Mumak wrote on 2023-08-11, 10:08:

For the invalid ATAPI device, what drive details does it show in that case? Is there some other mismatch in the info?

IIRC it showed the details of the optical drive twice; they were correct from what I could tell.

I didn't notice any other mismatches, and the "double ATAPI" thing happens fairly randomly, maybe only like 10% of the time? I was not able to get a report with it happening.

And yeah, thanks for the clarification re: the Vibra. It's not a particularly great card anyway, so, whatever, man. 😄 (Sorry if that hurt someone's feels.)

1. Does actual HWiNFO detect CGA Plantronics and Olivetti/ATT 6300, IBM PGC, Secondary FDC controller, streamer devices on FDC port?
2. If VESA/DDC multidisplay in HWiNFO detection will be made, I recommend X-VESA utility by Marco Pistella which can detect two monitors on same videocard (f.e. Geforce 2/3/4/5 series VGA+DVI).
3. It is possible to detect FDC controller version, see here https://github.com/torvalds/linux/blob/master … /block/floppy.c

b_w wrote on 2023-08-13, 13:02:

1. Does actual HWiNFO detect CGA Plantronics and Olivetti/ATT 6300, IBM PGC, Secondary FDC controller, streamer devices on FDC port?
2. If VESA/DDC multidisplay in HWiNFO detection will be made, I recommend X-VESA utility by Marco Pistella which can detect two monitors on same videocard (f.e. Geforce 2/3/4/5 series VGA+DVI).
3. It is possible to detect FDC controller version, see here https://github.com/torvalds/linux/blob/master … /block/floppy.c

I'm not aware of any software possibility to detect various CGA adapters. AFAIK all those have same register sets. If you maybe know how to do that, let me know.
FDC controller version - I wasn't aware of that but let me look at that code...
Streamer devices can't be detected by HWiNFO.

Detect Plantronics ColorPlus and Olivetti https://forum.vcfed.org/index.php?threads/is- … software.79115/

Also it's a good idea to make HWiNFO graphics frames work in EGA text mode (since it has reprogrammable character set also) not only it regular (S)VGA, if you using BIOS calls to do it - it should work right without great modifications, you need 8x14 font. You only has to detect that this system is EGA. Either color or mono.

Hercules plus also has reprogrammable character set, see sample below and use font.bin file sized 14x256=3584 bytes:

1#include <dos.h>
2#include <conio.h>
3#include <stdio.h>
4#include <stdlib.h>
5#include <string.h>
6#include <stddef.h>
7#include <math.h>
8#include <memory.h>
9
10int main (int argc, char *argv[])
11{
12char * p;
13char i;
14unsigned short j,k,l, m;
15union REGS regs;
16struct SREGS sregs;
17char fname[255];
18FILE* f;
19
20regs.x.ax = 0x07;
21int86x(0x10, &regs, &regs, &sregs);
22printf("HGC FontSet. Press a key to start.\n");
23if (argc < 2) { printf("USAGE: hgcset [font.fnt], font.fnt - 8x14 font file\n"); return 1; }
24
25strcpy(fname,argv[1]);
26printf ("Using font: '%s' ... ",fname);
27
28f = fopen(fname,"r");
29if (f)
30{
31printf ("OPENED\n");
32
33asm{
34mov dx, 0x3BA
35in al,dx
36mov i,al
37}
38if (i & 0x80)
39{
40i &= 0x70;
41switch (i)
42{
43case 0x00: printf ("HGC detected\n"); break;
44case 0x10: printf ("HGC+ detected\n"); break;
45case 0x50: printf ("HGC InColor detected\n"); break;
46default: printf ("HGC unknown detected\n"); break;
47}
48
49while (!getch());
50
51asm{
52mov dx, 0x3B4
53mov ax,0x0114
54out dx,ax
55mov dl,0xBF
56mov al,1
57out dx,al
58}
59///////////////////////
60p = malloc(8*14*256);

61if (fread(p, 1, 8*14*256, f))
62{
63m = 0;
64for (j = 0; j < 256; j++)
65for (k = 0; k < 14; k++)
66{
67i = p[m]; m++; 
68l = (j << 4) + k;
69asm{
70mov ax,0xB400
71mov es,ax
72mov ax, l
73mov di, ax
74mov al, i
75stosb
76}
77}
78
79}
80if (p) free(p);
81///////////////////////
82asm{
83mov dx, 0x3BF
84mov al,0
85out dx,al
86}
87
88}
89else
90{
91printf("HGC not detected.");
92}
93
94fclose(f);
95}
96else
97printf ("BAD\n");
98
99return 0;
100}

b_w wrote on 2023-08-13, 15:34:

Detect Plantronics ColorPlus and Olivetti https://forum.vcfed.org/index.php?threads/is- … software.79115/ Also it's a good idea […]
Show full quote

Detect Plantronics ColorPlus and Olivetti https://forum.vcfed.org/index.php?threads/is- … software.79115/
Also it's a good idea to make HWiNFO graphics frames work in EGA text mode (since it has reprogrammable character set also) not only it regular (S)VGA, if you using BIOS calls to do it - it should work right without great modifications, you need 8x14 font. You only has to detect that this system is EGA. Either color or mono.
Hercules plus also has reprogrammable character set, see sample below and use font.bin file sized 14x256=3584 bytes:

Spoiler

Copy code to clipboard

1#include <dos.h>
2#include <conio.h>
3#include <stdio.h>
4#include <stdlib.h>
5#include <string.h>
6#include <stddef.h>
7#include <math.h>
8#include <memory.h>
9
10int main (int argc, char *argv[])
11{
12char * p;
13char i;
14unsigned short j,k,l, m;
15union REGS regs;
16struct SREGS sregs;
17char fname[255];
18FILE* f;
19
20regs.x.ax = 0x07;
21int86x(0x10, &regs, &regs, &sregs);
22printf("HGC FontSet. Press a key to start.\n");
23if (argc < 2) { printf("USAGE: hgcset [font.fnt], font.fnt - 8x14 font file\n"); return 1; }
24
25strcpy(fname,argv[1]);
26printf ("Using font: '%s' ... ",fname);
27
28f = fopen(fname,"r");
29if (f)
30{
31printf ("OPENED\n");
32
33asm{
34mov dx, 0x3BA
35in al,dx
36mov i,al
37}
38if (i & 0x80)
39{
40i &= 0x70;
41switch (i)
42{
43case 0x00: printf ("HGC detected\n"); break;
44case 0x10: printf ("HGC+ detected\n"); break;
45case 0x50: printf ("HGC InColor detected\n"); break;
46default: printf ("HGC unknown detected\n"); break;
47}
48
49while (!getch());
50
51asm{
52mov dx, 0x3B4
53mov ax,0x0114
54out dx,ax
55mov dl,0xBF
56mov al,1
57out dx,al
58}
59///////////////////////
60p = malloc(8*14*256);

…Show last 40 lines

61if (fread(p, 1, 8*14*256, f))
62{
63m = 0;
64for (j = 0; j < 256; j++)
65for (k = 0; k < 14; k++)
66{
67i = p[m]; m++; 
68l = (j << 4) + k;
69asm{
70mov ax,0xB400
71mov es,ax
72mov ax, l
73mov di, ax
74mov al, i
75stosb
76}
77}
78
79}
80if (p) free(p);
81///////////////////////
82asm{
83mov dx, 0x3BF
84mov al,0
85out dx,al
86}
87
88}
89else
90{
91printf("HGC not detected.");
92}
93
94fclose(f);
95}
96else
97printf ("BAD\n");
98
99return 0;
100}

Thanks! I see the CGA test requires native CGA mode to be set, which is not supported by HWiNFO GUI. I'm not even sure how HWiNFO for DOS will run/look on a CGA adapter.

For latest version 7.62 noticed missing info for windows 98 - the Computer Description is not there. It shows on newer OS's though.

It's in the registry at HKLM\System\CurrentControlSet\Services\VxD\VNETSUP - keyname is Comment

Robbbert wrote on 2023-09-11, 14:17:

For latest version 7.62 noticed missing info for windows 98 - the Computer Description is not there. It shows on newer OS's though.

It's in the registry at HKLM\System\CurrentControlSet\Services\VxD\VNETSUP - keyname is Comment

Thanks, will be fixed in the next build.

Thanks, tried 7.64 on Win98Se - works.

Just to satisfy my curiosity, I tried running HWINFO32.exe in Windows for Workgroups, but I got a message to say it was incompatible with win32s. Then I tried with the installer and got the same error.
I imagine that the hooks you use to gather the information might not exist. Just wondering if there's any thought or possibility of HWINFO32 working with WIN32S at some time?

HWiNFO32 requires at least Windows 95 to run. Support for earlier versions like 3.x would require significant effort which is currently not planned. For such systems I rather recommend using HWiNFO for DOS.

Thanks for answering. I thought that might be the case. Not a problem. 😀

I notice the new version 766 will not install on Win98SE.

Does that mean you've removed support for various old platforms?

Can you list what you do support now?

Robbbert wrote on 2023-11-13, 12:05:

I notice the new version 766 will not install on Win98SE.

Does that mean you've removed support for various old platforms?

Can you list what you do support now?

That wasn't intentional.
Is the problem installation or running the HWiNFO32 ?

Installing - didn't get to run it.

Hmm, the installer (InnoSetup) was updated to a newer version but I thought it maintained backwards compatibility. Are you still able to run the latest portable HWiNFO32?

The portable version works.

The installer starts up and asks for the language, but then says it can't run on this computer.

"This program does not support the version of Windows your computer is running".

Thanks, I will check that. Important is that HWiNFO32 still maintains the compatibility 😀

Thanks, I added the actual message above.

Main problem of Hwinfo for Dos are for me at least are system freezes, on some machines. Navratil and others are crashing machine so much.

Its possible write it more safely? Its guess that you are using some ASM calls to check some registry if there is something what you can identify and after you asking for details.. I suppose that ask for something what is not existing on machine is not freeze machine itself.. So problem would be probably something like unhandled exception, some values what you don't expect.. or what else could cause these fails?

This is a bunch of hacks and it simply cannot be clearly done. Some routine that works on some machine can crash/hang on other machine, it's hard to avoid this when you are trying to get maximum available info. Maybe you can e.g. use CPUID first and then decide if some specific hack make a sense to call for such a machine with that CPUID range or not, that can slightly lower the probability of hang but it cannot be avoided at all.

Hard to know what to improve when I don't know exactly where and on which systems it crashes. I suppose it's mostly the Super IO check which others don't check, so they don't crash. But this can be disabled in HWiNFO.
Problem with such legacy systems is that there's no universal standard or method that would help you detect many things. So you need to perform several test probes (accessing various registers and I/O) which on some other systems can use these locations for something else or not use at all which will cause a crash. PnP was the first attempt to somehow standardize things but even that doesn't work properly on many systems.