Re: shell

[Devon Sean McCullough]

Three ways to bury your shell buffer:

1) Waiting for input.
2) Waiting for output to start.
3) Waiting for output to finish.

Ideally, monitor the, er, output tempo and notify the user when it
differs from whatever it was during the time between the last user
input and the command to hide the shell buffer.  E.g., "I'll read mail
now, beep me when output starts or finishes" perhaps requiring a
comint-buffer-hide-hook or similar.
Password protection becomes a special case of output notify.

Easier to show more context with the password prompt.

                Peace
                        --Devon
         /~\
         \ /    Health Care
          X     not warfare
         / \

        Dubya won the digital vote
        Kerry won the popular vote

PS: FSF paperwork arrived, will sign and mail.
I use some other shell improvements, not yet ready.

PPS: LOL!  My sister Jen laughed too.

From: "Richard M. Stallman" <address@hidden>
In-reply-to: <address@hidden> (message from Devon Sean
        McCullough on Wed, 20 Jul 2005 13:28:13 -0400 (EDT))
Subject: Re: shell
Reply-to: address@hidden
Date: Sat, 23 Jul 2005 20:01:22 -0400

    Bury the shell buffer while it sleeps.

That's a horrible thing to do.  No wonder it gets confused
when it wakes up!

    $ sleep 1; echo -n 'Password: '; read -rs; echo -e "\nREPLY=\"$REPLY\""

    Bury the shell buffer while it sleeps.

    The mode line prompts Password: with no indication as to why
    nor which of several shell buffers and other processes want it.

What do you think it should do in this case?

It seems to me that the cleanest way to handle this case
would be to wait until the buffer is current again
before prompting the user.  But that isn't easy to do.