You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
DECUDK doesn't seem to work when terminal type is k95 #577
Program Shift+F6 with echo -e '\eP0;1|17/414243\e\\'; read
Press Shift+F6 to test the output
If it was working I'd expect to see ABC, but instead I see the escape sequence ^[[17;2~.
It works as expected when the terminal type is VT320, so I thought maybe it was intentionally disabled for k95. But if that were the case, I'd expect the DSR-UDK query to return 23 (i.e. extension not supported), and the DA1 query should also not be including extension number 8.
Btw, I hope you don't mind me flooding you with bug reports, but I've been having a lot of fun with this testing. Don't feel any obligation to fix anything in a hurry.
The bug reports are great! It would probably take ages for me to stumble upon these myself, and I'm sure there are plenty more!
This one I'm aware of. Or was aware of - I'd forgotten, but its probably somewhere in my "problems with the K95 terminal type keymap" notes. If you give K95 the command set key \885 \Kudkf6 it should make Shift+F6 behave the same way it does for the VT320 terminal type.
At the moment the K95 terminal type is setup with an xterm-like keymap (you can dump all keymaps to a file with save keymap foo.ksc to see if you want) because that's what everything seems to expect now, but I don't like just copying xterm like that. I've been tempted to change it to match VT520-with-a-pc-keyboard but I'm not sure how practical that would be. The terminfo file would help to an extent assuming people actually install it (probably unlikely), but then there are programs like emacs which just do their own thing and would need special config. At the very least I probably ought to remap the function keys back to the UDK Kverbs, and then have the UDK Kverbs default to sending whatever xterm sends for the function keys when the K95 terminal type is active.
There is also all the other keyboard stuff xterm (and others) do I guess I need to consider someday. At the moment xterm modes 1034 and 1036 are implemented as overlay keymaps which isn't really ideal, and then there is stuff like XTMODKEYS, etc.
If you give K95 the command set key \885 \Kudkf6 it should make Shift+F6 behave the same way it does for the VT320 terminal type.
I think this should be perfect for my needs. I typically have a bunch of settings that I configure for each terminal so I can get them into a state that's best suited for passing my tests. The main point of this exercise it to be able to recommend terminals to people that need a genuine terminal emulator, so I want to be able to tell them how to get the most out of each one.
And in Kermit's case, you would typically be setting the terminal type to the target terminal anyway (e.g. vt320), so this wouldn't be necessary. It's just easier for me to do a single test run on your "highest" terminal type than to test each type individually.
At the moment the K95 terminal type is setup with an xterm-like keymap
That seems fair enough, considering that's what modern apps would be expecting. And now that I think about it, Xterm probably doesn't have UDK supported by default either (my test setup notes have the keyboard as -kt vt220, which is probably what gets it to work). So my only recommendation then would be to update your DSR and/or DA1 reports to match that state, so apps can tell that the functionality is missing.
At the very least I probably ought to remap the function keys back to the UDK Kverbs, and then have the UDK Kverbs default to sending whatever xterm sends for the function keys when the K95 terminal type is active.
Yeah, I was going to recommend something like that. But it occurred to me that you may still want it disabled by default anyway because of the potential security risks. I'm happy as long a there's some way to turn on the functionality when it's needed, so your set key suggestion is good enough for now. Thanks.
Activity
davidrg commentedon Oct 18, 2026
The bug reports are great! It would probably take ages for me to stumble upon these myself, and I'm sure there are plenty more!
This one I'm aware of. Or was aware of - I'd forgotten, but its probably somewhere in my "problems with the K95 terminal type keymap" notes. If you give K95 the command
set key \885 \Kudkf6it should make Shift+F6 behave the same way it does for the VT320 terminal type.At the moment the K95 terminal type is setup with an xterm-like keymap (you can dump all keymaps to a file with
save keymap foo.kscto see if you want) because that's what everything seems to expect now, but I don't like just copying xterm like that. I've been tempted to change it to match VT520-with-a-pc-keyboard but I'm not sure how practical that would be. The terminfo file would help to an extent assuming people actually install it (probably unlikely), but then there are programs like emacs which just do their own thing and would need special config. At the very least I probably ought to remap the function keys back to the UDK Kverbs, and then have the UDK Kverbs default to sending whatever xterm sends for the function keys when the K95 terminal type is active.There is also all the other keyboard stuff xterm (and others) do I guess I need to consider someday. At the moment xterm modes 1034 and 1036 are implemented as overlay keymaps which isn't really ideal, and then there is stuff like XTMODKEYS, etc.
j4james commentedon Oct 18, 2026
I think this should be perfect for my needs. I typically have a bunch of settings that I configure for each terminal so I can get them into a state that's best suited for passing my tests. The main point of this exercise it to be able to recommend terminals to people that need a genuine terminal emulator, so I want to be able to tell them how to get the most out of each one.
And in Kermit's case, you would typically be setting the terminal type to the target terminal anyway (e.g. vt320), so this wouldn't be necessary. It's just easier for me to do a single test run on your "highest" terminal type than to test each type individually.
That seems fair enough, considering that's what modern apps would be expecting. And now that I think about it, Xterm probably doesn't have UDK supported by default either (my test setup notes have the keyboard as
-kt vt220, which is probably what gets it to work). So my only recommendation then would be to update yourDSRand/orDA1reports to match that state, so apps can tell that the functionality is missing.Yeah, I was going to recommend something like that. But it occurred to me that you may still want it disabled by default anyway because of the potential security risks. I'm happy as long a there's some way to turn on the functionality when it's needed, so your
set keysuggestion is good enough for now. Thanks.