OK, thx.
I can tell from your video that WinAGI is able to open the WORDS.TOK file, because it shows the word and group counts in the properties window. I think there's a problem with the words editor form- maybe a control on it that is not compatible with Windows XP or a property being set incorrectly. But why it would do that in XP but not any other Windows versions, I don't know.
I might have access to a Windows XP machine. I'll try to do some testing to see if I can figure this out.
Windows XP actually did complain about the WinAGI installer making changes to my system for my... "security." It appeared the installer wasn't able to register ole controls. Although it seemed to work otherwise, what you said suggests to me this could be related?
After about a month of Windows 9x and emulation and other messing around... well I don't know if this will be any useful information to you but here goes.
I installed Windows 95 within Dosbox-X on a Raspberry Pi (ARM64) system. The emulation speed wasn't the most impressive, no dynamic recompiler it turns out, but it was nearly at the speed of a 486-66. I installed the Windows Installer service and the Visual Basic 6 runtime and... WinAGI 2.1.16 wouldn't install. In fact I tested multiple versions and 1.1.22 was the newest version that seems to install on any 9x system. Note that the latest version of Windows Installer available for 9x is 2.0... though, I don't know if you
intend WinAGI to presently run on 9x anyway. 1.1.22 crashed when I ran it.
Since I had a real system set up with Windows Me (yes, Me) on it, I decided to play around with it a little. I installed VB6, (turns out Me already includes Windows Installer) and... it ran, (even the words.tok editor worked!
) although it crashed upon exit. (From what I can tell, WinAGI also stays running after this and takes a few Windows components with it... this prevents from Windows from shutting down properly, and Scandisk will complain about erroneous FAT entries for temporary files.)
Since I had it (well, an old version) running on a real Windows Me system, next I wanted to know if the crashing was a flaw in Windows 95 or Dosbox. I installed Windows 95 in PCem (on my x64 desktop) and... WinAGI ran just the same way as on my real Windows Me system. So 1.1.22 at least runs on 9x, though with a crashing bug... and it's not as if using such an old version is ideal anyway. WinAGI not working in Dosbox-X must be a bug with its CPU interpreter. I also tried 2.1.16 with the files manually copied into the PCem Windows 95 install, but it just beeped at me when I tried to run it. Haven't tried that on ME yet... I'm not expecting improved results.
I'm not sure by what other means WinAGI would be usable, especially on an ARM system. More accurate emulators such as PCem will probably run even slower on such a system, much too slow to run Windows, and I still haven't gotten it to work in Wine. (Even tried installing IE6.) It appears that AGS will work but... well...
I hardly have to do things this way... but wouldn't it be cool if I did?
I did get
Qt AGI Studio to run on Raspberry Pi... in Wine. Yes. Although Linux builds are available, they fail to run or compile due to missing Qt 3 and 4 libraries. The last update in 2012, (before being revived by cromer) it switched to being Qt4... with Qt3 compatibility modules. 10 years later, and now Qt4 is deprecated, and unavailable in Debian Bullseye's repos! (Might work in Buster?) That sure doesn't motivate me to write software against Qt! (But, Qt AGI Studio isn't what we're here for anyway...)