Pfthb

I'm attempting to build a simple terminal UI that displays a string that is updated every second (somewhat like watch -n 1 <command> would do but without doing a full screen refresh) and I'm running into a problem when the content to be refreshed is larger than the terminal emulator screen. I feel like it should be solvable without using curses but I can't figure it out.

The program I've written is pretty simple, the logic works something like this (this is pseudocode in python, the actual script is in Haskell but that shouldn't change the problem).

ansi_clearline = "33[2K"
ansi_upline = "33[1A"

printed_value = fetch_value() # String formatted for display to the terminal
print(printed_value)
while True:
 num_lines = len(printed_value.split("n"))
 control_chars = ansi_upline.join(num_lines * [ansi_clearline])
 printed_value = fetch_value()
 print(control_chars)
 print(intial_value)

That is, for every line in the last output we print a control sequence to clear the line, then a control sequence to move up a line, except for the first line where we only print a clear line.

This works fine as long as all the content fits into the terminal emulator screen. Unfortunately it breaks if the output (printed_value) is longer than the emulator screen. Presumably this is because the ansi control sequences don't know anything about moving into the scroll back buffer and don't let you move up one line once you're at the top of the screen. What this looks like is the following:

lineone 
linetwo
 ------top of screen
linethree
linefour
linefive
 -------Bottom of screen

Scroll up a bit with the mouseo

lineone 
linetwo
 ------top of screen
lineone 
linetwo
lineone 
 -------Bottom of screen
linetwo
linethree
linefour
linefive

Hopefully that graphic makes clear what is happening. Is there any way to achieve what I want to do (effectively using control sequences to clear part of the scrollback buffer) without using curses?