Keycode error

[miland]

Hi, thank you very much for your quality work. vDos is a great program.
Unfortunately I have a problem with the keycode in config.txt.
My old DOS application works in CP 895, which is not supported in Czech Windows with CP 852. The display is without problems thanks to the CP 895 font in config.txt. Keyboard support programs in vDOS do not work. When using the keycode, the enclosed "Option is already set" error is displayed. Is there any way to fix this error and recode the keyboard using keycode? Normal key recoding works, only recoding of special "Czech" keys reports this error.

Attachments:

[Jos]

(DOS) keyboard support programs indeed don’t work in vDos. It relies on Windows settings for language related items like keyboard layout and character display. Those would of course already be setup correctly.

The KEYCODE option is mostly provided to correct a Windows default comma of the NumPad decimal point (DOS expects) for some languages. KEYCODE can only be used once (for one specific key combination). The error message just says you used it more than once.

Your (second) KEYCODE = 0231=0228 would substitute the “1” key of the upper row to a “(“. If code page 895 opposed to 852 requires another than the default keyboard layout, you should install (in Windows) an appropriate keyboard layout for code page 895 and select that running vDos.

Jos

[miland]

Thank you very much for your answer. You're right, if I use the keycode once, I can recode any key. Unfortunately, this cannot be used for my problem when I need to recode the whole keyboard.
The problem of the CP 895 is not in the layout of the keys, it is the same as in the CP 852, but the same special "Czech" key has a different ASCII code in the CP 895 than in the CP 852.
Maybe I'll find another solution so I can use vDos. So far, I'm using an older vDosplus in which I get a "page fault" again.

[Jos]

You lost me, if CP 895 is that different in keyboard layout (Czech special), there should one for that in Windows?

KEYCODE was never intended to remap the whole keyboard layout.
I still expect an appropriate Windows keyboard layout for your situation, you just have to find it?

Eventually use the On Screen Keyboard to verify the key assignments.

Jos

[miland]

In "Czech" Windows special "Czech" keys return ASCII key codes in CP 852, I would need to translate them into CP 895. The layout of the keys is the same, there is only one "Czech" keyboard. This is approximately 40 characters.
In vDosPlus I solved it with a special program for the keyboard, of course it would be more elegant to use a standard keyboard from Windows recoded from CP 852 to CP 895.

[miland]

For completeness, I enclose a translation table from CP895 to CP852 and back.

Attachments:

895-852.txt (5.94 KB)

[Jos]

One would expect a Windows keyboard driver that handles your CP 895, but probably that was found too obsolete, DOS specific, MS omitted it?

Since a DOS keyboard redefinition utility doesn’t work in vDos, you would have to use a Windows one.
Did you try the MS Keyboard Layout Creator (https://www.microsoft.com/en-us/download/details.aspx?id=102134)?

Jos

[Jos]

EDIT:

If the Czech keyboard already returns the correct Unicodes for CP 852, the mentioned MS Keyboard Layout Creator of course won’t help. You would need to set CP 895 by the CHCP command in vDos, but Windows doesn’t provide that codepage. Perhaps the corresponding NLS file is somewhere on the Internet?

Jos

[miland]

Thank you very much for the answer, I did not find a language pack for Windows 10 with CP 895 on the Internet. I would like to find a solution without interfering with Windows.
My programs are in DOS Foxpro 2.6b, I will try to make a PLB library for transcoding keyboard characters from CP852 to CP895.
Otherwise, if possible, I would like to pay for this feature in vDOs. In config.txt there could be a string of 256 characters, from which the resulting character would be selected by the ascii code of the pressed key.
In addition, after resolving this issue, I purchase a network license immediately :-).

[Jos]

You’re in luck: I discussed some weeks ago with Edward Mendelson a solution for DOS code pages that have no equivalent one in Windows. That didn’t lead to anything since the projected user would also need a RTL display of characters.
I had some spare time since I stopped major modifications to vDos shortly for its release on May 1st. So I came with this solution:

Using CHCP <codepage>, vDos will still inquire Windows to build the ASCII-Unicode translation table. Missing characters remain as before those of the previous active codepage.

Added that vDos will also try to open the file C_<codepage number>.txt in its start directory.
If present, the first 256 lines are read, supposed to define the (hexadecimal) Unicode values substituting what Windows eventually has defined.
The first 128 lines would normally be empty since DOS codepages generally only define the upper 128 characters differently.
The 129th line of for instance C_895.txt would however define ASCII 128 by: 10C.

So CHCP 895 will be accepted if C_895.txt is present, and substitute the characters of the previous active codepage with those defined in C_895.txt

I’ll send you by wetransfer a pre-release with this functionality.

Jos

[miland]

That's great news, thank you very much and I'm already looking forward to it