app-shells/bash: add 5.2_p26-r8 with non-preliminary XTWINOPS support

The 5.2_p26-r7 revision contained (harmless) changes that were not yet
intended to be applied to the gentoo repo, owing to a miscommunication
between Sam and I. This commit applies the changes as were eventually
intended and, for this reason, the remainder of this commit message
shall be written accordingly. That is, as if no preceding commit had
been involved to get to this point.

...

Xterm is able to push and pop window titles to a stack and there are
several other terminal emulators that can do so, such as alacritty, foot
and tmux. Take advantage of this feature so as to reinstate automatic
window title setting in the case that the PTY is owned by sshd(8).
Unfortunately, there are a lot of terminal emulators that falsely
advertise themselves as being xterm-compatible, making it impossible to
reliably identify xterm itself. However, we can reliably identify
alacritty, foot and tmux so let's support those three to begin with.

The benefits conferred upon tmux are of a distinct nature, since it was
already the case that it was being whitelisted for title support.
Specifcally, the benefits are as follows:

- title restoration is supported even where tmux(1) is launched prior to ssh(1)
- title restoration is supported for nested instances of tmux

It should be noted that tmux does not forward titles to the outer
terminal emulator by default. Such can be arranged for with the
following configuration.

  set -g set-titles on
  set -g set-titles-string "#T"

Don't enable title setting for GNU screen in the case that the PTY is
owned by sshd(8) and screen(1) was launched prior to connecting with
ssh(1). This is a distinction that can be made by checking whether the
WINDOW variable is set in the environment.

Have the genfun_set_win_title function export a variable named
SHELL_SETS_TITLE upon the first occasion that it is called. Presently,
nothing responds to this variable but the intention is to eventually
have portage respond to it. Portage implements heuristics and behaviours
that are horrifyingly broken. For instance, it considers the mere
presence of PROMPT_COMMAND as somehow proving that the interactive shell
uses it for nothing other than to set the title, despite the fact that:

- the contents of PROMPT_COMMAND may be arbitrarily defined by the user
- the purpose of PROMPT_COMMAND is whatever the user may wish it to be
- nobody in their right mind would export PROMPT_COMMAND
- PROMPT_COMMAND can be an array since 5.1 (making it unexportable)

Worse still, in the event that portage is somehow able to ascertain the
value of PROMPT_COMMAND, it takes its first element and proceeds to
inject its value into an invocation of either sh, $SHELL or bash -c,
irrespective of the consequences. No, I'm not making this up.

As such, the purpose of the SHELL_SETS_TITLE variable is to act as a
straightforward indicator that an interactive shell exists as an
ancestor process and that it will take it upon itself to set a fresh
window title upon its primary prompt being displayed.

Signed-off-by: Kerin Millar <kfm@plushkava.net>
Signed-off-by: Sam James <sam@gentoo.org>

0 files changed, 0 insertions, 0 deletions