VOICE Home Page: http://www.os2voice.org
[Newsletter Index]
[Previous Page] [Next Page]
[Features Index]

July 1999

OS/2 Tips

We scan the Web, Usenet and the OS/2 mail lists looking for these gems. Have you run across an interesting bit of information about OS/2 recently? Please share it with all our readers. Send your tips to editor@os2voice.org

Editor's note: these tips are from OS/2 users and in some cases can not be verified by myself. Please heed this as a warning that if you are not sure about something, don't do it.

June 15, 1999 - Funny how all of a sudden one sees the same subject come out in various places at the same time. For a while this month, everywhere I looked I saw complaints about extra drive objects. Here is one answer on comp.os.os2.setup.storage from our old friend Sam Detweiler:

But you HAVE a drive B, even when you don't THINK you do..
Back when there was a single drive system, imaging copying a file from one diskette to another...
So, when you only have ONE drive, since 1981 ALL systems have an A: and a LOGICAL B: drive. copy a:*.* B: reads files from one diskette, prompts to change diskettes and then copies to the new diskette..

June 17, 1999 -On comp.os.os2.networking.tcp-ip, Lorne Sunley offered this tip in response to a request for a network that copy's directly from one drive to another without passing the clients memory.

There is a command NET COPY That does exactly what you want. It is documented in the Warp command reference, under the LAN commands.

June 17, 1999 - Brandon S. Allbery posted the logic for calculating a leap year on the POSSI Mail List:

The rule is:

years divisible by 4 are leap years, except
years divisible by 100 are not leap years, except
years divisible by 400 are leap years, except
years divisible by 1600 are not leap years.

2000 % 400 = 0, 2000 % 1600 = 400; so 2000 is a leap year.

You're hearing one of the things that's likely to be a common Y2K failure mode.

June 17, 1999 - Again on the POSSI list, this time Ronald A. Boschelli asked how to remove a file named LPT1 from his system. Randy Petersen had the following successful suggestion:

Have you tried temporarily remming out the BASEDEV=PRINT01.SYS statement in config.sys, rebooting and then trying to delete the file?

June 22, 1999 - Remeber the issue of the extra drive objects? Here is another possible explanation for too many drive objectrs if you have a Thinkpad. A tip of the hat to Bill Fullam on the TeamOS/2 Mail List:

What you are experienceing are the "drives" installed by the PC Card Director software. Look for the ATA PCMCIA driver entry in your config.sys. You may rem out that line as it adds support for ATA devices such as harddrives. You may have to play with the entries to rem out the correct lines for your setup.

YMMV, so backup everything before you start playing. The drives you see are "ghosts" but can cause problems if you are also logging into a LAN. My current network card (IBM 10/100 CardBUS) requires I suppress the PC Card support on my 770x, so I have remmed out all of my PCMCIA entries. The temp 3COM card I once used exhibited this behaviour.

In conclusion, you have not done anything wrong. In fact you have done it all right. You, and most people just do not need this "feature".

June 25, 1999 - On TeamOS/2 Help List Rob Burton offers this tip for people installing Star Office 5.1 and who don't want SO to install itself as the default for all HTML access:

Testing here I found that the issue Vic raises about association of URL and HTM/L objects with SO5.1 arises if you accept SO5.1's offer to make itself your default Internet application. I found that if I uninstalled SO5.1 and re-install NC/2 4.04, the proper association is re-established for the url objects in the web site folders of Connections. Upon reinstallation of SO5.1 the association with Communicator was retained. However, I find that there is an apparently hidden association to HTM objects. Maybe this is a job for Keldar's Association Editor.

June 26, 1999 - Guy Robinson had this advise for LS-120 users on the Aurora Beta Mail List:

Here is my scenerio an what solved the problem for me. I have an aptiva 2137-e26 and had installed the LS-120 and downloaded the drivers and readme files from the IBM web site. There are (I think) four statements that have to be put in/change in config.sys. Using the original 'A' drive and these files the Aptiva under Warp 4 recognized the floppy part of the drive as drive B: . Later I decided I wanted to install a two floppy ( 1.44m and 1.2m) drive unit in the system. I moved the LS-120 to where the original floppy went and then installed the dual floppy under the CD rom drive bay. I went into setup and told the system 'A' is a 1.44m and 'B' is a 1.2m drive. At this point the system wanted to recognize BOTH the LS-120 and the 1. 2m drive as drive 'B' (what complicated matters more was the vendor who sold me the dual floppy had pinned it out 180 degrees opposite of normal - good thing my other computer had one for comparison in it!)

So what is the solution? The readme file gave me the switch I needed:
Modify the statement - BASDEV=OS2DASD.DMD to read:
BASDEV=OS2DASD.DMD /MP:*,1 It works! The drive will then be assigned a higher letter - in my case the drive letter is next to the one for the cdrom drive. Using the new drive letter you will still be able to use the floppy in the LS-120.

June 26, 1999 - Brandon S. Allbery offered this REXX script on the POSSI Mail List for those times when Desktop goes missing:

Try this:

boot to command line
type SET DESKTOP=C:\Desktop <- note the drive letter!

If your desktop comes up, then the WP_DESKTOP entry in OS2.INI is missing or
damaged. This REXX script should restore it:

call SysLoadFuncs
call SysSetObjectData 'C:\DESKTOP', 'OBJECTID=<WP_DESKTOP>'

(Replace "C:" with your boot drive if necessary.)

If that doesn't do it, then possibly the WPS is hung due to garbage in
OS2.INI or OS2SYS.INI. Try renaming them and running MAKEINI to create new

If that doesn't get you a functional (if minimal) desktop, then I'm stumped.

June 26, 1999 - On comp.os.os2.networking.tcp-ip James Knott offered this assistance for some poor soul looking for help with installing NETBIOS over TCP/IP:

The TCP/IP adapter number does not affect the NetBIOS adapters. Also the NetBIOS adapter number should be lower NetBIOS/IP adapter number. This is so that NetBIOS will be tried first. Also, you don't often need both NetBIOS & NetBIOS/IP on a small network. Choose one and stick with it on all systems. NetBIOS (NetBEUI in the Micros~1 world) is faster and easier to set up than NetBIOS/IP. On my work computer I have NetBIOS, TCP/IP and 802.2 (SNA) all on adapter 0 and NetBIOS/IP on 1. The main reason for NetBIOS/IP, is to get past routers. NetBIOS is non-routable. If you use NetBIOS/IP, you'll need an RFCNAMES.LST file (similar to the Windows LMHOSTS file) to map computer names to IP addresses. The reason for the different adapter numbers for NetBIOS and NetBIOS/IP, is that you can't have two instances of the same protocol type with the same number.

June 27, 1999 - On TeamOS/2 Help List Steven Levine had this tip for testing the integrity of zip files:

Both unzip and pkzip have the ability to test zip files for integrity.
Commands like:

unzip -t javaintk.exe
pkzip /test javaintk.exe
are your friends. Also, a simple:
javaintk /h
will tell you
javaintk /test
will get you the same results.

June 30, 1999 - Are you as confused as I am about all the variations on SCSI? Patric Bechtel had this rather clear explanation on the Aurora Beta Mail List:

First: there is the wire (cabling). Here are three factors which are important:

- signalling type: differential, single-ended, lvd (low-voltage differential)
- signalling frequency: async, sync5, fast10, fast20 (commonly ultra), fast40 (commonly ultra2), fast80 (coming soon)
- signalling width: 8 (50pin), 16 (68pin) or (VERY seldomly used) 32 (80pin) bit.

Second: protocol standard
- SCSI-1: the very first one. Only usable (ok, the ARE exceptions) for hard drives or tapes. Doesn´t know SCSI_INQUIRE and some commands commonly needed by modern devices. Knows of signalling freq async and sync5.
- SCSI-2: second version of the protocol. Was needed when more devices were found to need SCSI. Devices have classes and subclasses, removables are possible, signalling freq up to fast20, width up to 32, signal types SE and D.
- SCSI-3: relative new, needed for even more standardisation. AFAIK declares all known signalling types, freqs and width's.

Something special are SCSI-Bridges (First seen on the Adaptec U2W): looks for the controllerchip as if it is the same bus; whereas the bridge allows it to connect LVD-devs on the one side, on the other side standard SE-components. Very special. I prefer two controllers: One doing lvd-things and one for all SE´s; because on LVD-Bus, the command processing is made on LVD-mode too; whereas on standard SE commands always are interchanged in async mode (which increases command overhead and latencies).

July 1, 1999 - Here is Deepak Datta's rather radical cure to fix printing problems when a printer is not fully deleted before installing a new one which can cause problems for some apps like PMMail:

Unimaint is the right thing to use. If you haven't used it before, please be very careful!

Start Unimaint > load user ini file > look for application PMMail in the list on the left > Delete application > close Unimaint

When you next start PMMail, it would have forgotten all about itself in re OS/2, i.e. its position on the desktop, which printer it prints to etc. It will look up the available resources and attach to the correct printer.

You might also need to fill in the registration info.

This worked for me when I had removed all printers and re-installed my printer. PMMail, Galleria and CopyShop, could not print any more. After this procedure all found the correct printer and printed correctly.

Your problem might be completely different, but if you follow this correctly you won't be causing any harm.

July 1, 1999 - In response to a question about errors in DOIP: "error : [NETW] Invalid fcs", Brandon S. Allbery says they are:

"Invalid frame checksum" --- line noise or etc.

July 3, 1999 - On comp.os.os2.beta, Fred Springfield offered this suggestion for chirping systems if you just installed the DANIS506 busmastering IDE driver:

While beta testing for Dani, long before the Gamma releases, I had a "chirping" problem with one of my 2 systems which I was using. It was a full tower system, and I was able to eliminate the chirping by putting in a new and shorter cable (24" vs the old 30"). So-- if you have chirping, you can reduce the PIO level as you did, or you can also take a look at your hardware, especially the cables. This might not work for everyone, or even all of the time, but it's a place to start.

July 4, 1999 - On the NS4OS2 List, Peter Skye offers this insight into the differences between Java and javascript:

The Java code which I write doesn't look anything like the JavaScript code I write. JavaScript is embedded in the browser, while Java is external and not part of the browser. JavaScript is not executed by the Java Run Time engine.

Thus, JavaScript will work even if you haven't installed Java.

But the JavaScript code can call Java Applets, and the Java Applets use the Java Run Time engine.

You can write your HTML code and include JavaScript, and you can then have the JavaScript call Java Applets.

Thus, if you turn off JavaScript, the Java Applets embedded in the JavaScript will not be called and the Java Run Time won't be executed. Thus, no failure.

However, the HTML programmer can call a Java Applet directly without using JavaScript by using the APPLET tag, hence you can _also_ execute the Java Run Time engine without having any JavaScript at all in your HTML. And these direct calls via APPLET to the Java Run Time engine aren't affected when you turn off JavaScript in your browser.

So you're both right. :)

To summarize: JavaScript will work even if you haven't installed Java. Java Applets will run even if you've turned off JavaScript. And JavaScript can run Java Applets.

Hope this helps. It _is_ confusing, to be sure.

July 6, 1999 - From the OS/2 List here is David Eckard's explanation of what may be in your HOSTS file if you have TCP/IP installed:

The first Item is the IP address. the second item is the host name. The third item is an alias. The last item I believe needs the # sign is a comment. for example... orion orion #main computer nebula nebula #bedroom computer hobbes.nmsu.edu hobbes ##where os2 files are

I have cut most of the space out. I just use the same name for the alias and for the hostsname on my two computers The third example would make hobbes the same as hobbes.nmsu.edu and I could ping both equally well. Check out tcpcfg.exe (i usually type it from the command line as it takes too long to find it otherwise). This brings up the settings. One page is hostnames. The second page of that is where you can put shortcuts and the other machines on your network. Another cool thing described in OS/2 Ezine that you can do is to put something like this in. www.adnet.com

I haven't tried this yet but OS/2 Ezine list about 10 addresses where those banner ads come from and this will cause it to lookup local host first, thus, never finding the things.

Puting your popular sites in it will eliminate one step from your browsers work and speed up access to that site.

July 4, 1999 - Yet another helpful IBMer, Irv Spalten, offers this pointer on the comp.os.os2.apps news group for people having problems creating Utility Disks:

It has NOT been fixed. The 'real' reason is that the Adaptec drivers grew SIGNIFICANTLY and just will NOT fit with all
the other dd's on a diskette. Some people 'die' on diskette 1, others on 2 or 3. Depends on the individual h/w you have. We do have manual instructions to build them though, please see URL

July 6, 1999 - On comp.os.os2.games, Brad Barclay, YAI (Yet Another IBMer) had this tip for people having problems with Links/2 after changing Video resolution:

Back when Links/2 was released, it had a problem that I wrote a quick patch for (IIRC, if your default window settings had the top view minimized, it would cause the game to crash after trying to start a new round). The patch doesn't have anything to do with your problem, but during its development I learned a bit about how certain window size and position information is saved.

What you should probably try to do is get an INI editor and edit your OS2.INI file. In it there are several applications setup for storing Links/2 child window size/position information. Delete all of these, and start up Links/2. Hopefully this should cause Links/2 to revert to it's defaults for your resolution. I won't guarantee that this will work - but it certainly can't hurt to try :)

July 7, 1999 - Ever wonder why you have to install a bootable partition entirely before the end of the 1024th cylinder? Jon Saxton had this explanation on the TeamOS/2 Help List:

When you start your computer the primary bootstrap code in the ROM BIOS gets executed. The processor is in real mode with a 1Mb memory address space limit. Disk accesses are executed via software interrupts (int 13h). A read request, for example, requires that a request block be filled. The request block 24 bits wide and specifies the cylinder, head and sector number to be read. There are 10 bits for the cylinder, 6 bits for the head and 8 bits for the sector so the cylinder number can range from 0 to 1023, the head from 0 to 63 and the sector from 1 to 255 (sector number 0 is not used).

All of those numbers are, in a sense, fake. Modern disk controllers do not need or use cylinder/head/sector specifications to access the disk. All they need is a sector number (ranging from 0 to the end of the disk) and the controller takes care of the rest. The controller provides a reverse CHS mapping simply to satisfy the ancient code in the typical PC BIOS. All LBA does is to provide an alternative reverse mapping which, in the extreme case, extends the addressability of the disk to about 8+ gigabytes. This limit is reached when all 24 bits of the request block are filled with ones.

The ROM BIOS cannot address a larger disk using int 13h calls and since the ROM BIOS code is what loads your operating system it follows that your operating system must reside within the first 8+ Gb.

Once your operating system is loaded it can use whatever method it wants to access the disk so the entire disk can be accessible.

I believe we'll see some fixes to this entire issue soon. One option is for the BIOS to detect intelligent disk controllers and in the presence thereof access the disk directly rather than via the int 13h mechanism. I believe that 3rd-party add-on software such as OnTrack disk manager works that way.

July 12, 1999 - This months tips are filled with help from IBMer's. So no surprise that the last tip of the month comes from someone at Big Blue as well. On news://ncc.hursley.ibm.com/ibmpub.java.os2, Chris Tomlin of the IBM Java Technology Centre posted this pointer to a schedule of expected release dates for JVM ports including 1.2 (OS/2 preview of 1.2 expected in 4th Qtr 1999, with the GA in Early 2000).

Availability dates can be found at: http://www.ibm.com/java/jdk/other/portingplans.html

[Previous Page ] [ Index] [Next Page
VOICE Home Page: http://www.os2voice.org