E. C. Masloch

https://github.com/FDOS/freecom/blob/7b7889c367f0742f269dfa2bfd6586993837a50d/tools/50.asm This is the call that changes into 50-line mode which I think is also what changes the height of the cells.

By the way, if I focus a window other than dosemu then the "transparent" cursor shown in the dosemu window is correct.

Insight sets the cursor size using https://fd.lod.bz/rbil/interrup/video/1001.html in https://hg.pushbx.org/ecm/insight/file/fb36068d7047/src/windows.inc#l1493

Didn't test your patch yet. But doesn't Insight's int 10h service 01h call invoke `set_cursor_shape` again at https://github.com/dosemu2/dosemu2/blob/80882dda567de16070e1187ca6e5150ab5bbbe6b/src/base/bios/int10.c#L809 overwriting whatever your patch sets?

I already linked it: https://hg.pushbx.org/ecm/insight/file/fb36068d7047/src/windows.inc#l1493 Quoted: ``` no_cursor: mov cx,2020h ;### 2000h ? jmp set_cursor ; small_cursor: mov cx,0607h ;### 0d0eh ? jmp set_cursor ; large_cursor: mov cx,000dh ;### 000fh...

Yes, it does. The large cursor function is called unconditionally in the Shift+D mode at https://hg.pushbx.org/ecm/insight/file/fb36068d7047/src/iface_2.inc#l1903 So I assume Insight should choose a smaller (but still large) cursor in case...

I added a configuration item to select the value in `cx` used for the set cursor calls: https://hg.pushbx.org/ecm/insight/rev/4d7e14ca09a9 Choosing 000Bh or smaller down to 0004h (inclusive) results in the desired...

In 25-line mode values from 0004h to 000Bh also seem to be fine, at least in `dosemu -X`. So I'm tending towards 000Bh as the default value to use for...

~~I could also get the initial cursor value as set by DOS/BIOS(?) when 50-rows mode is selected, and use that as "large cursor" default value. In that case your suggested...

The Interrupt List says at https://fd.lod.bz/rbil/interrup/video/1001.html > applications which wish to change the cursor by programming the > hardware directly on EGA or above should call INT 10/AX=1130h or >...