Applied, thanks.
I don't remember why we change to track what the pane wants but there
was probably a reason.
On Mon, Mar 11, 2013 at 03:24:50AM -0500, Chris Johnsen wrote:
> Along with the arrow keys, xterm sends different sequences for the
> Home and End keys depending on whether it is in "normal cursor keys"
> mode or "application cursor keys" mode.
>
> tmux puts xterm clients into "normal cursor keys" mode or
> "application cursor keys" mode, depending on the active pane's
> "application keypad" mode (initially, "normal cursor keys" mode via
> rmkx in tty_start_tty; then "tracking the active pane" via smkx or
> rmkx in tty_update_mode from server_client_reset_state).
>
> However, the xterm terminfo entry only provides the "application
> cursor keys" mode sequences (in khome and kend). Thus, tmux will
> fail to recognize a client xterm's Home and End key sequences unless
> the active pane is in "application keypad mode".
>
> This inconsistent recognition is a problem if Home or End are used
> as a prefix or in a binding: it may be impossible to type the prefix
> or trigger the binding. It may also cause problems for programs
> running in a pane: (unless they send screen's smkx sequence, which
> they really should!) they will see the foreign xterm-style "normal
> cursor keys" mode sequences for Home and End (e.g. Home's "^[[H"
> leaks through tmux as two keystrokes: "M-[" and "H") instead of
> being translated to the *different* screen-style sequences.
>
> This inconsistency does not happen for the arrow keys because
> tty_default_raw_keys supplies "hard coded" support for both variants
> of the arrow key sequences. To fix this problem for Home and End,
> give them similar support. According to xterm's ctlseqs
> documentation, these are the only other keys affected by the cursor
> keys mode.
>
> Another approach would be to always put the client terminal in
> "keyboard transmit" mode (smkx). It looks like tmux actually
> operated like this in the past, but it was changed to the current
> "track the pane mode" in OpenBSD patchset 579 (2009-12-04, ddb5bb80
> in the current Git repository).
> ---
> tty-keys.c | 7 +++++++
> 1 file changed, 7 insertions(+)
>
> diff --git a/tty-keys.c b/tty-keys.c
> index 575920e..25e0ef8 100644
> --- a/tty-keys.c
> +++ b/tty-keys.c
> @@ -82,6 +82,13 @@ const struct tty_default_key_raw tty_default_raw_keys[] = {
> { "\033[C", KEYC_RIGHT },
> { "\033[D", KEYC_LEFT },
>
> + /* Other (xterm) "cursor" keys */
> + { "\033OH", KEYC_HOME },
> + { "\033OF", KEYC_END },
> +
> + { "\033[H", KEYC_HOME },
> + { "\033[F", KEYC_END },
> +
> /* rxvt-style arrow + modifier keys. */
> { "\033Oa", KEYC_UP|KEYC_CTRL },
> { "\033Ob", KEYC_DOWN|KEYC_CTRL },
> --
> 1.8.1.3
>
>
> ------------------------------------------------------------------------------
> Symantec Endpoint Protection 12 positioned as A LEADER in The Forrester
> Wave(TM): Endpoint Security, Q1 2013 and "remains a good choice" in the
> endpoint security space. For insight on selecting the right partner to
> tackle endpoint security challenges, access the full report.
> http://p.sf.net/sfu/symantec-dev2dev
> _______________________________________________
> tmux-users mailing list
> tmux-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/tmux-users
------------------------------------------------------------------------------
Symantec Endpoint Protection 12 positioned as A LEADER in The Forrester
Wave(TM): Endpoint Security, Q1 2013 and "remains a good choice" in the
endpoint security space. For insight on selecting the right partner to
tackle endpoint security challenges, access the full report.
http://p.sf.net/sfu/symantec-dev2dev
_______________________________________________
tmux-users mailing list
tmux-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/tmux-users
