Re: [dev] [st] Keyboard patches

From: Roberto E. Vargas Caballero <>
Date: Tue, 13 Nov 2012 22:06:33 +0100

> The patches are removing the hardcoded values in st and make them more
> configurable to how terminals work. In the future important shortcuts

Not only does this. There is some conditions that can not be handle with the
previous code. All the keypad keys can generate two values depend of the
keypad mode, one of them is the ansi sequence that the own terminal
understand for this function (KX_KP_Up generates \033[A, that is the
sequence for CUU), or one code that identify to the key (KX_KP_Up generates
\033Ox). This is done in this way because usual programs don't have to worry
about them, because they generate correct sequences and they don't need to
know if the sequence come from the key or the output of the program. But
when you are in raw mode you want to know what key was pressed, and the ansi
sequence does't give to you this information. you couldn't put any of these
values in the Key array because they depend of the keypad mode or cursor
mode, so the Key array was not useful.

> that were missed but some asshole is using them hardcoded in its so‐
> called »ncurses application« will be easier to add. But yes, most of it
> is compatibility bloat.

I use a lot of these sequences. I know that it is not the typical usage of a
terminal emulator by a novice, but some of us have to use terminals for
other things. When you use telnet or ssh they disable the kernel echo in the
terminal and let to the other side the task of generating the
echo. Sometimes I have to connect to systems which don't generate the echo
(due to some reasons that are not important here) and It is very important
for me see what I am writing.

Best regards,
Received on Tue Nov 13 2012 - 22:06:33 CET

This archive was generated by hypermail 2.3.0 : Tue Nov 13 2012 - 22:12:04 CET