DOS/4G error (2001): exception 0Bh (segment not present)

[modelo64]

Hello.

My name is Enrique and I am from Argentina.

In the company where I work it was decided to replace all Windows 8 by Windows 10 due to the end of service of Windows 8.

In my sector we have computers with Windows 8 - 32 bits where we have running a system developed in RM-Cobol-85 version 5.36 for DOS.

That is why I am trying to make the migration to Windows 10 - 32 bits.

I was evaluating different emulator options (DOSBox, DOSBox-X, vDos) and due to the way it is implemented and the CMD calls that the system has, the only one that allows me those calls is vDos.

But I have a serious problem and it is related to the management of virtual windows inside RM-Cobol.

As I said, the system is currently working. And in the tests I did in DOSBox and DOSBox-X it also works.

But unfortunately in vDos it gives me an error when trying to "close" those virtual windows and abruptly terminates the program.

I attach the error screen and the log file.

I sincerely hope you can give me some help.

Thank you very much.

Best regards.

Attachments:

vDos.log (3.55 KB)

[Jos]

You would migrate to Windows 64-bit, 32-bit still has NTVDM to run DOS apps.

DOS extenders like DOS/4G are complex beasts. The log file doesn’t report relevant information, the screenshot of the DOS/4G error isn’t really helpful.

I can however reconstruct the error with another RM-Cobol-85 program. That seemingly worked fine up to vDos 2016: There’s still a saved report printed by the program.
As of vDos 2016, a similar error is thrown by DOS/4G at just quitting the login. I don’t have documentation anymore how to start the program (login).

Will have a look at what causes the exception 0Bh by DOS/4G…

Jos

[modelo64]

Jos, thank you for your reply.

From what I understand, your suggestion is to try to run the system developed in RM-Cobol on a 64 bit Windows. Isn't it?

Do you think it could work in that environment? Because if so I might give it a try. I really need to find a solution as soon as possible.

Anyway, hopefully you can figure out what causes the 0Bh DOS/4G exception.

I look forward to your comments.

Thank you very much.

model64

[Jos]

For vDos it doesn’t matter running in Windows 32 or 64-bit. RM-Cobol then runs on the same emulated CPU and DOS environment.
RM-Cobol won’t run ‘directly’ on Windows 64-bit because that has no NTVDM. Windows 32-bit however does.

I’ll look into the RM-Cobol program at hand for the DOS/4G exception. Curiously it seemingly once ran w/o problems up to vDos 2016.

Jos

[modelo64]

Then I can only hope that you can figure out what causes the exception.

I think I have no way to continue with the migration until I have a possible solution. At least with the current RM-Cobol system.

Obviously one option is to modify the programs by not using the virtual windows, but that would be practically reprogramming 50% of the system... and at this moment it would not be possible.

The other option is to decline the possibility of running the system on 32 / 64 bit Windows (10 or higher)... which would make me look very bad to the leadership of the company.

But as I always say, "it's what it is.... quite a little bit by the way!" ¯\_(ツ)_/¯

Hopefully you can find a solution to be able to use the RM-Cobol system in vDos!

Regards!

modelo64

[Jos]

Could be I found the RM-Cobol mishap.

COULD, I can’t get pass the login (don’t have the password) of the tested program. The exception isn’t thrown anymore quitting the program.

I’ll send you a modified vDos.exe by wetransfer.com. Mind, it is my development version, not fully tested by users. So be careful with that.

Jos

[modelo64]

Jos:

First of all, thank you very much for taking the time to analyze the problem and try to solve it.

I received the modified version of the program and I can tell you that it WORKED!

I have to keep testing in more detail, but at least the problem of the virtual windows seems to be solved.

I will keep you updated. Any news or problem (I hope not) I will post it. Also the confirmation (I hope so) that finally everything is working fine.

Again thank you very much for your knowledge and time!

Regards!

> modelo64

[modelo64]

Topic status update.

Hi all.

I have tested the whole RM-Cobol system implemented in '90, and with the new version of vDos (2023.xx.xx), I can confirm that it works perfect!

Once Jos gave me the new version, the problem with RM-Cobol virtual windows was solved.

The next step was to test with the new version of vDos all the programs and routines that I had originally compiled in Windows 8 (32 bits).

Result: they worked fine.

Then, to check that the vDos environment would allow me to work with RM-Cobol without problems, I recompiled all the programs and routines in the new vDos environment.

The result was also that everything worked perfectly!

I can say that this new version of vDos (2023.xx.xx) is fully compatible with RM-Cobol version 5.xx for DOS.

Now I still have to solve some "foreign" issues to the RM-Cobol programming, but that are part of the implemented system.

Keep in mind that the system began to be developed in 1990, so it uses several resources or "tricks" to be able to do certain things.

All of them are done through cmd calls to run certain processes (printing, copying files, generating PDF files, etc.).

Among those things to fix, are the calls to RM-Cobol programs that must be run from the Windows task scheduler.

But this I will be exposing in a new topic.


Jos:
I hope this improvement will be included in the next public release.
Again, thank you very much for your knowledge and your dedication to solve the problem.

> modelo64