Client Properties

              notebook is displayed.

          14.  Click General tab.

          15.  Type the name of the default logon IBM OS/2 Warp Server domain.

          16.  Click the Advanced tab.

          17.  Select the System Policy settings you want to use.

    IBM Networks Client for Windows 95                          Page 3 of 5

    컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴

          18.  Click OK.  The Network notebook is displayed.

          19.  Click OK.  The System Settings Change window is displayed.

          20.  Click Yes to restart your system.

    -----------------------------------------

    4.0  Installing from IBM OS/2 Warp Server

    -----------------------------------------

        To install from IBM OS/2 Warp server:

          1.  From your IBM OS/2 Warp Server system:

              a.  Create a directory named d:\IBMWIN95

              b.  Copy all of the files from the IBM Networks Client for

                  Windows 95 diskette into the d:\IBMWIN95 directory.

              c.  Create a shared resource, type:

                    NET SHARE IBMWIN95=d:\IBMWIN95

              d.  Set Access Controls, type:

                    NET ACCESS d:\IBMWIN95 /ADD GUESTS:RX USERS:RX

          2.  From your Windows 95 system:

              a.  Follow steps 1 through 6 listed in section 3.0 Installing

                  from Diskette.

              b.  From the Install From Disk window, click Browse...  The

                  Open window is displayed.

              c.  Click Network...  The Map Network Drive window is displayed.

              d.  From the Drive list box, select an available drive letter.

              e.  In the Path list box specify the UNC path to the IBM OS/2

                  Warp Server network resource where the IBM Networks

                  Client for Windows 95 files are located.  For example:

                    \\ServerName\IBMWIN95.

              f.  Click OK.  The Open window is displayed.

              g.  From the Drive list box, select the drive letter that was

                  mapped in step 2d.

              h.  Verify that the IBMNET32.INF file is displayed in the File

                  name box.

              i.  Click OK.  The Install From Disk window is displayed.

              j.  Verify that the mapped network drive letter is displayed

                  in the Copy manufacturer's files from box.

              k.  Click OK.

              l.  Continue with steps 8 through 19 as listed in section 3.0

                  Installing from Diskette.

    IBM Networks Client for Windows 95                          Page 4 of 5

    컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴

    ------------------

    5.0 Up and Running

    ------------------

        5.1  How to Logon

            To logon:

              1.  After completing the installation process, restart your

                  Windows 95 session.  The IBM Networks Client logon

                  window is displayed.

              2.  The default user name and domain name are displayed, but

                  they can be changed.

              3.  Type your network password.

              4.  Click OK to begin validation of your logon.

            Note: If you did not select IBM Networks Client as the

                  Primary Logon Provider, you can logon to the domain

                  through the Graphical User Interface.

        5.2  Starting the Graphical User Interface

            To start the graphical user interface:

              1.  Click Start button, and then click Programs.

              2.  Click IBM Networks Client.

              3.  Click IBM Networks Client.

            Note: You can use Windows 95 Network Neighborhood to view and

                  connect network resources.

        5.3 Displaying the Users Guide

            To view the users guide, do the folowing:

              1.  Click Start button, point to Programs, and then click

                  IBM Networks Client.

              2.  Click Users Guide.

    IBM Networks Client for Windows 95                          Page 5 of 5

    컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴컴

    -------------------------------------

    6.0 How to remove IBM Networks Client

    -------------------------------------

        To remove IBM Networks Client:

          1.  Click Start button, point to Settings, and then click

              Control Panel.

          2.  Double-click Network.  The Network notebook is displayed.

          3.  Click Configuration tab.

          4.  From The following network components are installed list, click

              IBM Networks Client.

          5.  Click Remove... button.

          6.  Click OK.

          7.  Restart your system.

          8.  Click Start button, and then click Run...  The Run window is

              displayed.

          9.  In the Open box type:

              d:\WINDOWS\IBMNDEL.BAT

              All of the files for IBM Networks Client program are deleted.

END OF DOCUMENT.

***********

작성일: 2004/05/09 07:50:30

작성자: 순수백수 

  eCS 1.1 Entry 시디2에 있는 프로그램입니다. 오투판 VNC라고 보시면 되

겠네요. 원격 터미널에서의 화면은 별롭니다. VNC와 다른 점은 원격지 컴

퓨터에서 호스트에 접속하는 방식이 웹을 이용한다는 정도일것 같네요.

주의 하실 점은, 접속후 호스트 컴의 데스크탑이 팝업창으로 뜨기 땜에 웹

브라우저의 팝업 차단 기능은 끄셔야 한다는 것입니다.

ps. 원격 컴에서 모질라 웹브라우저로는 수월치가 않군요.-_- 

작성자: 좀딱 

  - v0.1.4 ( 2004/01/18 )

    .모질라에서 가끔 한/영 변환이 되지 않는 것을 고쳤습니다.

    .특수문자/한자를 입력할 때, 초점이 바뀌어서 나타나던 문제가 사라졌습니다.

    .E 에서 한글 입력 도중에 마우스를 클릭하면 마우스 위치에 한글이 나타나는 

    것을 고쳤습니다.

    .오딘에 대한 지원이 강화되었습니다. 따라서 오딘에 바탕을 둔 InnoTek 의 

    Win32 포팅 프로그램들(Java 2 v1.4.2, OpenOffice 1.1) 에서도 한글 입력이 

    수월해졌습니다.

  - v0.1.3 ( 2003/03/04 )

    .입력 포커스가 바뀌면 그에 대응하여 상태창의 입력 상태도 바뀝니다.

    .입력 중에 마우스를 클릭하면 입력 중인채로 글자가 남아 있는 문제를 

    고쳤습니다.

    .영어판 Warp 에서 코드페이지 949 로 전환했을 때, PM 상에서 '\' 를 

    입력할 수 있습니다.

  - v0.1.2 ( 2002/03/18 )

    .Caps Lock 이 켜져 있을 때, Shift 키가 눌려져 있는 것처럼 동작하는

    것을 고쳤습니다.

    .영어판 워프에서 Num Lock 이 켜져 있을 때, 화살표키를 누르면 키패

    드에 있는 화살표 키를 누른 것처럼 숫자가 나타나는 것을 고쳤습니다.

  - v0.1.1 ( 2002/01/25 )

    .TaskList 에 KIME 을 등록하지 않습니다. 따라서 TaskList 와 관련된

    문제들이 나타나지 않습니다.

    .Ztelnet 으로 입력 초점이 바뀌고나면, 다른 윈도의 한글 입력상태가

    Ztelnet 과 같은 한글 입력 상태로 변하는 것을 고쳤습니다.

    .Ctrl 글쇠 조합이 있을 때, 특수문자 입력창이 나타나는 것을 고쳤습

    니다. 단, Alt/Shift 글쇠를 제외한 나머지 글쇠를 누른 후 마지막에

    Ctrl 글쇠를 누르면 여전히 특수문자 입력창이 나타납니다. 

        예. \ + Ctrl.

  - v0.1.0 ( 2001/11/27 )

    .특수문자/한자 입력을 지원합니다.

    .글꼴연관을 설정할 수 있습니다.

    .한글 입력 상태에서 함수키(Fxx)가 처리되지 못하는 것을 고쳤습니다.

  - v0.0.0 ( 2001/11/08 )

    .PM 에서 한글을 입력할 수 있고, KZP 의 기능을 통합했습니다. 

  등 록 자 : KOMH    

등록일자 : 2003.03.05 15:37  

찬 성 : 0  

반 대 : 0  

키 워 드 :  

KIME013.ZIP (164535 Bytes) 

  K Input Method Editor v0.1.3

0. 소개

  이 프로그램은 한국어용으로 제작되지 않은 OS/2 에서 제한적으로 한글을

입력할 수 있도록 해주는 프로그램입니다. 아울러, Ztelnet 에서 세벌식 390

자판을 쓸 수 있도록 해줍니다.

1. 제작 환경 

  - 한국어판 OS/2 Warp 4 with FixPak #5

  - EMX 0.9d fix #4 + gcc 2.8.1

2. 실행에 필요한 환경

  한글 입력이 가능하기 위해서는 적어도 한글 출력이 가능해야 합니다.

특히 코드페이지 949 가 지원 가능해야 합니다. 다행히 Warp 4 FixPak #5

부터 모든 코드페이지가 지원됩니다.

  CONFIG.SYS 에서 다음 줄을 찾아서 코드페이지 949 를 추가해주셔야 합

니다.

CODEPAGE=850,437    

에서

CODEPAGE=850,949

  이렇게 바꾸시면 됩니다. 물론 주코드페이지를 949 로 맞추어도 되지만,

시스템 폴더를 쓰지 못하든지하는 문제가 생길 수 있습니다. 그리고 코드

페이지와 관련된 기능이 제대로 작동하지 않을 수도 있으니, 보조코드페

이지로 맞추어 놓고 쓰시기 바랍니다.

  그리고 한글 출력을 위해서는 당연히 한글 글꼴이 필요합니다. OS/2 에서

사용가능한 한글 글꼴을 등록하셔야 합니다. 자료실에 보면 Times New Roman

WT-K 글꼴이 있는데, 임시로 이 글꼴을 쓰셔도 됩니다.

  이정도의 조건을 갖추고 있으면 기본적인 환경은 마련한 셈입니다.

  그리고 당연히 프로그램이 실행이 될 때는 코드페이지 949 에서 실행이 되

어야 합니다. 다음과 같은 배치 파일이 도움이 될 수 있을 것입니다.

---------- 949run.cmd

@echo off

chcp 949

start /n /f %1 %2 %3 %4 %5 %6 %7 %8 %9

chcp 850

----------

2-1. 테스트 환경

  이 프로그램은 영어판 WSeB FixPak #3 와 한국어판 Warp 4 FixPak #15 에서

테스트되었습니다. 한국어판 워프에서도 기존 IME 와 충돌없이 쓸 수 있습니

다.

3. 기능

  - 사용하는 키보드 자판에 상관없이 한글 입력이 가능합니다.

  - 두벌식, 세벌식 390, 세벌식 최종 자판을 지원합니다.

    단, Ztelnet 에서는 세벌식 최종 자판은 지원되지 않습니다.

  - 한글 입력 상태를 윈도별로 관리합니다.

  - 한글 입력이 필요없는 프로그램을 지정할 수 있습니다.

  - Ztelnet 을 쓸 때, 한글 입력 상태에서 채팅 라인을 쓸 때 사이띄우개를 

    두번 눌러야 하는 벌레를 패치합니다

  - Over The Spot 형태의 입력을 지원합니다.

  - 글꼴 연관 기능을 제공합니다.

  - 특수문자/한자 입력을 지원합니다.

  - 한국어판 Warp 에서 Shift+PageDown 를 인식하지 않는 IME 의 벌레를 수 

    정합니다.

  - 영어판 Warp 에서 코드페이지를 949 전환했을 때, PM 상에서 '\' 를 입  

    력할 수 있게 합니다.

4. 설치

  압축 파일을 아무 디렉토리에다 풀면 됩니다. ^^ 단, kimehook.dll 은 

LIBPATH 가 지정되어 있는 디렉토리에 넣어두면 됩니다.

5. 사용법

  WPS 나 명령행에서 kime.exe 를 실행시키면 됩니다.

5-1. 사용되는 단축키

  Shift-Space : KIME 의 한/영 상태를 변환합니다.

  Ctrl-Space  : KIME 의 한글 자판을 변환합니다.

                단, Ztelnet 에서는 한/영 상태를 강제 변환합니다.

  Ctrl-F3    : Ztelnet 에서 글자/라인 상태를 강제 변환합니다.

  오른쪽 CTRL : 한글을 입력하고 있을 때는 '한자' 입력 대화상자가 나타나

                고, 한글을 입력하고 있지 않을 때는 '특수문자' 입력 대화

                상자가 나타납니다.

5-2. 사용되는 마우스 기능

  오른쪽 버튼 : 팝업 메뉴가 나타납니다.

  왼  쪽 버튼 : 한/영 또는 글자/라인 상태를 변환시킵니다.

5-3. 메뉴

  Hide    : KIME 상태 상자를 숨깁니다.

  Options : KIME 의 설정을 바꿉니다.

  Exit    : KIME 의 실행을 끝냅니다.

5-4. 명령행 옵션

  "--kbd390"      : KIME 을 세벌식 390 자판으로 시작합니다.

  "--kbd3f"      : KIME 을 세벌식 최종 자판으로 시작합니다.

  "--no-3bul"    : Ztelnet 에서 세벌식 패치를 하지 않습니다.

  "--no-chatline" : Ztelnet 에서 채팅 라인 패치를 하지 않습니다.

  "--hanjafont"  : KIME 에서 한자 입력을 지원하기 위한 한자 글꼴을 

                    설정합니다.

                    한자 글꼴은 HanEdit/2 에서 쓸 수 있는 글꼴이면

                    모두 가능합니다. 

5-5. KIMEOPT 환경 변수

  명령행 옵션을 이용하지 않고, KIMEOPT 라는 환경 변수를 이용할 수도 있

습니다. KIMEOPT=--kbd390 가 설정된 상태에서 KIME 를 실행하는 것은 

KIME --kbd390

으로 실행시킨 것과 같은 효과를 나타냅니다.        

6. 제한 사항 또는 알려진 벌레

  - 어느 윈도의 클래스 이름이 "ClientWindowClass" 일 경우, KIME 이 

    오동작을 할 수 있습니다.

    Ztelnet 의 클라이언트 윈도의 클래스 이름이 "ClientWindowClass" 이고,

    KIME 에서 현재 윈도가 Ztelnet 인지 판단하기 위해 윈도의 클래스 이름

    을 쓰기 때문입니다.

  - Ztelnet 세션과 KIME 사이에 한/영 및 글자/라인 상태 동기화가 

    이루어지지 않을 수 있습니다.

    이것은 Ztelnet 이 접속되어 있지 않은 경우, 키입력을 처리하지 않지만,

    KIME 에서는 접속 여부를 확인할 수 없기 때문에 일어나는 문제입니다.

    따라서, KIME 의 상태를 강제로 변환시켜야 합니다.

  - Ztelnet 에서 세벌식으로 입력할 때 도깨비불 현상이 나타납니다.

    기본적으로 두벌식 입력 방식에 키맵핑 형식으로 세벌식 입력을 

    구현하였기 때문에 나타나는 현상입니다.

  - 넷스케이프에서 한글을 입력하는 도중에 방향키 또는 엔터기를 누르면

    키처리 후에 한글이 입력됩니다.

  - 현재 입력하고 있는 윈도의 코드 페이지 확인을 하지 못하기 때문에, 949

    이외의 코드페이지에서도 한글 입력을 시도합니다. 당연히 입력된 한글이

    제대로 처리되지 않습니다.  

  - 가상키( Alt, Ctrl, Shift, Fxx ) 를 누르면 상태창이 나타납니다.

  - 예외 목록 파일은 현재 작업 디렉토리를 기준으로 처리합니다.

  - 상태창을 이용해서 한/영 또는 글자/라인 상태를 변환시킬 수 없습니다. 

  - PM 상에서만 한글 입력이 지원됩니다.

  - 특수문자/한자를 입력을 할 때 입력 창이 바뀌어도 특수문자/한자 

    입력창이 사라지지 않습니다.

  - 포커스를 새로 얻을 때 전체 선택 기능이 있는 컨트롤에서 특수문자/한자

    입력을 하면 기존의 내용이 모두 사라지고, 새로 입력한 특수문자/한자만

    남습니다.

        예. 넷스케이프 Location 부분

  - 유니코드 글꼴은 글꼴연관에 쓰일 수 없습니다.

    단, FreeType/2 한글 지원판에서는 글꼴연관에 유니코드 글꼴을 쓸 수 

    있습니다.

  - E 에서 한글 입력중일 때 마우스를 클릭하면 클릭한 위치에 글자가 

    나타나고, 글자 옆에 선택 상태처럼 반전된 부분이 나타납니다.

7. 할 것들

  - 상태창을 통한 입력 상태 바꾸기

8. 예외 목록 파일

  예외 목록 파일이란 일부 자체 한글 프로그램에서 한글 변환을 위해서 쓰는

키가 KIME 가 겹치는 경우가 있는데, 이런 혼동을 피하기 위해서 또는 한글입

력이 필요 없는 프로그램에서 아예 한글 입력을 막기 위해서 해당 프로그램 

이름을 적어 놓은 파일입니다.  

  예외 목록 파일은 KIME 이 있는 디렉토리에 EXCEPT.DAT 라는 텍스트 파일이

기 때문에, 일반 텍스트 에디터로 편집하면 됩니다. 편집 방법은 KIME 이 

입력을 시도하지 않아야 하는 프로그램의 이름을 줄단위로 적어주면 됩니다.

  참고로, 기본적으로 EXCEPT.DAT 에는 다음의 내용이 있습니다.

---------- EXCEPT.DAT

HANEDIT.EXE

HWPPM.EXE

FOOL.EXE

ZTELNET.EXE

MED.EXE

VPC.EXE

----------  

9. 판번호 보기

  - v0.1.3 ( 2003/03/04 )

    .입력 포커스가 바뀌면 그에 대응하여 상태창의 입력 상태도 바뀝니다.

    .입력 중에 마우스를 클릭하면 입력 중인채로 글자가 남아 있는 문제를 

    고쳤습니다.

    .영어판 Warp 에서 코드페이지 949 로 전환했을 때, PM 상에서 '\' 를 

    입력할 수 있습니다.

  - v0.1.2 ( 2002/03/18 )

    .Caps Lock 이 켜져 있을 때, Shift 키가 눌려져 있는 것처럼 동작하는

    것을 고쳤습니다.

    .영어판 워프에서 Num Lock 이 켜져 있을 때, 화살표키를 누르면 키패

    드에 있는 화살표 키를 누른 것처럼 숫자가 나타나는 것을 고쳤습니다.

  - v0.1.1 ( 2002/01/25 )

    .TaskList 에 KIME 을 등록하지 않습니다. 따라서 TaskList 와 관련된

    문제들이 나타나지 않습니다.

    .Ztelnet 으로 입력 초점이 바뀌고나면, 다른 윈도의 한글 입력상태가

    Ztelnet 과 같은 한글 입력 상태로 변하는 것을 고쳤습니다.

    .Ctrl 글쇠 조합이 있을 때, 특수문자 입력창이 나타나는 것을 고쳤습

    니다. 단, Alt/Shift 글쇠를 제외한 나머지 글쇠를 누른 후 마지막에

    Ctrl 글쇠를 누르면 여전히 특수문자 입력창이 나타납니다. 

        예. \ + Ctrl.

  - v0.1.0 ( 2001/11/27 )

    .특수문자/한자 입력을 지원합니다.

    .글꼴연관을 설정할 수 있습니다.

    .한글 입력 상태에서 함수키(Fxx)가 처리되지 못하는 것을 고쳤습니다.

  - v0.0.0 ( 2001/11/08 )

    .PM 에서 한글을 입력할 수 있고, KZP 의 기능을 통합했습니다.

10. 소스 컴파일 하기

  소스를 컴파일 하기 위해서는 EMX 0.9d fix #4 + gcc 2.8.1 이 필요하며, 

사용된 Make 프로그램은 GNU Make 3.76.1 이상입니다.

11. 하고 싶은 말

  이 프로그램은 문유성(kwisatz)님께서 만드신 한글 입력 루틴 HIA 를 이용

해서 만들어졌습니다. 매번 느끼는 것이지만, 유성님의 탁월한 설계 능력에 

감탄합니다. 다시 한 번 감사드립니다. 그리고 저에게 WSeB 환경을 제공해 주

셔서, KIME 을 시험해 볼 수 있는 환경을 제공해 주신 최민순(josuwa77)님께

도 깊은 감사의 말씀을 드립니다.

12. 하고 싶은 말이 있을 때...

  e-mail        : komh@chollian.net

  하이텔, 천리안 : komh

                                                  만든이 : 고명훈

인터날 리비전 : 14.097h

출처 :  ftp://testcase.software.ibm.com/ps/fromibm/os2

the OS/2 SES facility, you need a new sesdd32.sys dated April 2001 or later

to use this kernel. Also, a matching dump formatter (df_ret.exe), provided

separately, is required, if you want to analyze a dump from this kernel.

Finally, note that os2ldr is only required for the >64mb memory detection

mentioned below.

***New as of 2002/04/16: I'm including an updated os2dump module, because the

   one that comes with mcp2/acp2 can possibly OVERWRITE your partition(s)!

   If you plan to use standalone dump and you have mcp2 or acp2, use this!

   Caveat: This os2dump will *not* work on non-LVMized systems (i.e. Warp 4+

   FP13-16). On the other hand, the bad os2dump wasn't distributed for FP13-16.

New function:

    1) Support for >64m memory using int15 func e820. Note, this support is

       enabled ONLY for Intel Pentium Pro or later, or non-Intel Pentium-class

       machines. This is done in an effort to prevent problems on older PCs.

    2) The presence of \os2\boot\os2norev.$$$ suppresses the display of 

       internal revision during boot.

    3) Adding "EARLYMEMINIT=TRUE" to config.sys will allow device drivers, etc.,

       access to the memory above 16mb early in boot. Previously, this was only

       available after DD and IFS init was completed. This has various

       implications when enabled:

  a) AHA154X.ADD may do bad things to your system. Don't even ask.

  b) There may be some settings of HPFS386 cache that are incompatible.

    4) Trap messages will now give the module name instead of the device name.

       For example, "Exception in module: CLOCK01" instead of "Exception in

       Device Driver: CLOCK$". 

***change from 0220/0223: Traps in 32 bit code now show the modname, too

    5) Systems with a 486SX chip running FP15 will function correctly. Note,

       they will NOT work with ACP or (necessarily) on MCP.

    6) SESDD32 will now work correctly (I hope) for both uni and smp

    7) All process dumps should now be considerably faster. paddr(all) dumps

       should take about 40% less time and others around 80-90% less time.

    8) EXE names will now also be displayed in alt-f2/alt-f3 processing (note

       that CMD file names, e.g. MPTSTART, will NOT be displayed, only cmd.exe)

    9) alt-f4 is like alt-f2 plus it pauses after displaying each names (does

       not pause after BASEDEV= or PSD=) and, by pressing F1, allows you to

       skip loading/running the file displayed.  Alt-f5 is like alt-f4+alt-f1.

    10) The dump formatter (in dfxxx.zip) should be faster and also work on

        dumps >= 2gb, which was not formerly possible.

  04/16/2002:

    11) Added "1" as an OK setting for flags in config.sys. Now, things like

        EARLYMEMINIT=TRUE can also be YES,ON, or 1 (or FALSE,NO,OFF, or 0).

I wish I could make stuff like this NLV-independent but that is 

ABSOLUTELY IMPOSSIBLE without HUGE changes.

 2002/01/02:

    Support for 16bit filesystems with >2gb file support (e.g.the forthcoming

    >2gb dump support ifs)

 9/13/01

    11) If you also install the enclosed clock01.sys to \os2\boot, you can 

        add "CLOCKSCALE=x" to config.sys. The choices for x are 1 (the default,

which is the old behavior), 2, and 4. The number represents a scale 

factor for the frequency of clock ticks. The theory is that CLOCKSCALE=4

will speed up some apps (maybe Netscape and/or Java).

Bug Fixes:

  Fixes from earlier "dated" kernels:

    Fixed a trap 000d in w_OpenCreate (140:4dfd or so in uni/w4 krnls, don't

              have regs for SMP). This was a post mcp/acp regression

    Fixed strange loader i/o problems specific to the 0111 kernel.

    Fixed a bootup sys31xx problem specific to the 0123 kernel.

    Fixed a problem that was (I think) specific to  Janauary uni & W4 kernels

          which resulted in odd ring 3 (maybe ring 0?) traps and hangs

    Fixed some trap and other peculiar problems specific to the 0206 kernel.

          Those were the result of a build problem.

    Fixed some trap and other problems specific to the 0220/0223 kernels

    Fixed a regression from JR15038 that caused a trap in CWBBS.IFS and

    PCSFLR0.IFS (AS/400 shared folders access).

    Fixed a regression in the 0330 kernel that messed up getting procdumps

    Fixed a regression in the 0416 kernel that caused some hangs and also

        affected APM

    Fixed a regression in the 0502 kernel that prevented booting altogether or

            caused other very bad problems.

    Fixed a regression in the 0502/0503 kernels that caused some odd trap d/e's

            in the scheduling code

    Don't use the 0601 kernels.

    Don't use the 0615 kernels, either. There seems to have been some kind of

            build problem, there.

    Removed the int 3's I'd accidentally left in some EA code in 0727 build

    The 0730-0801 kernels all had problems due to various attempts to fix

            PJ28078. Sorry about that!

    10/16: the 1015 os2ldr was broken on some CPUs

    11/29: Trap 000e (DS will be 0) in 11/28 kernels (only)

           Some peculiarities deleting files

    01/18: Some hang/trap problems relating to the 01/17 kernel

    01/25: Fixed regression from DosDelete change where some files weren't

       getting deleted if DELDIR was set.

    02/21: Strange problems on SMP kernel, particularly with HPFS386 (related to

       0201 kernel, only)

    02/22: 2/21 kernel fix for PJ28434 fix was bad. Unknown possible side-

           effects during bootup.

    03/15: Chgfileptr (seek) ops may fail on FAT, particularly on a vdisk

    04/01: df_ret/df_deb trap on some memory analysis operations. note, you can

           use this df_ret.exe (& df_deb.exe) on dumps of kernels back to about

   Sep 2001 or so. Just always use the symbols that match the dump.

    04/16: Added test fix for trap 000e in _ldrSegFixAddIterRec

           - *May* have fixed some strange program load behavior with the

     implementation of the latest JR15038 fix.

  Fixes to problems that exist in released code:

    JR15494: Delayed response of Ctrl-C or Ctrl-S (pause) in os/2 windows

    JR15405: (probably) DosQueryAsyncTimer fails on timecritical thread after

         mscount rolls over after 49.7 days of uptime

    JR15030: TRAPDUMP QUERY doesn't show PD (needs a new trapdump.exe, too)

    JR13335: Some VDMs won't run, complaining of too-long command lines

    JR15194: XCOPY fails with SYS1186 or SYS1192 on (at least) MO drives 

    JR15038: Trap 8 in h_DeviceHelpDM running an app from a tcpbeui drive

    JR15229: Files left open after ERROR_TOO_MANY_OPEN_FILES returned to app

    PJ27700: SMP systems may get a trap 000e in _tkStopScan, particularly 

             during a procdump operation.

    PJ27676: Causes hangs on some systems running UNI or W4 kernels

    PJ27554: Hang or (on allstrict) trap 3 on SMP (only) during beeps

    PJ27678: SMP kernels could not be RIPL-booted. As part of this, there is

     a new config.sys parameter I13PAGES= available. Adding I13PAGES=1

     to config.sys (note, no SET) reduces the memory usage of the mini

     VDM processes used by ibm1s506.add and the GRADD drivers. This

     becomes important in RIPL-boot situations. It may be helpful for

     UNI RIPL boot, as well.

    PJ27677: SMP systems hang with multiple threads of same process

     entering a critical section (affects ADSM client)

    PJ27757: Trap 000e running Theseus/4 (note, theseus4.exe may still end up

             with a sys3175, but at least the system doesn't crash).

    PJ27736: Trap in _PLD_WriteReturnCode

    PJ27777: Set BEGINLIBPATH= (in config.sys ONLY) may be ignored. Also, 

             it would lead to a bogus copy of BEGINLIBPATH= in the environment.

    ;PJ27821: Trap 000e in SchedNextRet2 on some APM-enabled machines

    ;         *removed as of 0508 -- original fix caused regressions. new fix

    ;         on hold, for now

    PJ27906: Trap 000e in pvwRegKrnlInstance running SPM/2 (may also cause

             unknown odd behavior in other performance monitors).

    PJ27919: Complete system hang running some apps (same as JR12136)

    PJ27544: Panic in SwitchNPX

    PJ27972: trap 000e in _SecExecPost (the actual fix is in ses0604 or later)

    PJ27952: DosFindFirst returns incorrect returncode for various bufsizes

             when querying extended attributes

    PJ28005: ProcDump command returns error code 6 (and the procdump is invalid)

    PJ28013: read of >2k from named pipe on a timecritical thread can cause

         loss of pipe data or session hang

    PJ28027: trap in KMQueryClientRegister reallocating a trace buffer

    PJ28068: trap 000e in clock$ where sslim is ffffffff (during JFS activity)

    PJ28076: "trapdump [PD|NOPD]" disables the procdump.exe command. Also, it

         causes a trap 000e if you initiate a dump with ctrl-alt-f10-f10

     (or numlock-numlock).

    PJ28077: Some VPM apps will refuse to initialize

    PJ28078: Various traps with floppy disk hard errors when UDF is loaded

    JR13789: DRIVES does not reflect change in directory entries

    PJ28102: Performance tracing slow on Pentium 4 CPUs (this is the only 

             APAR listed here fixed in os2ldr)

  10/16:

    PJ28199  SES QUERYFILEINFO buffer pointer always -1

    PJ28207  SES callgate hooks don't work

  10/26

    PJ28215  Added HMA flag to DevHelp_GlobalToProcess. The new flag 20h (in

             EAX for the call to the DevHelp) allows a device driver to map

     global memory to a process' high private arena.

    PJ264136 Trap 000e in the kernel when theseus gets system working set. Also

             fixed a theseus problem on systems with HPFS386.

  11/28 (MCP2/ACP2 kernel plus the following)

    IC32279  Trap 000e in TKDeclareInversion after killing a process while it's

             doing DosDelete processing.

  01/17 (happy new year) revision 14.086c

    PJ28355  SYS3175 (or other problems?) when booting a Pentium 4 machine when

             initializing testcfg.sys or apm.sys

    **Note: For those of you with ACP1 or earlier, the os2ldr contained here is

            also mandatory to fix the P4 problem.

  01/18 revision 14.086d

    PJ28362  traps in os2krnl running theseus/4

  01/25 revision 14.086e

    PJ28379  disallow alt-f4 and alt-f5 in the presence of altf1sec.cmd

  02/01 revision 14.086g

    PJ28392  trap in f_w16_DoPageIO ***** see 03/06 *******

  02/21 revision 14.088a

    PJ28434  swapper.dat can be trashed from a VDM session

    JR15405: Second attempt to fix this

  02/26  revision 14.088c

    PJ28450  fix random problems on SMP Pentium PRO machines that started as

             of 2001/08/21

    PJ28355 part 2 -- some P4 machines STILL had boot problems

  0306   revision 14.088d

    PJ28392   Again! Somehow, this fixed got dropped from 0226

  0311   revision 14.088e

    PJ28468   Undelete not working since CP2

  0315   revision 14.088f

    PJ28472   Trap 000e in THEArena_Get running theseus/4 (may also be in 0311)

  0401   revision 14.088g

    JR15405 part 2: more likely fix for 49.7 day hang (still in test)

    JR15038 part 2: more trap 8s running an app from a tcpbeui drive

  0408   revision 14.088h

    JR15038 part 3: more trap 8s running an app from a tcpbeui drive

  0416   revision 14.089b

    PJ28541 sys0008 trying to start up lots of huge DLLs

    JR15038 new version: rather than slowing executable loads for everyone, I've

            gone back to using the faster stack-based buffer allocations most

    of the time. If you have loaded an ISS (security) or if you

    explicitly add "LDRSTACKOPT=TRUE" (No "SET ,,," please) to

    config.sys, we will try to minimize ring 0 stack usage by

    allocating buffers off the heap.  This is slower, but prevents

    ring 0 trap 0008s in some cases. If you use TCPBEUI, particularly

    with the IBMTRP driver, you should probably add this parameter

    to config.sys.

  0426   revision 14.089c

    Note: JR15038  from 0416 temporarily removed, as I think it was causing

          unstable behavior (including trap 000E and trap 0008). I left the

  LDRSTACKOPT option in, but for now it does nothing.

    PJ28547 trap 000e in _ldrSegFixAddIterRec running Oracle 6 and other apps

    PJ28566 trap 000e at _PLD_WriteReturnCode + 169 or so during procdump.

    PJ28567 trap 0003 in _PLDClose during procdump.

  20020515   revision 14.089d  (note the new naming convention)

    PJ28600 non-zero return from SecOpenPost ISS callout after file create

        operation leaves zero length files around

    PJ28602 internal thread names wrong in df and debug krnl on 1 cpu SMP

    PJ28603 module name missing or incorrect in r0 trap screen

  20020516   revision 14.089e

  A build problem caused the 20020515 PJ28600 fix to cause a trap. You

will only see this if you have a security device driver installed.

  20020802   revision 14.091a

  - not clear whether PJ28392 was included in recent kernels before this

- IRQ sharing problem in os2ldr caused hang on reboot for some USB

  controllers sharing IRQs

- Yet another attempt (the last, we hope) to fix the 49.7 day hang

  20020822   revision 14.091b

  - backed out the os2ldr junk from 20020802

JR15405 (final) the 49.7 day change might not have been in 20020802.

  20021010   revision 14.091c

        - put back in the IRQ fix in os2ldr. seems to be ok

- took clock01 out of the package.

  20021028   revision 14.093a

  PJ28871 trap with TRACEBUF= .../D=ALL in config.sys

- fix an obscure trap in KernelFaultEntry on debug kernel only

- fix a trap if Kern32ReadFileAt called and an ISS is installed

  20021205 14.093c

- fix a trap if Kern32ReadFileAt called and an ISS is installed

  (the fix was not correct in 1028)

- trap in w_dup trying to do DosDupHandle to target handle outside range

  of current max file handle number

PJ28941 trap 000e routine VMAliasToAlias doing procdumps on some systems

  20031226 14.093d

  - back out all experimental stuff including the w_dup trap and trap 8

  test fixes

  20030213 14.093e

- close some files that had been left over by some inconsiderate DD

  init code. the extra handles choked some apps, like MS C6

- Prevent trap under some condtion I don't understand while using a

  debugger (like codeview or ipmd or whatever) in DbgSeverConnection

JR17862 fix a problem with a thread hanging (but not the whole system) 

  after the 49.7 day rollover doing a DosSleep (*Not* fixed in 1226)

          Also, DosRequestMutexSem with 0 < timeout < ffffffff may hang if the

  thread wakes up after rollover and still doesn't own the sem.

PJ29032  trap in thesmte_get or themte_get running theseus4 or using

         theseus4 APIs

  20030303 14.093f

       - backed out the file closing change in 0213, as it seems to be causing

         problems on some systems. the underlying problem will be investigated

later in individual drivers.

       - change to fix from JR17862 in 0213 kernel due to possible regression

         that would result in a semaphore not timing out properly

       PJ29066 possible fix for a rare trap in _qsSem running pstat

       JR17842 various dump formatter rc:5 problems (fixed in df_ret/df_deb)

  20030306 14.093g

       - w_dup fix back in

       - another shot at fixing the trap 8 problems while running an ISS or

      using tcpbeui. If you're using tcpbeui (*especially *under SMP), you

      should not use this. If you're using an ISS, we ignore this. For others,

      you may get some speed improvement in loading apps by adding

      LDRSTACKOPT=FALSE to config.sys (the default being true).

  20030415 14.093h

       - LDRSTACKOPT back out. It caused too much confusion and I'm not 100%

         sure about its stability. Sigh.

       - Some enhancements to the ".pb" command in the dump formatter

  20030501 14.093i

       - I kept forgetting to take out the code that closes file handles > 2

         before starting PROTSHELL. The good news is that doing the close

reduces the number of leaked file handles by a few in some systems.

The bad news is that it also makes PMDD$ trap on boot on other

systems. So, if that happened to you with 14.093h, get this kernel.

As of now, add CLOSEFILES=TRUE if, like me, you need the changed

behavior (in my case, Microsoft C 6 kept aborting).

       PJ29180  Trap 000E while taking a process dump

  20030620 14.093i1

  - don't use this one. I put it out by accident due to packaging bug.

  20030621 14.094

  - nothing new, but this is the first "official" (i.e. one where we'll

  actually have full source code control) in a while. It leaves out

  the CLOSEFILES keyword, though.

  20030822 14,096

  PJ29349  Trap in ThePD_Get (another theseus/4 trap)

PJ28318  problems accessing non-US filenames (regression due to PJ29234)

         (related fix to cmd.exe to back out bogus changes will be 

  available in next fixpack)

  20031031  14.096b

- CLOSEFILES= back in

- fix 2 traps in w_DupHandle

   Seems to affect Netfinity remote access.

PJ29849  trap 000c in w_NewSize during procdump processing

  20031103  14.096c

        - incorrect parameters being reported in some devhelp system traces

  20031110  14.096c

        - 1103 system trace fix not built into some kernels

- fix named pipe regression in 1031 and 1103 kernels

- os2ldr is now the version that "works" with shared interrupts. 

  HOWEVER, if your system has shared interrupts that didn't used to

  get noticed, then you may suddenly get much slower. This happens on

  my personal Intellistation which shares an IRQ for the USB and audio.

  20031113  14.096e

- trace fix didn't get into previous w4 and uni kernels

- EXPERIMENTAL: Removed a bunch of redundant locking and unlocking of

  some pages in filesystem code that's already pinned, anyway. This

  ought to speed up a number of filesys-related APIs. Also, re-enabled

  the un-pinning of numerous pages of memory from init code. This

  memory is now available for other uses, which should reduce paging on

  many systems. This had been turned off since WSeB!

PJ29472: process hangs (and other things?) if named pipe handle is

   closed while another thread is blocked in DosConnectNmPipe.

  20031117  14.097b  (post xr_c004)

  - Some of the pages in the 20031113 package may possibly need to stay

  resident to avoid any danger of a trap 000e if they're needed during

  page fault processing or interrupt time. So, this time, I left those

  pages resident,too.

  20040116 14.097d

  - the code lock stuff seems to be working well, so it's no longer 

   "experimental."

- possible DosConnectNmPipe regression test fix

- Support for loading pure 32 bit code and data objects in DLLs above

  the 512mb line. More details on that to come separately.

  20040119 14.097d  (still)  *DON'T USE THIS OS2LDR*

  No kernel changes from 0116, but OS2LDR changes:

  - the interrupt sharing fix had, at some point, got taken back out.

  users with seeming IRQ issues should try this

- possible problem booting up P4 Xeon 4 gHz (hang at boot blob)

  20040123 14.097e

  - fix bad cFH count in DosQuerySysState/QProcStatus with level=0

- attempt to fix intermittent trap 000e resulting from highmem loadin

  in 116 kernel

    - bad fix for boot blob hang in 20040119 caused "unable to operate

  hard disk" failures with os2ldr from 20040119 packages

  20040126 14.097f

  - fix bad file handle array in DosQuerySysState/QProcStatus with level=0

  - more changes related to boot hangs and other weirdness from 1/23

  20040126 14.097g

  - more on file handle array in DosQuerySysState/QProcStatus with level=0

  - another attempt to update os2ldr to work with everything from 486SX to P4 Xeon >3gHz....

PJ29595: trap 000e at SftFileSize + 66 or so

  20040206 14.097h

  PJ29602 trap 000d at h_DOS_ChDir + 28 or so after media change

- backed out bogus os2ldr changes (again)

  20040323 14.098a

  PJ29697 add debug support for disappearing browser windows

- fix DevIOCtl passthru support

- fixed the "k" command in the dump formatter (at least it's better)

  20040330 14.098b

  PJ29712 (os2ldr) final fix for boot blob hang on fast (around 3+ gHz) CPUs

- backed out some kernel stuff that accidentally got built

Following are rough instructions for applying the os2krnl fix:

In the directory from which you boot (C:\ or d:\ or whatever), type these

commands. Note that os2krnl should have come with this file.

Also, there are some REXX scripts written by third parties that might make

installation/archiving of your kernels easier. Go to http://hobbes.nmsu.edu

and search for "moritz", for example. This is unsupported by IBM!

1. attrib -r -s -h os2krnl

2. attrib -r -s -h os2ldr

3. copy os2krnl os2krnl.sav

4. copy os2ldr os2ldr.sav

5. copy \os2\system\trace\*.tdf  \save       (or wherever)

5. unzip -o xxx.zip

If you want, you can turn the os2krnl and os2ldr attributes back on, but

it's not necessary to do so.

Trace files (*.tdf)  have been included.

An os2krnl.sym file has been provided with this fix. Just save it somewhere

(root directory or anywhere). It is only needed in case OS/2 

service requests a standalone dump of your system in the future.

the OS/2 SES facility, you need a new sesdd32.sys dated April 2001 or later

to use this kernel. Also, a matching dump formatter (df_ret.exe), provided

separately, is required, if you want to analyze a dump from this kernel.

Finally, note that os2ldr is only required for the >64mb memory detection

mentioned below.

***New as of 2002/04/16: I'm including an updated os2dump module, because the

   one that comes with mcp2/acp2 can possibly OVERWRITE your partition(s)!

   If you plan to use standalone dump and you have mcp2 or acp2, use this!

   Caveat: This os2dump will *not* work on non-LVMized systems (i.e. Warp 4+

   FP13-16). On the other hand, the bad os2dump wasn't distributed for FP13-16.

New function:

    1) Support for >64m memory using int15 func e820. Note, this support is

       enabled ONLY for Intel Pentium Pro or later, or non-Intel Pentium-class

       machines. This is done in an effort to prevent problems on older PCs.

    2) The presence of \os2\boot\os2norev.$$$ suppresses the display of 

       internal revision during boot.

    3) Adding "EARLYMEMINIT=TRUE" to config.sys will allow device drivers, etc.,

       access to the memory above 16mb early in boot. Previously, this was only

       available after DD and IFS init was completed. This has various

       implications when enabled:

  a) AHA154X.ADD may do bad things to your system. Don't even ask.

  b) There may be some settings of HPFS386 cache that are incompatible.

    4) Trap messages will now give the module name instead of the device name.

       For example, "Exception in module: CLOCK01" instead of "Exception in

       Device Driver: CLOCK$". 

***change from 0220/0223: Traps in 32 bit code now show the modname, too

    5) Systems with a 486SX chip running FP15 will function correctly. Note,

       they will NOT work with ACP or (necessarily) on MCP.

    6) SESDD32 will now work correctly (I hope) for both uni and smp

    7) All process dumps should now be considerably faster. paddr(all) dumps

       should take about 40% less time and others around 80-90% less time.

    8) EXE names will now also be displayed in alt-f2/alt-f3 processing (note

       that CMD file names, e.g. MPTSTART, will NOT be displayed, only cmd.exe)

    9) alt-f4 is like alt-f2 plus it pauses after displaying each names (does

       not pause after BASEDEV= or PSD=) and, by pressing F1, allows you to

       skip loading/running the file displayed.  Alt-f5 is like alt-f4+alt-f1.

    10) The dump formatter (in dfxxx.zip) should be faster and also work on

        dumps >= 2gb, which was not formerly possible.

  04/16/2002:

    11) Added "1" as an OK setting for flags in config.sys. Now, things like

        EARLYMEMINIT=TRUE can also be YES,ON, or 1 (or FALSE,NO,OFF, or 0).

I wish I could make stuff like this NLV-independent but that is 

ABSOLUTELY IMPOSSIBLE without HUGE changes.

 2002/01/02:

    Support for 16bit filesystems with >2gb file support (e.g.the forthcoming

    >2gb dump support ifs)

 9/13/01

    11) If you also install the enclosed clock01.sys to \os2\boot, you can 

        add "CLOCKSCALE=x" to config.sys. The choices for x are 1 (the default,

which is the old behavior), 2, and 4. The number represents a scale 

factor for the frequency of clock ticks. The theory is that CLOCKSCALE=4

will speed up some apps (maybe Netscape and/or Java).

Bug Fixes:

  Fixes from earlier "dated" kernels:

    Fixed a trap 000d in w_OpenCreate (140:4dfd or so in uni/w4 krnls, don't

              have regs for SMP). This was a post mcp/acp regression

    Fixed strange loader i/o problems specific to the 0111 kernel.

    Fixed a bootup sys31xx problem specific to the 0123 kernel.

    Fixed a problem that was (I think) specific to  Janauary uni & W4 kernels

          which resulted in odd ring 3 (maybe ring 0?) traps and hangs

    Fixed some trap and other peculiar problems specific to the 0206 kernel.

          Those were the result of a build problem.

    Fixed some trap and other problems specific to the 0220/0223 kernels

    Fixed a regression from JR15038 that caused a trap in CWBBS.IFS and

    PCSFLR0.IFS (AS/400 shared folders access).

    Fixed a regression in the 0330 kernel that messed up getting procdumps

    Fixed a regression in the 0416 kernel that caused some hangs and also

        affected APM

    Fixed a regression in the 0502 kernel that prevented booting altogether or

            caused other very bad problems.

    Fixed a regression in the 0502/0503 kernels that caused some odd trap d/e's

            in the scheduling code

    Don't use the 0601 kernels.

    Don't use the 0615 kernels, either. There seems to have been some kind of

            build problem, there.

    Removed the int 3's I'd accidentally left in some EA code in 0727 build

    The 0730-0801 kernels all had problems due to various attempts to fix

            PJ28078. Sorry about that!

    10/16: the 1015 os2ldr was broken on some CPUs

    11/29: Trap 000e (DS will be 0) in 11/28 kernels (only)

           Some peculiarities deleting files

    01/18: Some hang/trap problems relating to the 01/17 kernel

    01/25: Fixed regression from DosDelete change where some files weren't

       getting deleted if DELDIR was set.

    02/21: Strange problems on SMP kernel, particularly with HPFS386 (related to

       0201 kernel, only)

    02/22: 2/21 kernel fix for PJ28434 fix was bad. Unknown possible side-

           effects during bootup.

    03/15: Chgfileptr (seek) ops may fail on FAT, particularly on a vdisk

    04/01: df_ret/df_deb trap on some memory analysis operations. note, you can

           use this df_ret.exe (& df_deb.exe) on dumps of kernels back to about

   Sep 2001 or so. Just always use the symbols that match the dump.

    04/16: Added test fix for trap 000e in _ldrSegFixAddIterRec

           - *May* have fixed some strange program load behavior with the

     implementation of the latest JR15038 fix.

  Fixes to problems that exist in released code:

    JR15494: Delayed response of Ctrl-C or Ctrl-S (pause) in os/2 windows

    JR15405: (probably) DosQueryAsyncTimer fails on timecritical thread after

         mscount rolls over after 49.7 days of uptime

    JR15030: TRAPDUMP QUERY doesn't show PD (needs a new trapdump.exe, too)

    JR13335: Some VDMs won't run, complaining of too-long command lines

    JR15194: XCOPY fails with SYS1186 or SYS1192 on (at least) MO drives 

    JR15038: Trap 8 in h_DeviceHelpDM running an app from a tcpbeui drive

    JR15229: Files left open after ERROR_TOO_MANY_OPEN_FILES returned to app

    PJ27700: SMP systems may get a trap 000e in _tkStopScan, particularly 

             during a procdump operation.

    PJ27676: Causes hangs on some systems running UNI or W4 kernels

    PJ27554: Hang or (on allstrict) trap 3 on SMP (only) during beeps

    PJ27678: SMP kernels could not be RIPL-booted. As part of this, there is

     a new config.sys parameter I13PAGES= available. Adding I13PAGES=1

     to config.sys (note, no SET) reduces the memory usage of the mini

     VDM processes used by ibm1s506.add and the GRADD drivers. This

     becomes important in RIPL-boot situations. It may be helpful for

     UNI RIPL boot, as well.

    PJ27677: SMP systems hang with multiple threads of same process

     entering a critical section (affects ADSM client)

    PJ27757: Trap 000e running Theseus/4 (note, theseus4.exe may still end up

             with a sys3175, but at least the system doesn't crash).

    PJ27736: Trap in _PLD_WriteReturnCode

    PJ27777: Set BEGINLIBPATH= (in config.sys ONLY) may be ignored. Also, 

             it would lead to a bogus copy of BEGINLIBPATH= in the environment.

    ;PJ27821: Trap 000e in SchedNextRet2 on some APM-enabled machines

    ;         *removed as of 0508 -- original fix caused regressions. new fix

    ;         on hold, for now

    PJ27906: Trap 000e in pvwRegKrnlInstance running SPM/2 (may also cause

             unknown odd behavior in other performance monitors).

    PJ27919: Complete system hang running some apps (same as JR12136)

    PJ27544: Panic in SwitchNPX

    PJ27972: trap 000e in _SecExecPost (the actual fix is in ses0604 or later)

    PJ27952: DosFindFirst returns incorrect returncode for various bufsizes

             when querying extended attributes

    PJ28005: ProcDump command returns error code 6 (and the procdump is invalid)

    PJ28013: read of >2k from named pipe on a timecritical thread can cause

         loss of pipe data or session hang

    PJ28027: trap in KMQueryClientRegister reallocating a trace buffer

    PJ28068: trap 000e in clock$ where sslim is ffffffff (during JFS activity)

    PJ28076: "trapdump [PD|NOPD]" disables the procdump.exe command. Also, it

         causes a trap 000e if you initiate a dump with ctrl-alt-f10-f10

     (or numlock-numlock).

    PJ28077: Some VPM apps will refuse to initialize

    PJ28078: Various traps with floppy disk hard errors when UDF is loaded

    JR13789: DRIVES does not reflect change in directory entries

    PJ28102: Performance tracing slow on Pentium 4 CPUs (this is the only 

             APAR listed here fixed in os2ldr)

  10/16:

    PJ28199  SES QUERYFILEINFO buffer pointer always -1

    PJ28207  SES callgate hooks don't work

  10/26

    PJ28215  Added HMA flag to DevHelp_GlobalToProcess. The new flag 20h (in

             EAX for the call to the DevHelp) allows a device driver to map

     global memory to a process' high private arena.

    PJ264136 Trap 000e in the kernel when theseus gets system working set. Also

             fixed a theseus problem on systems with HPFS386.

  11/28 (MCP2/ACP2 kernel plus the following)

    IC32279  Trap 000e in TKDeclareInversion after killing a process while it's

             doing DosDelete processing.

  01/17 (happy new year) revision 14.086c

    PJ28355  SYS3175 (or other problems?) when booting a Pentium 4 machine when

             initializing testcfg.sys or apm.sys

    **Note: For those of you with ACP1 or earlier, the os2ldr contained here is

            also mandatory to fix the P4 problem.

  01/18 revision 14.086d

    PJ28362  traps in os2krnl running theseus/4

  01/25 revision 14.086e

    PJ28379  disallow alt-f4 and alt-f5 in the presence of altf1sec.cmd

  02/01 revision 14.086g

    PJ28392  trap in f_w16_DoPageIO ***** see 03/06 *******

  02/21 revision 14.088a

    PJ28434  swapper.dat can be trashed from a VDM session

    JR15405: Second attempt to fix this

  02/26  revision 14.088c

    PJ28450  fix random problems on SMP Pentium PRO machines that started as

             of 2001/08/21

    PJ28355 part 2 -- some P4 machines STILL had boot problems

  0306   revision 14.088d

    PJ28392   Again! Somehow, this fixed got dropped from 0226

  0311   revision 14.088e

    PJ28468   Undelete not working since CP2

  0315   revision 14.088f

    PJ28472   Trap 000e in THEArena_Get running theseus/4 (may also be in 0311)

  0401   revision 14.088g

    JR15405 part 2: more likely fix for 49.7 day hang (still in test)

    JR15038 part 2: more trap 8s running an app from a tcpbeui drive

  0408   revision 14.088h

    JR15038 part 3: more trap 8s running an app from a tcpbeui drive

  0416   revision 14.089b

    PJ28541 sys0008 trying to start up lots of huge DLLs

    JR15038 new version: rather than slowing executable loads for everyone, I've

            gone back to using the faster stack-based buffer allocations most

    of the time. If you have loaded an ISS (security) or if you

    explicitly add "LDRSTACKOPT=TRUE" (No "SET ,,," please) to

    config.sys, we will try to minimize ring 0 stack usage by

    allocating buffers off the heap.  This is slower, but prevents

    ring 0 trap 0008s in some cases. If you use TCPBEUI, particularly

    with the IBMTRP driver, you should probably add this parameter

    to config.sys.

  0426   revision 14.089c

    Note: JR15038  from 0416 temporarily removed, as I think it was causing

          unstable behavior (including trap 000E and trap 0008). I left the

  LDRSTACKOPT option in, but for now it does nothing.

    PJ28547 trap 000e in _ldrSegFixAddIterRec running Oracle 6 and other apps

    PJ28566 trap 000e at _PLD_WriteReturnCode + 169 or so during procdump.

    PJ28567 trap 0003 in _PLDClose during procdump.

  20020515   revision 14.089d  (note the new naming convention)

    PJ28600 non-zero return from SecOpenPost ISS callout after file create

        operation leaves zero length files around

    PJ28602 internal thread names wrong in df and debug krnl on 1 cpu SMP

    PJ28603 module name missing or incorrect in r0 trap screen

  20020516   revision 14.089e

  A build problem caused the 20020515 PJ28600 fix to cause a trap. You

will only see this if you have a security device driver installed.

  20020802   revision 14.091a

  - not clear whether PJ28392 was included in recent kernels before this

- IRQ sharing problem in os2ldr caused hang on reboot for some USB

  controllers sharing IRQs

- Yet another attempt (the last, we hope) to fix the 49.7 day hang

  20020822   revision 14.091b

  - backed out the os2ldr junk from 20020802

JR15405 (final) the 49.7 day change might not have been in 20020802.

  20021010   revision 14.091c

        - put back in the IRQ fix in os2ldr. seems to be ok

- took clock01 out of the package.

  20021028   revision 14.093a

  PJ28871 trap with TRACEBUF= .../D=ALL in config.sys

- fix an obscure trap in KernelFaultEntry on debug kernel only

- fix a trap if Kern32ReadFileAt called and an ISS is installed

  20021205 14.093c

- fix a trap if Kern32ReadFileAt called and an ISS is installed

  (the fix was not correct in 1028)

- trap in w_dup trying to do DosDupHandle to target handle outside range

  of current max file handle number

PJ28941 trap 000e routine VMAliasToAlias doing procdumps on some systems

  20031226 14.093d

  - back out all experimental stuff including the w_dup trap and trap 8

  test fixes

  20030213 14.093e

- close some files that had been left over by some inconsiderate DD

  init code. the extra handles choked some apps, like MS C6

- Prevent trap under some condtion I don't understand while using a

  debugger (like codeview or ipmd or whatever) in DbgSeverConnection

JR17862 fix a problem with a thread hanging (but not the whole system) 

  after the 49.7 day rollover doing a DosSleep (*Not* fixed in 1226)

          Also, DosRequestMutexSem with 0 < timeout < ffffffff may hang if the

  thread wakes up after rollover and still doesn't own the sem.

PJ29032  trap in thesmte_get or themte_get running theseus4 or using

         theseus4 APIs

  20030303 14.093f

       - backed out the file closing change in 0213, as it seems to be causing

         problems on some systems. the underlying problem will be investigated

later in individual drivers.

       - change to fix from JR17862 in 0213 kernel due to possible regression

         that would result in a semaphore not timing out properly

       PJ29066 possible fix for a rare trap in _qsSem running pstat

       JR17842 various dump formatter rc:5 problems (fixed in df_ret/df_deb)

  20030306 14.093g

       - w_dup fix back in

       - another shot at fixing the trap 8 problems while running an ISS or

      using tcpbeui. If you're using tcpbeui (*especially *under SMP), you

      should not use this. If you're using an ISS, we ignore this. For others,

      you may get some speed improvement in loading apps by adding

      LDRSTACKOPT=FALSE to config.sys (the default being true).

  20030415 14.093h

       - LDRSTACKOPT back out. It caused too much confusion and I'm not 100%

         sure about its stability. Sigh.

       - Some enhancements to the ".pb" command in the dump formatter

  20030501 14.093i

       - I kept forgetting to take out the code that closes file handles > 2

         before starting PROTSHELL. The good news is that doing the close

reduces the number of leaked file handles by a few in some systems.

The bad news is that it also makes PMDD$ trap on boot on other

systems. So, if that happened to you with 14.093h, get this kernel.

As of now, add CLOSEFILES=TRUE if, like me, you need the changed

behavior (in my case, Microsoft C 6 kept aborting).

       PJ29180  Trap 000E while taking a process dump

  20030620 14.093i1

  - don't use this one. I put it out by accident due to packaging bug.

  20030621 14.094

  - nothing new, but this is the first "official" (i.e. one where we'll

  actually have full source code control) in a while. It leaves out

  the CLOSEFILES keyword, though.

  20030822 14,096

  PJ29349  Trap in ThePD_Get (another theseus/4 trap)

PJ28318  problems accessing non-US filenames (regression due to PJ29234)

         (related fix to cmd.exe to back out bogus changes will be 

  available in next fixpack)

  20031031  14.096b

- CLOSEFILES= back in

- fix 2 traps in w_DupHandle

   Seems to affect Netfinity remote access.

PJ29849  trap 000c in w_NewSize during procdump processing

  20031103  14.096c

        - incorrect parameters being reported in some devhelp system traces

  20031110  14.096c

        - 1103 system trace fix not built into some kernels

- fix named pipe regression in 1031 and 1103 kernels

- os2ldr is now the version that "works" with shared interrupts. 

  HOWEVER, if your system has shared interrupts that didn't used to

  get noticed, then you may suddenly get much slower. This happens on

  my personal Intellistation which shares an IRQ for the USB and audio.

  20031113  14.096e

- trace fix didn't get into previous w4 and uni kernels

- EXPERIMENTAL: Removed a bunch of redundant locking and unlocking of

  some pages in filesystem code that's already pinned, anyway. This

  ought to speed up a number of filesys-related APIs. Also, re-enabled

  the un-pinning of numerous pages of memory from init code. This

  memory is now available for other uses, which should reduce paging on

  many systems. This had been turned off since WSeB!

PJ29472: process hangs (and other things?) if named pipe handle is

   closed while another thread is blocked in DosConnectNmPipe.

  20031117  14.097b  (post xr_c004)

  - Some of the pages in the 20031113 package may possibly need to stay

  resident to avoid any danger of a trap 000e if they're needed during

  page fault processing or interrupt time. So, this time, I left those

  pages resident,too.

  20040116 14.097d

  - the code lock stuff seems to be working well, so it's no longer 

   "experimental."

- possible DosConnectNmPipe regression test fix

- Support for loading pure 32 bit code and data objects in DLLs above

  the 512mb line. More details on that to come separately.

  20040119 14.097d  (still)  *DON'T USE THIS OS2LDR*

  No kernel changes from 0116, but OS2LDR changes:

  - the interrupt sharing fix had, at some point, got taken back out.

  users with seeming IRQ issues should try this

- possible problem booting up P4 Xeon 4 gHz (hang at boot blob)

  20040123 14.097e

  - fix bad cFH count in DosQuerySysState/QProcStatus with level=0

- attempt to fix intermittent trap 000e resulting from highmem loadin

  in 116 kernel

    - bad fix for boot blob hang in 20040119 caused "unable to operate

  hard disk" failures with os2ldr from 20040119 packages

  20040126 14.097f

  - fix bad file handle array in DosQuerySysState/QProcStatus with level=0

  - more changes related to boot hangs and other weirdness from 1/23

  20040126 14.097g

  - more on file handle array in DosQuerySysState/QProcStatus with level=0

  - another attempt to update os2ldr to work with everything from 486SX to P4 Xeon >3gHz....

PJ29595: trap 000e at SftFileSize + 66 or so

  20040206 14.097h

  PJ29602 trap 000d at h_DOS_ChDir + 28 or so after media change

- backed out bogus os2ldr changes (again)

  20040323 14.098a

  PJ29697 add debug support for disappearing browser windows

- fix DevIOCtl passthru support

- fixed the "k" command in the dump formatter (at least it's better)

  20040330 14.098b

  PJ29712 (os2ldr) final fix for boot blob hang on fast (around 3+ gHz) CPUs

- backed out some kernel stuff that accidentally got built

Following are rough instructions for applying the os2krnl fix:

In the directory from which you boot (C:\ or d:\ or whatever), type these

commands. Note that os2krnl should have come with this file.

Also, there are some REXX scripts written by third parties that might make

installation/archiving of your kernels easier. Go to http://hobbes.nmsu.edu

and search for "moritz", for example. This is unsupported by IBM!

1. attrib -r -s -h os2krnl

2. attrib -r -s -h os2ldr

3. copy os2krnl os2krnl.sav

4. copy os2ldr os2ldr.sav

5. copy \os2\system\trace\*.tdf  \save       (or wherever)

5. unzip -o xxx.zip

If you want, you can turn the os2krnl and os2ldr attributes back on, but

it's not necessary to do so.

Trace files (*.tdf)  have been included.

An os2krnl.sym file has been provided with this fix. Just save it somewhere

(root directory or anywhere). It is only needed in case OS/2 

service requests a standalone dump of your system in the future.

(C) Chris Wohlgemuth 2002-2003

http://www.os2world.com/cdwriting

http://www.geocities.com/SiliconValley/Sector/5785/

This package contains drivers to use USB mass storage

class compliant devices with OS/2 and eCS.

The mass storage driver supports other, maybe more, devices

than the IBM drivers in particular the 'Archos Jukebox

Recorder 20' (In-System Design USB 2.0 bridge), Olympus-C1

and Casio digital cameras.

Included are:

CWUSBUHC.SYS  UHCI compliant USB host controller driver 

CWUSBOHC.SYS  OHCI compliant USB host controller driver

CWUSBD.SYS    USB driver

CWUSBMSD.ADD  USB mass storage class compliant driver.

HowTo for using the 'Archos Jukebox Recorder 20' with OS/2.

License:

This package is provided 'as is' without any

warranty. See documentation for more.

작성자: 좀딱 

1. 소개

  이 판은 FreeType/2 에 DBCS 특히, 한글 관련 기능을 강화시킨 것입니다. 원판에 

덧붙여진 점은 다음과 같습니다.

  첫째, 주코드페이지에 맞추어 글꼴 이름을 보여줍니다.

  둘째, 영문으로 인코딩된 글꼴 이름에 DBCS 코드가 있을 경우 이를 제대로 

처리합니다.

  섯째, 한글 완성형 글꼴과 유니코드 글꼴 지원 기능이 추가되었습니다.

  넷째, TTF 글꼴의 언어 ID 와 인코딩이 다를 경우, 예를 들어 언어 ID 는 

영어이지만, 인코딩은 완성형인 경우, 글꼴 이름을 제대로 처리합니다.

  다섯째, TTF 글꼴의 문자 배열 인코딩 순서가 Apple Roman 과 Wansung 으로 되어 

있을 때, Apple Roman 으로 인식하는 문제를 해결했습니다.

  여섯째, 한자 글꼴이 포함되어 있지 않은 경우, 한자음을 보여줍니다.

  일곱째, UGL(Universal Glyph List) 처리를 추가해서, 한글 글꼴에서 보이지 

않던 글자들을 볼 수 있습니다. 예를 들어, '\'(원화) 를 볼 수 있습니다.

  여덟째, '\' 에 대한 처리를 강화하여, FreeType/2 가 관리하는 글꼴에서는 '\' 

를 제대로 볼 수 있습니다.

  아홉째, 윈도용 Arial 글꼴을 DBCS 글꼴로 인식하던 문제를 고쳤습니다.

  열째, 프로그램이 UCONV.DLL 에서 자주 죽는 문제를 고쳤습니다.

  열한번째, 포함하고 있는 GLYPH 수로 DBCS 글꼴을 판단하는데, 이 기준값을 더 

크게 잡았습니다.

  열두번째, 내장 비트맵(sibt)을 쓸 경우에 글자 반쪽이 잘리는 현상을 

고쳤습니다.

  열세번째, 글꼴에 변형(회전/기울이기)이 가해질 경우에는 내장 비트맵(sbit)을 

사용하지 않습니다.

* 열네번째, 글꼴 이름을 인식하지 못할 경우, PM 으로 넘어가지 못하는 현상을 

고쳤습니다. eCS 1.1 을 설치할 때, 중국어 글꼴을 설치하는 경우에 나타나는 

현상입니다. 이 경우 해당 글꼴의 이름은 'Wrong Name Font' 가 됩니다.

* Arial Bold Italic 글꼴이 Arial Bold Ital 로 인식되는 것을 고쳤습니다.

* 글꼴 이름이 너무 길 경우에는 적당한 길이로 자릅니다.

  예) Monotype Sans Duospace WT K Bold -> Monotype Sans Duospace WT K Bol

작성자: 좀딱 

1. 소개

  이 판은 FreeType/2 에 DBCS 특히, 한글 관련 기능을 강화시킨 것입니다.

원판에 덧붙여진 점은 다음과 같습니다.

    첫째, 주코드페이지에 맞추어 글꼴 이름을 보여줍니다.

    둘째, 영문으로 인코딩된 글꼴 이름에 DBCS 코드가 있을 경우 이를 제대

로 처리합니다.

    섯째, 한글 완성형 글꼴과 유니코드 글꼴 지원 기능이 추가되었습니다.

    넷째, TTF 글꼴의 언어 ID 와 인코딩이 다를 경우, 예를 들어 언어 ID 

는 영어이지만, 인코딩은 완성형인 경우, 글꼴 이름을 제대로 처리합니다.

    다섯째, TTF 글꼴의 문자 배열 인코딩 순서가 Apple Roman 과 Wansung 

으로 되어 있을 때, Apple Roman 으로 인식하는 문제를 해결했습니다.

    여섯째, 한자 글꼴이 포함되어 있지 않은 경우, 한자음을 보여줍니다.

    일곱째, UGL(Universal Glyph List) 처리를 추가해서, 한글 글꼴에서 보

이지 않던 글자들을 볼 수 있습니다. 예를 들어, '\'(원화) 를 볼 수 있습니

다.

    여덟째, '\' 에 대한 처리를 강화하여, FreeType/2 가 관리하는 글꼴에서

는 '\' 를 제대로 볼 수 있습니다.

    아홉째, 윈도용 Arial 글꼴을 DBCS 글꼴로 인식하던 문제를 고쳤습니다.

    열째, 프로그램이 UCONV.DLL 에서 자주 죽는 문제를 고쳤습니다.

*** 열한번째, 포함하고 있는 GLYPH 수로 DBCS 글꼴을 판단하는데, 이 기준값을 더 

    크게 잡았습니다.

*** 열두번째, 내장 비트맵(sibt)을 쓸 경우에 글자 반쪽이 잘리는 현상을 

    고쳤습니다.

*** 열세번째, 글꼴에 변형(회전/기울이기)이 가해질 경우에는 내장 

    비트맵(sbit)을 사용하지 않습니다.

OS/2 Warp 4.0에 설치해보았습니다. 설치 CD를 넣으면 설치 프로그램이 시작됩니다. 윈도우용과는 배경부터 해서 완전히 다르네요. 계속을 누릅니다. 

등록 정보를 입력합니다. 사용자명과 회사명, 제품 번호를 입력하고 계속을 누릅니다. 

입력한 정보를 확인한 후 예를 누릅니다. 

설치 방법을 선택합니다. 기본 설치를 선택합니다. 

시스템의 드라이브를 스캔 중입니다. 저 눈이 살짝 무섭네요. 

만약 사용자 설치를 하게 되면 이렇게 설치할 구성 요소를 선택할 수 있습니다. 계속을 누릅니다. 

한글이 설치될 새로운 디렉토리를 만듭니다. 확인을 누릅니다. 

파일 복사 중입니다. CD지만 CD 안에 디스켓 내용 구성물이 다 있기 때문에 이렇게 나옵니다. 

파일 복사가 거의 마무리되었습니다. 48번까지 있다니, 플로피 버전으로 설치했다면 생각만 해도 정말 끔찍하네요. 

이전 버전의 config.sys 파일을 config.ins 파일로 복사했다고 뜹니다. 

config.sys 파일을 수정할 건지 묻습니다. 수정한다고 하고 OK를 누릅니다. 

config.sys 파일이 수정되었습니다. 

한글 3.0b의 설치가 끝났습니다. 시스템을 재시작합니다. 

설치 후 작업표시줄의 한글과컴퓨터 항목을 열어보았습니다. 한글 3.0b와 HNC 구성, readme 파일로 구성되어 있습니다. 

한글 3.0b를 실행해보았습니다. 로고는 윈도우용과 동일합니다. 

OS/2용 한글 3.0b의 모습입니다. 아이콘의 배치만 다를 뿐이지 외양은 거의 같습니다. 

한글 입력 또한 매우 잘 됩니다. 

메뉴를 열어보았습니다. 저 특유의 둥글둥글한 글꼴까지 똑같습니다! 

설치 종료 시 readme 파일을 꼭 열어보라고 해서 한번 열어봤습니다. 말 그대로 안내문입니다. 

한글 3.0b 정보입니다. 저 금색 3.0b 로고는 언제 봐도 고급진 것 같습니다. 

이상 OS/2용 한글 3.0b 설치기였습니다.

                                                         Process Commander Version 1.0

                                                          November 6, 1996

                                (c) 1995-96 Stardock Systems Incorporated

 Installation:

 -------------

     To install the software, run the program INSTPC.EXE from the A:

     drive or a directory you copy the contents of A: to.  You may

     selectively install or uninstall components, and you can

     specify the target directory.  The installation takes about 6MB

     (3 MB temporary space).

     The installation has been carefully tested. If the installation

     does not work sucessfully try the following:

     (P=possible problem, A=Answer)

     P: No files installed at all

     A: Your installation disk may be damaged or there is not enough space

        on the target disk.

     P: Workplace Shell integration not successfully installed

     A: 1. deinstall WPS integration (no matter whether it is installed or not).

        2. reboot

        3. install WPS integration

     P: Menu extension not installed

     A: Do not forget to rename pcframe.bak to pcframe.dll (cf. manual)

     P: PC cannot be activated

     A: Is the PC driver installed? run pcstatus.exe to find out.

        On systems that are not compatible with the PC driver you

        have to use pckbd.sys. (cf. manual, WWW pages)

     If you contact Stardock with installation problems please have the file

     instpc.log at hand or email it to support@stardock.com

 Configuration:

 --------------

     You can also use the installation program INSTPC to change the

     configuration and to remove Process Commander components from

     the system.

 Uninstallation:

 ---------------

     To uninstall Process Commander run INSTPC.EXE and select "uninstall".

     The uninstall does not remove files from your system. So you can

     easily remove all references to Process Commander in your config.sys

     with the option to reinstall easily later.

     After a successful deinstallation you delete the complete Process

     Commander directory tree. We do not recommend to delete files before

     using the deinstall.

 Special Instructions for OS/2 2.x Users, Matrox users, japanese OS/2

 --------------------------------------------------------------------

     The OS/2 keyboard driver on OS/2 2.x and japanese OS/2 may not

     be compatible with the Process Commander keyboard driver. Some

     users reported problems with special versions of the Matrox

     display adapter, because those adapter have their own keyboard

     driver. In those cases please restore the old driver (cf.

     \OS2\BOOT\KBDBASE.TXT) and insert the line:

        DEVICE=C:\PC\TMP\PCKBD.SYS KEYB:CA-A

     to activate with Ctrl-Alt-A. Hotkeys for other things than

     activation are currently not supported by PCKBD.SYS

 Technical Support:

 ------------------

       WWW:          Stardock Homepage: http://www.stardock.com

                     PC Homepage:       http://www.stardock.com/pc.html

                     Trouble Shooting:  http://www.stardock.com/pcbugidx.htm

       CompuServe:   On CompuServe type GO STARDOCK.

       IBMLINK:      On the IBM IBMLINK service, there is a STARDOCK CFORUM.

       Stardock BBS: The Stardock BBS (313)/453-1845 in the Process Commander

                     file and message area.

       email:        support@stardock.com

       Updates to Process Commander will be available on the

       BBS and the World Wide Web site.

       If you need to speak with a technical support representative,

       call (313) 453-0328 Monday through Friday, 9 a.m. to 4 p.m.,

       Eastern Standard Time.

 Sales:

 ------

        Site license pricing for Process Commander is available.

        Please contact us at (313) 453-0328 or sales@stardock.com

        for more information about our site license programs for

        Process Commander.

 Copyrights and Trademarks:

 --------------------------

        OS/2 is a trademark of International Business Machines

        Corporation

        Process Commander is a trademark of Stardock Systems, Inc.

        (c) 1996 Stardock Systems (SDS) Inc.

GAMMA Version 1 (March 31, 1997)

THIS PROGRAM IS FOR IBM INTERNAL USE ONLY!

Thank you for your interest in this Beta program.  Please read these

notes before installing and using the Beta version of CHKDSK.

CONTENTS

________

This file contains the following information:

1.0 HIGHLIGHTS OF THIS VERSION OF CHKDSK

2.0 INSTALLING THE CHKDSK BETA

3.0 REMOVING THE CHKDSK BETA

4.0 USING THIS VERSION OF CHKDSK

5.0 TRADEMARKS

1.0 HIGHLIGHTS OF THIS VERSION OF CHKDSK

___________________________________

This version of CHKDSK offers significant improvements for use

with High Performance File System (HPFS) drives:

1.  Single pass correction of errors.

    You can run CHKDSK once to fix all errors on an HPFS drive.

2.  Performance improvements.

    CHKDSK execution is now up to two times faster on non-RAID

    drives, and up to 11 times faster on large RAID drives.

3.  Reduced memory requirement.

    CHKDSK now uses less memory to check large drives (> 4GB).

4.  Improved FOUND directory format.

    Files and directories recovered by CHKDSK and placed in the

    FOUND directory now have their original names (as much as can

    be recovered) appended to the names assigned by CHKDSK.

5.  More sophisticated error detection.

    This version of CHKDSK recognizes errors that previous versions

    did not.

6.  More sophisticated error correction.

    This version of CHKDSK will recover files and directories in cases

    where previous versions would have deleted them.

7.  Binary log file created in the root directory for use by Service.

    File name is chkdsk.log and a backup is kept in chkdsk.old.

    Service has a tool (PMCHKLOG) to format the logfile.

Improvements in this version of CHKDSK are for HPFS drives only.

CHKDSK operation for File Allocation Table (FAT) drives has not

changed from prior versions.

This enhanced CHKDSK operates on IBM OS/2 Warp 3.0 and higher,

including IBM Warp Server 4 and IBM Warp Server 4 SMP.  IBM OS/2 2.1

and IBM OS/2 2.1 SMP are NOT supported.

2.0 INSTALLING THE CHKDSK BETA

__________________________

To install this version of CHKDSK:

        1) Backup your system.

        2) At an OS/2 prompt, type INSTALL and press enter.

        3) Follow the prompts to install CHKDSK on your boot drive.

Installing this version of CHKDSK results in the following changes:

        a) Two files are added to the drive being checked:

                   chkdsk32.exe

                   chkdsk32.dll

        b) The existing uhpfs.dll file is replaced

        c) The original uhpfs.dll is renamed uhpfs.sav

These changes cause the Beta version to run instead of the previously

installed version of CHKDSK (including AUTOCHECK operation).

If you want to be able to choose between the Beta and the previous

CHKDSK version, take the following actions:

        1) Make sure there is a .; entry in your LIBPATH statement in

            config.sys ahead of \os2\dll; .

        2) Copy the beta uhpfs.dll to a new subdirectory of your choice;

            the new subdirectory cannot be in your LIBPATH statement.

        3) Restore the original uhpfs.dll from uhpfs.sav .

With these changes, the Beta version of CHKDSK operates when initiated

from the new subdirectory; otherwise, the previously  installed version of

CHKDSK operates.

3.0 REMOVING THE CHKDSK BETA

____________________________

The following actions will remove the CHKDSK Beta:

        1) On the drive on which you installed the Beta version, go to

            the \os2 subdirectory.

        2) Erase the following file:  chkdsk32.exe

        3) Go to the \os2\dll subdirectory on the same drive.

        4) Erase the following file:  chkdsk32.dll

        5) Verify uhpfs.sav is in this subdirectory.

        6)  Erase the following file:  uhpfs.dll

        7) Copy uhpfs.sav to uhpfs.dll

The CHKDSK Beta is no longer installed.  The previously installed version

of CHKDSK is again your current version.

4.0 USING THIS VERSION OF CHKDSK

____________________________

The interface to CHKDSK remains CHKDSK.COM and PMCHKDSK.  The

command line parameters and messages from CHKDSK are unchanged.

CHKDSK creates a binary log (chkdsk.log) on the root of the drive

being checked.  The previous log file is saved as chkdsk.old .

As with prior versions of CHKDSK, AUTOCHECK of drives over 8GB in

size is not supported.  Drives over 8GB in size should be checked using

the /C option of CHKDSK as follows:

        1) Place a call to CHKDSK in your STARTUP.CMD file and specify

           the drive to check, along with the /C.

        2) CHKDSK then examines the drive and checks the drive if

            needed.

The first time this Beta version of CHKDSK runs on a drive, it

will report errors if the drive was fast formatted.  These errors are

harmless and go away if CHKDSK is run with the /F option.

5.0 TRADEMARKS

_____________

The following terms are trademarks of the IBM Corporation in the

United States or other countries or both:

        IBM

        OS/2

THIS DOCUMENT IS PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND.

IBM DISCLAIMS ALL WARRANTIES, WHETHER EXPRESS OR IMPLIED,

INCLUDING WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF FITNESS

FOR A PARTICULAR PURPOSE AND MERCHANTABILITY WITH RESPECT TO

THE INFORMATION IN THIS DOCUMENT.  BY FURNISHING THIS DOCUMENT,

IBM GRANTS NO LICENSES TO ANY PATENTS OR COPYRIGHTS.

(C) Copyright IBM Corporation 1997.  All rights reserved.

Version:             2.4

Publisher:           CodeSmith Software, RR 1, S-2, C-23

                     Fernie, BC, Canada V0B 1M0  (Internet: mshill@elkvalley.net)

Preferred Filename:  XIT24.ZIP

System Requirements: OS/2 2.0 or later, any desktop manager (WPS, Filebar, etc)

Description:         Desktop enhancement utility. Add user defined titlebar

                     buttons on left or right side. Completely configurable.

                     Also supports keys and middle/right mouse button actions.

Keywords:            Utility, Desktop, OS/2, Xit

LIMITED DISTRIBUTION LICENSE:

Permission is granted to individuals charging no fees, and to BBSes,

User Groups, Mail-Order Disk Vendors, CD-ROM publishers, and BBS

File-Distribution Networks regardless of fees, to distribute Xit,

provided they convey a complete and unaltered copy of the shareware version

of this product, and provided they claim no ownership of these products.

Distribution by other channels, and distribution of modified versions, are

also allowed, but only with written permission from CodeSmith Software.

Permission to distribute these products is not transferable, assignable,

saleable, or franchisable.  Each vendor wishing to distribute these

products must independently satisfy the terms of this limited distribution

license.

CodeSmith Software may revoke any permissions granted here, by notifying

you in writing.

emxrt.doc    emx 0.9c FIX 00     USER'S GUIDE TO THE EMX RUNTIME    09-Sep-1996

===============================================================================

                                     Copyright (c) 1990-1996 by Eberhard Mattes

Table of Contents

=================

1     Introduction to the User's Guide to the emx Runtime

1.1     Hardware requirements

1.2     Files included in the emx runtime package

2     License

3     Installing the emx runtime package under OS/2

3.1     Summary for impatient OS/2 users

3.2     Detailed instructions for OS/2 users

4     Installing the emx runtime package under DOS

4.1     Summary for impatient DOS users

4.2     Detailed instructions for DOS users

5     Environment variables (OS/2 and DOS)

5.1     Environment variables (DOS)

6     DOS Notes

6.1     General notes

6.2     Emergency exit

6.3     Using rsx instead of emx

7     emx options

7.1     emx options (OS/2)

7.2     emx options (DOS)

8     The emxstack tool

9     The emxupd tool

9.1     Replacing a DLL

9.2     Deleting a DLL

9.3     Unpacking DLLs from a ZIP file and installing them with emxupd

10    Troubleshooting

10.1    Warning message `WARNING: emx 0.9c or later required'

10.2    unzip.exe says `cannot create emx/dll/emx.dll'

10.3    emx.dll cannot be deleted, renamed or overwritten

10.4    Error message `SYS1804: The system cannot find the file EMX.'

10.5    Error message `rsx not found, DPMI not supported by emx'

10.6    Error message `emx.dll: Stack size too small...'

10.7    DOS message `Sector not found'

11    emx messages

11.1    emx messages (OS/2)

11.2    emx messages (DOS)

12    Obtaining the complete emx package

13    The author of emx

1 Introduction to the User's Guide to the emx Runtime

=====================================================

This is the documentation for the emx runtime package.  The emx

runtime is an environment for 32-bit programs under OS/2 2.x, OS/2 3.x

(OS/2, in short), MS-DOS, and PC-DOS (DOS, in short) on machines with

a 386 CPU (or one of its successors).

Under OS/2, you can view the .inf version of this document by typing

    view \emx\book\emxrt

If you add \emx\book to the BOOKSHELF environment variable, you can

type

    view emxrt

to view that on-line book.

1.1 Hardware requirements

-------------------------

emx requires an 80386 CPU (80386DX, 80386SX, i486, or later)

1.2 Files included in the emx runtime package

---------------------------------------------

The emx runtime package includes the following files:

  \emx\bin\emx.exe              emx runtime for DOS

  \emx\bin\emxbind.exe          Tool for modifying emx programs (DOS and OS/2)

  \emx\bin\emxfpemu             Emulator for numeric coprocessor (DOS)

  \emx\bin\emxl.exe             emx loader (used by emxbind)

  \emx\bin\emxload.exe          Server for keeping OS/2 programs in memory

  \emx\bin\emxrev.cmd           Tool for displaying emx revisions (OS/2)

  \emx\bin\emxstack.exe         Tool for fixing the stack size of old programs

  \emx\bin\emxupd.exe           Tool for updating a DLL which is in use

  \emx\bin\emxuzdll.cmd         Unpack DLLs and install them with emxupd.exe

  \emx\book\emxrt.inf           The text you are looking at (VIEW version)

  \emx\dll\emx.dll              emx runtime for OS/2

  \emx\dll\emxio.dll            Hardware port I/O under OS/2

  \emx\dll\emxlibc.dll          Forwarder DLL to emxlibcm.dll (OS/2)

  \emx\dll\emxlibcm.dll         Multithread C library (OS/2)

  \emx\dll\emxlibcs.dll         Single-thread C library (OS/2)

  \emx\dll\emxwrap.dll          Interface to 16-bit OS/2 functions (OS/2)

  \emx\doc\COPYING              GNU General Public License

  \emx\doc\COPYING.EMX          emx copyright information

  \emx\doc\emxrt.doc            The text you are looking at (text version)

  \emx\etc\termcap.dat          Terminal data base for termcap-based programs

  \emx\etc\emclibcm.map         Map file for emxlibcm.dll

  \emx\etc\emclibcm.tdf         Trace definition file for emxlibm.dll

  \emx\etc\emclibcs.map         Map file for emxlibcs.dll

  \emx\etc\emclibcs.tdf         Trace definition file for emxlibs.dll

  \emx\etc\trc00f8.tff          Trace format file for emxlibcs.dll

  \emx\etc\trc00f9.tff          Trace format file for emxlibcm.dll

The .map, .tdf, and .tff files are used for debugging.  They are

included in emxrt.zip to keep them in sync with the DLLs.

2 License

=========

emx is free software; you can redistribute it and/or modify it under

the terms of the GNU General Public License as published by the Free

Software Foundation; either version 2, or (at your option) any later

version.

emx is distributed in the hope that it will be useful, but WITHOUT ANY

WARRANTY; without even the implied warranty of MERCHANTABILITY or

FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License

for more details.

You should have received a copy of the GNU General Public License

along with emx; see the file COPYING.  If not, write to the Free

Software Foundation, 59 Temple Place - Suite 330, Boston, MA

02111-1307, USA.

See \emx\doc\COPYING.EMX and \emx\doc\COPYING for details.

3 Installing the emx runtime package under OS/2

===============================================

3.1 Summary for impatient OS/2 users

------------------------------------

1.  Add \emx\dll to LIBPATH (always use the latest copy of the .dll

    files!)

2.  Add \emx\bin to PATH

3.  Reboot

3.2 Detailed instructions for OS/2 users

----------------------------------------

emx does not work with OS/2 1.x and in DOS mode of OS/2.

Before changing config.sys, make a backup copy of that file!

The emx runtime is implemented as a set of dynamic link libraries

under OS/2:

emx.dll

        emx run-time support for OS/2

emxio.dll

        access to hardware ports for emx programs under OS/2

emxlibc.dll

        forwarder DLL to emxlibcm.dll for existing applications that

        use emxlibc.dll (which has been renamed to emxlibcm.dll)

emxlibcm.dll

        multithread emx C library

emxlibcs.dll

        single-thread emx C library

emxwrap.dll

        32-bit wrappers for 16-bit OS/2 API functions

OS/2 looks for dynamic link libraries in the directories listed in the

LIBPATH statement of your config.sys file.  Note that LIBPATH is not

an environment variable!

At any time, at most one version of a dynamic link library is loaded.

If there are different versions of a dynamic link library in the

directories listed in the LIBPATH statement, the first one listed is

loaded.  If you already have installed the emx dynamic link libraries,

you should be careful to use the latest revision of those dynamic link

libraries.  Newer emx dynamic link libraries work with older emx

programs (however, you may have to apply emxstack), but newer emx

programs don't work with older emx dynamic link libraries.  To find

out about already installed emx dynamic link libraries, type

    cd \emx\bin

    emxrev

If no emx dynamic link libraries are installed, no output is produced.

If there are emx dynamic link libraries installed, the names and

revision numbers are displayed, for instance

    EMX : revision = 30

    EMXIO : revision = 30

    EMXLIBC : revision = 30

    EMXLIBCM : revision = 30

    EMXLIBCS : revision = 30

    EMXWRAP : revision = 30

Only directories listed in the LIBPATH statement are searched.  To

find out the revision numbers of the dynamic link libraries of this

package, type

    emxrev -d c:\emx\dll

if you have put the emx runtime package on drive C.  After running

emxrev, you should quit the command processor by typing

    exit

to release the dynamic link libraries (they have been loaded by

cmd.exe as emxrev is a REXX procedure).  Otherwise, you cannot rename,

overwrite, or delete those files.  This also happens if one of the emx

dynamic link libraries is currently used by a program.  If you get

error message SYS0032, or the message

    error:  cannot create emx/dll/emx.dll

from unzip when unpacking emxrt.zip, use

    pstat /L

to find out about the dynamic link libraries used by the programs

currently running.  Terminate all the programs using emx.dll,

emxio.dll, emxlibc.dll, emxlibcm.dll, emxlibcs.dll, or emxwrap.dll.

The command

    emxload -qw

will unload programs kept in memory by emxload.exe.

If that doesn't help, use emxuzdll to install the emx DLLs: Create a

temporary directory and run emxuzdll in that directory, giving the

path name of emxrt.zip and the path name of the emx DLL directory on

the command line:

    mkdir \emxtmp

    chdir \emxtmp

    emxuzdll a:\emxrt.zip c:\emx\dll

    chdir ..

    rmdir emxtmp

This will replace the DLLs in c:\emx\dll with the DLLs in emxrt.zip

even if the DLLs are in use.  The new versions of the DLLs become

active as soon as the last process referencing them is terminated, or

on the next reboot.  Therefore, you should reboot after running

emxuzdll, otherwise you won't know if and when the new versions become

active.  (Actually, you don't need a temporary directory, any

directory not containing any DLL files will do.)

Either copy the latest revisions of the emx dynamic link libraries

(that is, those with the highest revision numbers) to a directory

directory listed in the LIBPATH statement of config.sys or, better,

modify the LIBPATH statement to include the directory where the latest

revisions of the emx dynamic link libraries can be found.

The directories of the LIBPATH statement are separated by semicolons.

To add c:\emx\dll to the LIBPATH statement, append a semicolon -- if

there is not yet a semicolon -- to the end of the line of config.sys

that starts with

    LIBPATH=

Then, type the absolute name of the directory.  Example:

    LIBPATH=.;C:\OS2\DLL;C:\OS2\MDOS;C:\;C:\OS2\APPS\DLL;D:\EMX\DLL

Optionally, you can append a semicolon at the end.

Next, modify the line of config.sys starting with

    SET PATH=

to include the \emx\bin directory.  For instance, change

    SET PATH=C:\OS2;C:\OS2\SYSTEM;C:\OS2\MDOS\WINOS2;C:\OS2\INSTALL;

    C:\;C:\OS2\MDOS;C:\OS2\APPS

to

    SET PATH=C:\OS2;C:\OS2\SYSTEM;C:\OS2\MDOS\WINOS2;C:\OS2\INSTALL;

    C:\;C:\OS2\MDOS;C:\OS2\APPS;D:\EMX\BIN

As with LIBPATH, you have to put a semicolon between the path names

and a trailing semicolon is optional.  Note that the last two examples

are shown in two lines; in config.sys, however, the complete SET

statement must be put into one line.

You might want to add \emx\book to the BOOKSHELF environment variable

to simplify viewing the on-line documentation.  For instance, change

the line

    SET BOOKSHELF=C:\OS2\BOOK;

of your config.sys file to

    SET BOOKSHELF=C:\OS2\BOOK;C:\EMX\BOOK

as with PATH, you have to put a semicolon between the path names and a

trailing semicolon is optional.

Now, you should reboot to activate the changes to config.sys.  After

booting, run

    emxrev

to verify that the latest revision of emx is now active.

If OS/2 says

    SYS1804: The system cannot find the file EMX.

when running an emx program, you have not correctly set the LIBPATH

statement of config.sys.  See above for details.

When running an old program (which assumes that emx.dll creates a

stack object if the stack size is too small), you will get an error

message like

    emx.dll: Stack size too small.  Run

      emxstack -f D:\BIN\PROGRAM.EXE

    and try again.

Run emxstack as indicated by the error message to fix the stack size

of the program.  If you can't (because you don't have write permission

for the file), type

    set emxopt=-I

to make emx.dll ignore the problem.  However, the program will

probably crash due to insufficient stack size.  Using the -I option is

not recommended!

4 Installing the emx runtime package under DOS

==============================================

4.1 Summary for impatient DOS users

-----------------------------------

1.  Add \emx\bin to PATH or set EMX to point to \emx\bin\emx.exe

2.  If you are using a memory manager which implements DPMI, enable

    VCPI

3.  Remove BREAK=ON from config.sys

4.  Reboot

4.2 Detailed instructions for DOS users

---------------------------------------

emx programs for DOS come in two flavours: those which load emx.exe

from a separate file and those which have the emx runtime emx.exe

integrated in the executable file of the application program.

Before changing autoexec.bat and config.sys, make a backup copy of

these files!

If you want to use a program which loads emx.exe, you have to put

emx.exe into a directory which is listed in your PATH environment

variable or you have to change the PATH environment variable (in

autoexec.bat) to include the directory in which emx.exe can be found.

For instance, if your autoexec.bat file contains the line

    path c:\dos;d:\bin

you have installed the emx runtime on drive D, you should modify that

line to read

    path c:\dos;d:\bin;d:\emx\bin

You have to put a semicolon between the path names.

Alternatively, you can set an environment variable which directly

points to the emx.exe executable.  If you use this method, your

program can locate emx.exe more quickly.  If you have installed the

emx runtime on drive D:, you should add the following line to your

autoexec.bat file:

    set emx=d:\emx\bin\emx.exe

Note that a directory name is not sufficient -- you have to type the

complete path name of the emx.exe file.

Programs of the second flavour, having emx.exe integrated, don't

require \emx\bin added to the PATH environment variable or the EMX

environment variable being set.  The program itself, however, may need

changes to your autoexec.bat file.  See the documentation accompanying

the program.  The following steps are required for both flavours of

emx programs.

If you are using a memory manager (such as QEMM or EMM386 or 386max)

which implements DPMI (DOS Protected Mode Interface), you have to

change the invokation of that memory manager to enable VCPI as emx

does not support DPMI.  See the documentation of your memory for

details.  If your memory manager supports DPMI only (without VCPI),

you have to remove or disable the memory manager.  Alternatively, you

can use rsx, as indicated in section 6.3.  emx supports XMS and VCPI.

Examine your config.sys file: If there is a line

    BREAK=ON

in it, you should remove it as it causes trouble with emx and lots of

other programs.

If and only if you are using an `Intel Inboard 386/PC' expansion board

in your 8088-based PC, add the following line to autoexec.bat (it's

very unlikely that you are using such an expansion board):

    set emxopt=-m3

If the swap file (used to swap memory pages to disk when running out

of physical memory) should be in a specific directory or on a specific

disk drive, set the EMXTMP environment variable in autoexec.bat.

Example:

    set emxtmp=c:\tmp

Now, you should reboot to activate the changes to config.sys and

autoexec.bat.

5 Environment variables (OS/2 and DOS)

======================================

The operation of emx and emx programs is controlled by several

environment variables.  You might have to set one or more of the

following environment variables:

INIT    used by termcap to find termcap.dat

TERM    used by termcap to define the terminal type

EMXETC  used by termcap to find termcap.dat

EMXOPT  emx options (see below)

EMXSHELL

        overrides the COMSPEC environment variable.  You can use this

        to run an alternate command processor (shell) from emx

        programs

EMXPATH

        search path for programs (usually not relevant)

PATH    search path for program (used if EMXPATH fails)

TERMCAP

        used by termcap to find termcap.dat

5.1 Environment variables (DOS)

-------------------------------

EMX     path name of emx.exe, used by programs which are bound with

        emxl.exe (the emx loader) instead of emx.exe.  Set EMX to

        speed up loading of emx.exe or to load a specific version of

        emx.exe

EMXTMP  directory for the swap file

TMP     used if EMXTMP is not set

6 DOS Notes

===========

6.1 General notes

-----------------

- emx requires DOS 3.0 or later

- emx does not support DPMI

- emx does not support DESQview without VCPI (use QEMM-386 with

  DESQview)

- If emx doesn't work under DESQview (especially if it displays the

  message

      Nonmaskable interrupt (NMI)

  either set the protection level to 0 or use the WATCHDOG=0 option of

  QEMM-386.

- emx supports VCPI

- emx supports XMS.  Do not use emx with himem.sys 2.06 or earlier.

- emx checks for XMS version number 2.00 or later and for XMS driver

  revision 2.06 or later, as older himem.sys drivers don't work

  correctly.  You can override this check by using the -O emx option

  (for drivers using a different revision numbering scheme), but emx

  may not work with your XMS driver, anyway.  Actually, emx has not

  been tested with himem.sys 2.05; 2.04 fails, 2.06 works.

- emx supports vdisk.sys 3.3 and later and most other RAM disk drivers

- emx uses (or tries to use) all available memory, whether

  conventional, extended, or expanded memory.  If there isn't enough

  memory, emx will swap to disk.  The swap file will be created in the

  directory given by the EMXTMP environment variable.  If EMXTMP is

  not set, TMP will be used.  If both variables are not set, the swap

  file will be created in the root directory of the current drive.

- If a program uses floating point instructions and no numeric

  coprocessor (387) is installed, emx tries to load the FPU (floating

  point unit) emulator emxfpemu.  That file is first sought in the

  directory from which emx.exe was loaded (this does not apply if

  emx.exe is bound to the executable), then in the current working

  directory and the directories listed in the EMXPATH and PATH

  environment variables.  If emxfpemu cannot be found, the program

  will be aborted by signal SIGFPE.

6.2 Emergency exit

------------------

If the computer freezes while running emx, try hitting Ctrl-Break five

times in one scond.  That will terminate emx (and all emx processes

run by that instance of emx) unless it is blocked in a DOS call.  If

blocked in a DOS call, emx will terminate after return from that DOS

call.

6.3 Using rsx instead of emx

----------------------------

rsx is a DPMI-compliant DOS extender which is more or less compatible

to emx.  Moreover, rsx contains a floating point emulator.  rsx is

free software and has been written by Rainer Schnittker.  As of this

writing, rsx is assumed to be available for anonymous ftp as

    ftp.uni-bielefeld.de: /pub/systems/msdos/misc/dpmigcc5.zip

    ftp.uni-stuttgart.de: /pub/systems/os2/emx-0.9c/contrib/dpmigcc5.zip

    ftp.leo.org:          /pub/comp/os/os2/gnu/emx+gcc/contrib/dpmigcc5.zip

If the emx loader emxl.exe finds a DPMI server but no VCPI server, it

tries to run rsx.exe instead of emx.exe.  The RSX environment variable

will be used instead of the EMX environment variable for locating

rsx.exe.

To load rsx.exe instead of emx.exe even if there is a DPMI server and

a VCPI server (for instance if you need the floating point emulation

of rsx), set the EMX environment variable to point to rsx.exe.  Note

that rsx needs a DPMI server.

7 emx options

=============

You can customize emx by setting the EMXOPT environment variable.

This environment variable contains a list of options, similar to

command line options.  The options must be separated by at least one

blank.  Example:

    set emxopt=-c -h40

7.1 emx options (OS/2)

----------------------

-c      Disable core dumps caused by signals and exceptions

-h#     Set OS/2 file handle limit to #.  The number # must be between

        10 and 65536

-n      Suppress popups for exceptions.  As emx.dll displays an

        appropriate error message and dumps core, popups are not

        required

-q      All command line arguments passed to child processes will be

        quoted unconditionally, that is, wildcard expansion and

        response files won't work in child processes of processes for

        which the -q option is in effect

-r*     Prepend drive letter * to absolute path names.  If a path name

        starts with / but does not start with //, /dev/ or /pipe/, *

        followed by a colon will be prepended.  If -rd has been given,

        the file name \mydir\abc will be translated to d:\mydir\abc.

        Note: this option can cause unexpected effects

-t*     Truncate file names to 8.3 format.  Each part of a pathname is

        truncated to 8.3 format by taking the first 8 characters

        before the dot and the first 3 characters after the dot.  This

        is useful to compile programs on a FAT filesystem with minimal

        changes.

        Without argument, -t enables truncation on all drives and for

        UNC pathnames.

        The -t option takes an optional argument, listing the names of

        the drives on which file names should be truncated.  The

        special drive name `/' controls whether to truncate UNC

        pathnames.  With `-tc/', for instance, file names on drive C

        and UNC pathnames will be truncated.

        If the argument of the -t option is `-', truncation is

        disabled for all drives and for UNC pathnames.  This is the

        default setting.  If the argument starts with `-', truncation

        is disabled on all drives listed.  With `-t -t-d', for

        instance, UNC pathnames and file names on all drives except

        drive D will be truncated.

-x      Don't suppress wildcard expansion and response files if the

        `MKS Korn shell' method of passing command line arguments is

        used

-E      Run debuggee in same session (window) as debugger

-I      Don't complain about programs having too small a stack size.

        Without the -I option, emx.dll displays an error message for

        those programs, telling how to run emxstack to fix the stack

        size.  With the -I option, the program is run despite of the

        problem.  However, the program will probably crash due to

        insufficient stack size.  Using the -I option is not

        recommended!

-K      Don't use DosKillThread.  Due to bugs in OS/2, DosKillThread

        may cause problems.  emx.dll automatically avoids using

        DosKillThread for OS/2 2.1 and older.  For OS/2 2.11 and

        later, you can use the -K option to disable usage of

        DosKillThread.  Currently, DosKillThread is not used by

        emx.dll and this option does not have any effect.

7.2 emx options (DOS)

---------------------

-c      Disable core dumps caused by signals and exceptions

-d      Don't use extended memory

-e      Redirect the standard error handle (2) to standard output (1)

-h#     Set DOS file handle limit to #.  The number # must be between

        10 and 65536.  This option is ignored for DOS versions earlier

        than 3.30.  This option does not change the emx limit for the

        number of files per process -- that limit is always 40

-m1     Running on Fujitsu FMR70 (not implemented yet)

-m2     Running on NEC PC-98 (not implemented yet)

-m3     Running on Intel Inboard 386/PC

-o      Send the register dump of an exception to stdout.  Without -o,

        the register dump is sent to the CON device.  You need -o for

        redirecting the register dump to a file.

        emx now creates a core dump file (named `core') instead of a

        register dump.  Please keep that file in case the author of

        the program needs it for debugging.

-q      All command line arguments passed to child processes will be

        quoted unconditionally, that is, wildcard expansion and

        response files won't work in child processes of processes for

        which the -q option is in effect

-r*     Prepend drive letter * to absolute path names.  If a path name

        starts with / but does not start with //, /dev/ or /pipe/, *

        followed by a colon will be prepended.  If -rd has been given,

        the file name \mydir\abc will be translated to d:\mydir\abc.

        Note: this option can cause unexpected effects

-t*     Truncate file names to 8.3 format.  Each part of a pathname is

        truncated to 8.3 format by taking the first 8 characters

        before the dot and the first 3 characters after the dot.  This

        is useful to compile programs on a FAT filesystem with minimal

        changes.

        Without argument, -t enables truncation on all drives and for

        UNC pathnames.

        The -t option takes an optional argument, listing the names of

        the drives on which file names should be truncated.  The

        special drive name `/' controls whether to truncate UNC

        pathnames.  With `-tc/', for instance, file names on drive C

        and UNC pathnames will be truncated.

        If the argument of the -t option is `-', truncation is

        disabled for all drives and for UNC pathnames.  This is the

        default setting.  If the argument starts with `-', truncation

        is disabled on all drives listed.  With `-t -t-d', for

        instance, UNC pathnames and file names on all drives except

        drive D will be truncated.

-E      Don't check for 387 coprocessor.  Assume no coprocessor is

        present

-O      Override XMS version check.  By default, emx checks for XMS

        version number 2.00 or later and for XMS driver revision 2.06

        or later, as older himem.sys drivers don't work correctly.

        You can override this check by giving the -O option (for

        drivers using a different revision numbering scheme), but emx

        may not work with your XMS driver, anyway.  Actually, emx has

        not been tested with himem.sys 2.05; 2.04 fails, 2.06 works

-V      Display emx version.  On program start, the emx version will

        be displayed

-X      Disable usage of XMS memory.  Usage of XMS memory with VCPI is

        a new feature of emx.  If this happens to cause trouble, use

        the -X option.

8 The emxstack tool

===================

The emxstack tool is used for displaying and fixing the stack size.

Old programs assume that emx.dll allocates a stack object if the stack

size is too small.  As emx.dll no longer does this, the stack size of

the programs must be fixed.

    emxstack <command> [<options>] <file>...

The following commands are available:

-c

    Check whether the stack size should be fixed.  For each file, if

    the stack size is too small, a message will be displayed.  If the

    stack size is big enough, no message will be displayed.

-d

    Display the stack size.  For each file, the stack size in KByte

    will be displayed.

-f

    Fix the stack size.  For each file, the stack size is fixed if it

    is too small.

-s<size>

    Set the stack size.  For each file, the stack size is set to

    <size> KByte.

-u<size>

    Update the stack size.  For each file, the stack size is set to

    <size> KByte if it less than <size> KByte.  If the stack size is

    not less than <size> KByte, the stack size won't be changed.

The following options are available:

-p

    Act on all files in the PATH.  The command is applied to each file

    in a directory listed in the PATH environment variable.  No file

    names must be given on the command line.

-q

    Be quiet.  No messages will be displayed.  However, error messages

    won't be suppressed.

-v

    Be verbose.  Display additional messages.

Example:

    emxstack -f *.exe

9 The emxupd tool

=================

The emxupd tool can replace (and delete) DLLs which are currently in

use.  (It can do so for EXE files also.)  The old version of the DLL

stays in memory and in use for programs which currently use it until

all applications which use that DLL terminate.  Then, the new DLL will

be used.  Rebooting also switches to the new DLL.

9.1 Replacing a DLL

-------------------

To replace a DLL, type

    emxupd <source_file> <target_path>

This command copies <source_file> to the directory or file

<target_path>.  Examples:

    emxupd c:\tmp\emx.dll c:\emx\dll

This command copies c:\tmp\emx.dll to c:\emx\dll\emx.dll.

    emxupd abc.dll d:\dll\new.dll

This command copies abc.dll to d:\dll\new.dll.

9.2 Deleting a DLL

------------------

To delete a DLL, type

    emxupd -d <old_file>

where <old_file> is the path name of the DLL to delete.

9.3 Unpacking DLLs from a ZIP file and installing them with emxupd

------------------------------------------------------------------

To unpack all DLLs from a ZIP file and install them in a directory

with emxupd, replacing DLLs even if they are in use, type

    emxuzdll <zip_file> <dll_dir>

where <zip_file> is the path name of the ZIP file to unpack the DLLs

from and <dll_dir> is the directory where all the DLLs of the ZIP file

should be put.  emxuzdll can be used only if there are no DLL files in

the current directory.

10 Troubleshooting

==================

This section shows how to solve common problems involving the emx

runtime.

10.1 Warning message `WARNING: emx 0.9c or later required'

----------------------------------------------------------

If this message appears, an out-of-date version of emx.dll is active.

Make sure that the latest version of emx.dll is in a directory pointed

to by the LIBPATH statement in your config.sys file.  Use

    emxrev -p c:\config.sys

to list all emx dynamic link libraries reachable by the LIBPATH

statement of c:\config.sys (use the drive letter of your boot drive

instead of c:).  To simplify things, you should keep only one version

of each DLL.  Older programs work with newer emx DLLs, but not vice

versa.

10.2 unzip.exe says `cannot create emx/dll/emx.dll'

---------------------------------------------------

This happens when you try to unpack emxrt.zip while a program is

active which uses emx.dll.  Terminate all programs which use emx.dll

and try again.  If you run emxrev, the command interpreter keeps

emx.dll loaded until terminated.  Therefore, you should close all OS/2

windows and full-screen sessions in which you have run emxrev.  All

this also applies to emxlibcm.dll etc.

10.3 emx.dll cannot be deleted, renamed or overwritten

------------------------------------------------------

See the previous problem.

10.4 Error message `SYS1804: The system cannot find the file EMX.'

------------------------------------------------------------------

You have not correctly set the LIBPATH statement of config.sys,

therefore emx.dll could not be found.

10.5 Error message `rsx not found, DPMI not supported by emx'

-------------------------------------------------------------

The current version of emx cannot use DPMI.  Therefore, the loader

tried to use rsx instead, but couldn't find rsx.  Remedy: install rsx

or end the DPMI server (MS Windows).

10.6 Error message `emx.dll: Stack size too small...'

-----------------------------------------------------

You tried to run an old program (which assumes that emx.dll creates a

stack object if the stack size is too small).

Run emxstack as indicated by the error message to fix the stack size

of the program.  If you can't (because you don't have write permission

for the file), type

    set emxopt=-I

to make emx.dll ignore the problem.  However, the program will

probably crash due to insufficient stack size.  Using the -I option is

not recommended!

10.7 DOS message `Sector not found'

-----------------------------------

DOS seems to display that message if a file is opened twice (perhaps

one handle for reading, one for writing).  For instance, this happens

when GCC runs out of temporary files.  Look for files named like

cca00002, ccz00002 etc. (the digits may vary) in your temporary

directory and delete them.

11 emx messages

===============

11.1 emx messages (OS/2)

------------------------

core dumped

    The program cannot continue due to a fatal error (exception).  An

    image of the program's memory has been written to a file named

    `core'.  Please keep that file in case the author of the program

    needs it for debugging.

emx.dll: Stack size too small...

    You tried to run an old program (which assumes that emx.dll

    creates a stack object if the stack size is too small).  You

    should run emxstack to fix the stack size.

Invalid option in EMXOPT

    An option in the EMXOPT environement variable is not understood by

    emx.

Invalid syscall function code

    The program run by emx has called an invalid system function.

    Most likely, you're using an emx.dll which is out of date.

WARNING: emx 0.9c or later required

    The program requires emx.dll 0.9c or later.  However, an older

    version of emx.dll is active.  Make sure that the latest version

    of emx.dll is in a directory pointed to by the LIBPATH statement

    in your config.sys file.  Use

        emxrev -p c:\config.sys

    to list all emx dynamic link libraries reachable by the LIBPATH

    statement of c:\config.sys (use the drive letter of your boot

    drive instead of c:).  To simplify things, you should keep only

    one version of each DLL.  Older programs work with newer emx DLLs,

    but not vice versa.

11.2 emx messages (DOS)

-----------------------

Abnormal program termination

    The program received the SIGABRT signal.  The abort() function was

    called or SIGABRT was raised.

Bad environment

    The program name could not be found in the environment segment or

    the environment segment is longer than 32768 bytes.  This should

    not happen in practice.

Cannot create swap file

    The swap file could not be created.  Remedy: make sure the swap

    file can be created in the directory given by the EMXTMP (or TMP)

    environment variable.  If EMXTMP and TMP are not set, emx will

    create the swap file in the root directory of the current drive.

Cannot enable A20

    emx tried to enable address line A20 and failed.  Remedy: contact

    the author.

Cannot enable A20 via XMS

    emx tried to enable address line A20 by calling the XMS driver.

    Unfortunately, A20 couldn't be enabled.  Remedy: unknown.

Cannot open program file

    The program file given on the emx command line could not be found

    or opened.  Remedy: type the correct file name.  (This should not

    happen with bound programs.)

Cannot read program file

    An error occured while reading the program file.  Remedy: make

    sure the program file given on the emx command line is valid (this

    does not apply to bound programs).  Maybe some other program

    accesses that file.

core dumped

    The program cannot continue due to a fatal error (exception).  An

    image of the program's memory has been written to a file named

    `core'.  Please keep that file in case the author of the program

    needs it for debugging.

DPMI not supported

    The current version of emx cannot use DPMI.  Maybe some future

    version will do.  Remedy: end the DPMI server (MS Windows) or use

    rsx.

emx not found

    The emx loader (emxl.exe) failed to find emx.exe.  Make sure that

    emx.exe is in a directory listed in the PATH environment variable.

    Alternatively, you can set the EMX environment variable to the

    complete path name of emx.exe.

Exception ...

    The program run by emx tried to do something which is considered

    illegal by the processor or by emx.  Remedy: report to the

    supplier of the program.

Illegal moucall function

    The program run by emx has called an invalid system function.

    Remedy: report to the supplier of the program.

Illegal syscall function

    The program run by emx has called an invalid system function.

    Remedy: report to the supplier of the program.

Illegal viocall function

    The program run by emx has called an invalid system function.

    Remedy: report to the supplier of the program.

Invalid arguments for DOS function xx

    The program run by emx has used invalid arguments for a system

    function.  Remedy: report to the supplier of the program.

Invalid emx option

    An option in EMXOPT or set by emxbind is not understood by emx.

    Remedy: check EMXOPT.  Use emxbind -i to display the options set

    by emxbind.

Invalid program file

    The program file given on the emx command line (or bound to emx)

    is corrupt.  Remedy: reload the program from diskette.  If this

    doesn't help you should contact the supplier of the program.

Invalid special function code: xx

    The program run by emx has called an invalid system function.

    Remedy: report to the supplier of the program.

Linear address space exhausted

    The program run by emx uses far too much memory.  Remedy: complain

    to the author.

Nonmaskable interrupt (NMI) at xxxx:yyyyyyyy

    A nonmaskable interrupt occured.  Maybe this is due to a memory

    error.  Remedy: if you are using DESQview, either set the

    protection level to 0 or use the WATCHDOG=0 option of QEMM-386.

    Otherwise, run a program for testing memory.

Out of memory

    There isn't enough memory for running emx.  Remedy: remove some

    drivers and/or resident programs.  Or add memory to your machine.

Out of stacks

    emx ran out of local stacks.  Remedy: complain to the author.

Out of memory or swap space

    emx cannot write to the swap file because the disk is full.

    Remedy: make sure there is enough space on the drive used for the

    swap file.  Use the EMXTMP (or TMP) environment variable to tell

    emx where to create the swap file.  If neither EMXTMP nor TMP is

    set, emx will create the swap file in the root directory of the

    current drive.

Out of swap space

    emx cannot write to the swap file because the disk is full.

    Remedy: make sure there is enough space on the drive used for the

    swap file.  Use the EMXTMP (or TMP) environment variable to tell

    emx where to create the swap file.  If neither EMXTMP nor TMP is

    set, emx will create the swap file in the root directory of the

    current drive.

Process terminated by SIGINT

    You have stopped the program by hitting Ctrl-C or Ctrl-Break.

Process terminated by SIG...

    The program has been stopped due to a signal or exception.

    Usually, this indicates a fatal error in the program.

Program aborted

    You have aborted the program by answering `Abort' after a critical

    error had occured.

rsx not found, DPMI not supported by emx

    The current version of emx cannot use DPMI.  Therefore, the loader

    tried to use rsx instead, but couldn't find rsx.  Remedy: install

    rsx or end the DPMI server (MS Windows).

Swap file I/O error

    emx couldn't read or write the swap file.

This program does not run in DOS mode of OS/2

    DOS emulation of OS/2 doesn't support emx.  Remedy: boot DOS or

    run the program in a OS/2 window

This program requires an 80386 CPU

    You need an 80386DX, 80386SX, i486 or later CPU for running emx.

    Remedy: get another computer.

This program requires DOS 3.0 or later

    If emx is bound with an application program into one executable,

    DOS 3.0 or later is required for execution.

Too many processes

    emx supports up to 4 processes.  You tried to start 5 or more

    processes.

Unexpected interrupt

    This cannot happen.  Remedy: report to the author.

Unsupported VDISK.SYS version

    emx doesn't support the vdisk.sys device driver you are using.

    Remedy: remove VDISK.SYS or use a different version.  emx should

    work with vdisk.sys 3.3 or later.

Unsupported XMS version

    emx requires himem.sys 2.06 or later (older versions are too

    buggy).  Remedy: get a newer himem.sys.  If you are not using

    himem.sys, use the -O option of emx and tell the author whether

    this works or not.

Unusable interrupt vector mappings set by VCPI server

    The configuration of the VCPI server (EMS driver) is illegal.

    Remedy: read the manual of the EMS driver.

Use emxbind

    You tried to run emxl.exe, which doesn't make sense.  emxl.exe is

    designed to be bound into an executable file by emxbind.

Virtual mode not supported without VCPI

    A VCPI server is required to run emx if the CPU is in virtual

    mode.  You will get this message if you're using an EMS driver

    (EMS emulator) which doesn't support VCPI or if you have disabled

    EMS.  Remedy: remove the EMS driver, use an EMS driver which

    supports VCPI or enable EMS if it has been disabled (for instance,

    remove the NOEMS keyword from the command line of the EMS driver).

Wrong emx version

    The version of emx.exe does not match the version the program was

    compiled for.  Remedy: use the copy of emx.exe that comes with the

    program.  If you are using different programs which need different

    versions of emx.exe, use the EMX environment variable to set the

    path name of the correct version of emx.exe before running a

    program which requires a version different from the default

    emx.exe.  You can also use emxbind -u to put the correct version

    of emx.exe into the executable files.

12 Obtaining the complete emx package

=====================================

emx is available for anonymous ftp on

    ftp.uni-stuttgart.de: /pub/systems/os2/emx-0.9c

    ftp-os2.cdrom.com:    /pub/os2/lang/emx09c

    ftp-os2.nmsu.edu:     /os2/unix/emx09c

    src.doc.ic.ac.uk:     /pub/packages/os2/32bit/unix/emx09c

    ftp.leo.org:          /pub/comp/os/os2/gnu/emx+gcc

13 The author of emx

====================

The author of emx is

    Eberhard Mattes

    Teckstrasse 81                                      (TeX: Teckstra\ss e)

    D-71696 Moeglingen                                  (TeX: M\"oglingen)

    Germany

Internet: mattes@azu.informatik.uni-stuttgart.de  (subject to change)

No telephone calls please!  It must be stressed that the author does

not reply to letters if you don't include return postage

(international postal reply coupon if you are outside Germany) and a

self-addressed envelope.

--------------------------- END OF EMXRT.DOC --------------------------------

Problem Description:                                                          

 The NET STOP PEER traps in netwksta.200 in the GETTCB macro on a

 MCP machine with a SMP OS2KRNL of ACP machine. This was happening

 because we use CurrTCB variable of the OS2KRNL for getting the

 tcb information, while this variable is not set in a SMP kernel.

 And hence we ended up using the CurrTCB with value 0 on the SMP

 kernel and we trapped.

Action Taken:       

 TCB macros modified.                 

NET STOP PEER TRAPS IN NETWKSTA.200 ON A MCP MACHINE WITH OS2KRNL OF ACP.

There are 2 variations of the fix, smp8603.zip for an eCS GA running the

original version of Peer, IP08603, and ic31990.zip for an eCS GA that has

been updated to the MCP 2 level of Peer, IP08605.  The fix has only been

tested with the MCP 2 level of Peer.

To install the appropriate fix, back up the files to be replaced, unzip the

fix and copy the files as follows:

netapi.dll \muglib\dll\netapi.dll

netwksta.en \ibmlan\netprog\netwksta.200 (english)

netwksta.de \ibmlan\netprog\netwksta.200 (german)

peer.exe \ibmlan\services\peer.exe