Have you tried running it on the same system with a different graphics card, or the same card with a different DOS, to be 100% sure it is definitely the card at fault?

I have not yet tried it with a different card. I have a Matrox MGA 2064W available and may try it tomorrow. I have also tried MapEdit in MS-DOS 6.22 and Windows XP NTVDM and the graphical glitches appear in both. Also from looking at the actual ROM 8x8 font, MapEdit uses its own font, though it appears to be an 8x8 ASCII text mode font (the tiles themselves are ASCII characters, read from text files called OBJDATA and MAPDATA). MapEdit itself is a fullscreen graphical program--it has a graphical mouse pointer and Windows XP will not display it in a window.

It seems strange that the errors in the text panels are only the non alphabetic characters.

If you look at the map itself, some characters are outright different glyphs, like the centered yellow dots that represent flags in the DOSBox image are the character ƒ in the screenshot taken from the Athlon.

I would have thought the map editor would be using its own characters, not the card's. I'm no gamer though, so hopefully someone will have more ideas.

Oh it does use its own font, but the font uses code page 437.

8x8 ROM font can also be read from the motherboard BIOS and it's located at F000:FA6E. In that case, did you use the BIOS Patcher utility from rom.by? It can mess with the 8x8 ROM font if one is not careful.

Yes, I did, to accommodate the 1733 MHz Thoroughbred processor.

Woolie Wool wrote on 2021-10-29, 23:11:

Yes, I did, to accommodate the 1733 MHz Thoroughbred processor.

There is a command line switch which is /g to prevent the removal of the 8x8 ROM font but it's not perfect. That means that the original BIOS should be patched once again.

Gmlb256 wrote on 2021-10-29, 23:16:

Woolie Wool wrote on 2021-10-29, 23:11:

Yes, I did, to accommodate the 1733 MHz Thoroughbred processor.

There is a command line switch which is /g to prevent the removal of the 8x8 ROM font but it's not perfect. That means that the original BIOS should be patched once again.

If this is the problem it's a pretty big one since I patched the BIOS to accommodate the Thoroughbred processor, and if I have to take that out I might as well not even bother trying to fix it.

Looks more like a Borland EGAVGA.BGI problem to me. This is not text mode nor is it the actual BIOS font (Borland provides their own).

Tried other games that use the same libraries in the same mode i.e. Stellar Conquest III?

leileilol wrote on 2021-10-31, 10:57:

Looks more like a Borland EGAVGA.BGI problem to me. This is not text mode nor is it the actual BIOS font (Borland provides their own).

Tried other games that use the same libraries in the same mode i.e. Stellar Conquest III?

One game that could be tested is Block-Man by Soleau Software and will display the same symptoms as the MapEdit screenshot. Replacing the video card won't fix this.

The BIOS Patcher has also a virtual "dual BIOS", which allows one to temporarily switch back to the unpatched one by pressing the "-" key prior the POST screen appearing.

Just out of curiosity, aren't there any TSRs that can be used to upload fonts missing in the ROM also for the 8x8 font?

retardware wrote on 2021-10-31, 15:25:

Just out of curiosity, aren't there any TSRs that can be used to upload fonts missing in the ROM also for the 8x8 font?

I haven't found any that can handle the 8x8 font AFAIK. The ones available are for the 8x14 font which certain later video cards remove them to gain space for the VESA implementation.

leileilol wrote on 2021-10-31, 10:57:

Looks more like a Borland EGAVGA.BGI problem to me. This is not text mode nor is it the actual BIOS font (Borland provides their own).

Tried other games that use the same libraries in the same mode i.e. Stellar Conquest III?

Confirmed Stellar Conquest III has the same garbage font problem.

Gmlb256 wrote on 2021-10-31, 15:02:

The BIOS Patcher has also a virtual "dual BIOS", which allows one to temporarily switch back to the unpatched one by pressing the "-" key prior the POST screen appearing.

I tried this but nothing seemed to happen. Does it show any visual confirmation? Do you have to tap the key within a certain window of time or hold it while rebooting, or what?

Woolie Wool wrote on 2021-11-06, 02:34:

I tried this but nothing seemed to happen. Does it show any visual confirmation?

The only visual confirmation in your case would be that the AMD Athlon CPU won't be properly recognized during POST.

Do you have to tap the key within a certain window of time or hold it while rebooting, or what?

It's within a certain window of time, but try pressing repeatedly the key just after turning on the computer or a cold boot.

Here's an excerpt related to this feature from ABOUT_BP.TXT file that comes with the BIOS Patcher 4.23.

HOW SAFE IS IT? […]
Show full quote

HOW SAFE IS IT?

Patcher has several features against failures.

1. Virtual "DUAL-BIOS"

Because of keeping changes in one file patcher can load both BIOSes - original
and fixed. This Roll Back starts if you'll press "-" on keyboard during BOOT.
Module of changes reads pressed key and in case of minus simply skips fixed
BIOS and loads old BIOS.

2. Roll Back without keyboard.

Unfortunately, there is no universal initialization of keyboard controller for
all motherboards, that is why sometimes doesn't work read of pressed key. In
this case BIOS uses absence of keyboard. If module of changes can't get signal
from keyboard (DIN or PS/2) it should load old BIOS. Afterwards you can
connect keyboard and flash your old BIOS. But I think that you shouldn't use
this case.

As I count minimal reliabity of patcher is about 95%. IMHO this figure shows
reliability and a question about reliability solved (for people who has little
practise in flashing BIOSes).

Neither method works; the patched BIOS loads even with the keyboard disconnected (and at any rate, plugging a PS/2 keyboard into a powered-on system is AFAIK a big, big no-no that can damage your keyboard controller and keyboard).

However, now I'm interested in the Borland EGAVGA.BGI thing because I'm wondering if anyone else knows of problems with it and its character set.

Woolie Wool wrote on 2021-11-06, 21:44:

Neither method works; the patched BIOS loads even with the keyboard disconnected (and at any rate, plugging a PS/2 keyboard into a powered-on system is AFAIK a big, big no-no that can damage your keyboard controller and keyboard).

However, now I'm interested in the Borland EGAVGA.BGI thing because I'm wondering if anyone else knows of problems with it and its character set.

Seems that /s command line switch from the BIOS Patcher was used which removes that feature. Before moving on with EGAVGA.BGI (never really got problem with this though), could you patch the original BIOS again with the /g command line and perform a reflash?

For example:

1BP-4_23.EXE BIOS.BIN /g

This command line switch also works with later versions.