Is disabling ctrl-alt-backspace really such a good idea? - no.

[John Moser]

On Thu, Feb 12, 2009 at 8:41 PM, Charlie Kravetz
<cjk at teamcharliesangels.com> wrote:
> Okay, I have been reading this thread from the beginning. It seems like
> those making the most noise are the same individuals with the knowledge
> and ability to easily add the ability to use C-A-B back. Why should the
> thousands who do not need the ability be forced to have it to make it
> easier for the few that want it to be able to use it?
>
> Do you not customize your systems? This is just a very quick fix to
> re-enable the keypresses for yourselves. The many thousands who could
> care less are not even on the mailing list.
>

I actually removed zap myself ages ago due to randomly hitting it.  My
big problem is when I kill X I have 50 windows open and lose a LOT of
work.  It's disruptive.  I lose IRC, I lose the 10 tabs I have open in
firefox.  If you could make pidgin, xchat-gnome, and Firefox actually
restore properly (read:  Firefox RELOADS the tab it was in, it doesn't
come back up into the exact same state it shut down in!  This sucks!),
that would solve a large part of the problem for me.  Say, pidgin
crash-recovers by loading up all the same tabs with the same content
they had; xchat-gnome loads the same channels, queries, servers, with
the backlog all in place, and then complains it was disconnected and
reconnects for you.  That right there would make this whole thing far
less painless for me.

The big, unsolvable problem here for me is anything opened in a
terminal necessarily gets lost in a reboot.  In an X restart, it's on
me to use screen to save myself.  It's STILL worth more to me to NOT
accidentally zap X than it is to have to force a reboot when I could
have recovered.  Firefox, gaim, and xchat-gnome are annoying; with
openoffice and thunderbird, I recover documents and e-mail drafts
anyway; but with the terminal I can actually lose work.  I was zapping
by accident more often than on purpose so out it went.

This is precisely why I've been pushing for something that would TRY
to ask the user for confirmation, and give up if the client died or
got no response.  At least an ACCIDENTAL C-A-B would be abortable by
me.  The REAL solution is, yes, to fix the misbehaving programs (X or
otherwise); everyone has their own little list of programs that don't
recover from sudden X server kill to their liking, and I'm sure
they're all different from mine.