Here is the latest FAQ on FixPack 16 for OS/2.

A couple of things seem very important to have success with 
this FP. First, and this may seem obvious, is to READ THE 
DOCS. This FP has some subtle little quirks and the install 
notes can be a big help. I got burned on the first try as well as a 
lot of other people. The install from FP10 leaves some files lying 
about that will cause FP16 to fail to install. Excerpt from the IBM 
FixPack 16 install documentation:

'... check for the following files on your
  OS/2 boot drive in the \OS2\INSTALL directory and in your
  MMOS2\INSTALL directory wherever it is installed.
 
     LOGF0000.OS2
     LOGSTART.OS2
     LOGARCH.OS2
  
  Also check for the following file:
 
     C:\CSF_SEL.000
  
  If any of them exist then rename them and try the FixPak install
  again.  If it works then delete the renamed files....'


I had these on my system and they played havoc...


The second thing, and this seems especially relevant if you use 
MMOS/2, is to use the very latest CSD Boot Disks. These are 
available at your local FTP site as wkickr.zip.


Boot Trouble

If you have installed FP16 but cannot get past loading the shell 
and you have PMQSize in your config.sys you need to remove 
it. The new code from IBM and the patch implemented by PMQ 
doi not like each other so rem it out.


Hangs

This one is a " your milage may vary item". I had WatchCat 
installed on my machine and I had nothing but hangs until I 
removed WC from my config.sys. Many others have reported 
this too. However, a number of other users report no conflict 
with WC on their systems. It seems that newer machines are 
the ones that get around this problem. I don't know if it related 
to the specific BIOS or what. If anyone can shed light on this 
please send me e-mail.

IMPORTANT

I have read several posts and some e-mail from people that 
report trashed desktops/ini files from crashes. I was fortunate 
that this did not happen to me even though I had about a dozen 
reboots initially. Several wise souls recommend, and I agree, 
that you should backup your desktop BEFORE you tinker with 
this FP. I made a point of booting from utility disks after every 
crash and running CHKDSK [bootdrive] /F:2  repeatedly until I 
had a clean report. That was time-consuming but I didn't have 
any ini file troubles as a result.<g>


Video

Lots of people who use ATI cards report corruption and a host 
of other ills ranging from minor to major. I have seen some 
recommendations to re-install the video drivers after you install 
FP16. For some posters this is all that was needed, however, 
others had no luck after doing so. Again, your mileage may vary.
late note...This may work for errant Cirrus cards as well


I don't use an ATI card, in fact mine isn't even supported, but I 
have had no trouble in this department. Other than a couple of 
mentions about a Cirrus card and some S3 drivers most trouble 
seems to come from ATI.

If anyone gets there video working properly after having trouble 
please send the details and I will include it here< with credit of 
course...>

Again I am including a copy of a post made Dec22 regarding 
IBM working on a late bug fix for this FP. I am trying to find out 
more details but as of yet I can add nothing more...

<begin quoted post>

From: jkess@watson.ibm.com
Newsgroups: alt.org.team-os2
Subject: Fixpack 16 updated w/ a quick bugfix
Date: 22 Dec 1995 22:26:50 GMT
Message-ID: <4bfbba$fi0@watnews2.watson.ibm.com>

A problem in SEAMLESS.DLL has been found. Affects focus for 
seamless Windows apps. Fixpak is being re-built with a new 
SEAMLESS.DLL and will be uploaded soon.
For those who have already downloaded the fixpak, a 
PJ21367.ZIP file 
will be made available along with the updated fixpak. This ZIP 
will 
contain only the new version of SEAMLESS.DLL and 
instructions for 
replacing it.

The old version of the file should be on disk 5, and is 
recognizable as:
   12-04-95  11:37p     42624          49  SEAMLESS.DLL


(Josh White Jr., on being asked why folk musicians spend so 
much time 
tuning their
instruments: "We tune because we _care_.")
 < end quoted post>

Single Input Queue

It appears that a fix for the Single Input Queue has been 
effected. A number of people with the tools to measure the 
Queue have posted that indeed the Qsize is now at 3000. If you 
install FP16 and can't get the shell to load you should remove 
PMQSize from config.sys as it is not needed and conficts with 
the changed code. { as mentioned above }


A number of people have also reported that when an app fails 
to service the Q that there is a color change in the offending 
application and it has the focus removed therefore you get 
fewer hangs. Using ctrl+esc seems to be generally more 
effective but as always some patience is needed for it to work. 
This is another " your mileage may vary" item.


Development Products


VX-Rexx code broken by FP10 has been fixed but I am now 
told that there is another VX-Rexx code break. This one comes 
from : ve6yj@compusmart.ab.ca

< Begin quoted post>

Actually, I would humbly suggest that the VX-REXX issue is still 
current, but for different reasons.  Here are my findings.

Of course, the 'nasty' was that entry field that was only 1/16" 
high.  When I was at f/p 10, I simply regressed PMMERGE.DLL 
to f/p 5 and I was fine.

With f/p 16 I appear to get the WIDTH of VX-REXX dialog 
boxes in modal windows cropped.  For example, if you save a 
project that has not changed, you get a modal window with a "?" 
picture, text and two buttons.  The text says "Project has not 
change.  Save anyways?"  The buttons beneath are OK and 
CANCEL.  The text was fairly long, so the box was moderately 
wide.  WITH FIXPACK 16, the box is only as wide as required 
to support the two buttons side by side, and the text is cropped 
to say "Project has changed."  AND THAT'S IT!  So it appears 
we have a width problem now.

At f/p 16, I cannot regress PMMERGE.DLL to f/p 5; the system 
will crash during boot-up.

< end quote >

If anyone comes up with a solution please forward it me for 
inclusion here.

There is a serious FP16 bug in VAC++ as well. From what I 
gather if you so much as move the mouse over the action bar 
buttons you risk a hang or crash. Very nasty! 

If you get a solution please send it on yada yada yada...


SoundBlaster


Here is some advice from Troels:

<begin quoted mail>

I had a problem with the Soundblaster 16 driver; the one 
included in FP16 is older than the latest drivers from Creative 
Labs. The new drivers from Creative use another syntax
than the driver included in FP 16. Solution: Reinstallation of the 
latest SB16 drivers after FP16.
<edited mail>
Only problem with FP16 on my system: FleetStreet (Fidonet 
message editor) behaves strangely;the FS-windows look 
different right after starting the program, but they return to 
normal appearance as soon as I start using the program.
<end quoted mail>


ShareWare

This fixpack has caused some changes in the behavior of 
Shareware apps. If something you use has changed, better or 
worse, send me the details in a short note for inclusion here. 
Please tell about your system too.

Observations

I have noticed that UltiMail is quicker to load and my system is   
generally more responsive. I gather that it is those of us who run
Warp on pokey older systems who will see the improvement...

Some posters have indicated a change in some dialog boxes 
and pallette colour changes as well. The Dialog boxes are likely 
by design but if you have a tip on how to fix the colour changes 
let me know.

Others have noticed an improvement in Win/OS2 in the speed 
department. 



I have left out some tings in this file due to  time restrictions but 
I will try to keep things updated as best I can. 

I'll mention it once again, if you come up with a fix for a problem 
please send it it along to me for inclusion here. Include a 
description of your hardware, etc..

Here is a tip that I have used successfully:

Place this line in your config.sys

SUPPRESSPOPUPS=E:\
 where E is a drive of your choice.

when you get a crash or hang OS2 will write the details to 
popuplog.os2

in the details you will see something like this: 
CS:EIP=005b:17864fdf

you can identify the ring location of the crashed code from the 
CS portion.
The following applies:
last CS digit is: 0.4.8,or C and EIP <>7xxxxxxx
the trap was in the kernel.

last CS digit is: 0.4.8,or C and EIP =7xxxxxxx
the trap was in a virtual device.

last CS digit is: 1,5,9, or D
the trap was in ring one.

last CS digit is: 2,6,A, or E
the trap was in ring two

last CS digit is: 3,7,B, or F
the trap was in ring three.

Ring three is where application code runs and a crash in ring 
two indicates so form of I/O to a device was taking place.

Thanks to Dean Roddey for the ring two advice.


This may give you an idea of what happened when you went 
belly up. <g>


I hope to have more on FP16 for you before too long.

Send mail to tjerick@ibm.net

Here's to a good new year!

Tim Erickson

