[elinks-users] Ctrl-4 mapping -- kills elinks session?

[cga2000]

On Mon, Sep 18, 2006 at 12:40:14PM EDT, Miciah Dashiel Butler Masters wrote:
> On Sun, Sep 17, 2006 at 05:41:45PM -0400, cga2000 wrote:
> > I frequently issue a Ctrl-R to reload web pages and once in a while I
> > miss the R key and hit the "4" key -- just above the "E" and "R" keys
> > on a US keyboard.
> > 
> > What happens is that:
> > 
> > 1. I still see the original web page but it is overwritten by a shell
> >    prompt -- about 20% down the page.  As far as I can tell, in this
> >    instance, "CTRL-4" is equivalent to some form of SIGKILL, terminating
> >    the ELinks process and bypassing the "Do you really want to exit
> >    ELinks" popup screen.
> 
> Actually, CTRL-4 (also known as ^\) sends SIGQUIT. This is like ^C
> (SIGINT), except that it leaves a core file.

Well .. fancy that .. 

Did a "CTRL-V CTRL-4" and .. right you are ..

Weird I never ran into this before ..

> > 2. The shell session I end up in is frozen and does not accept a refresh
> >    screen command (Ctrl-L).  
> > 
> > 3. If I hit <Enter> after hitting Ctrl-L the screen is cleared but
> >    hitting <Enter> again displays a second prompt on the same line as
> >    the first one .. etc. ad Nauseam Aeternam ..
> > 
> > 4. If I try to use this shell session and issue commands such as "ls"
> >    .. "top" .. etc. they behave normally after I hit enter but the
> >    commands that I type are no longer echoed back to the terminal. 
> > 
> > 5. I need to issue a blind "reset" command to get the terminal to work
> >    normally again.
> > 
> > Has anyone seen this?
> 
> I haven't seen that behaviour, but it is understandable that ELinks
> leaves the terminal in a strange state. If you press Ctrl-C, ELinks
> catches the SIGINT signal and returns the terminal to its state from
> before ELinks was started, but since ELinks doesn't catch a SIGQUIT,
> it doesn't have the chance to clean up.

Makes very good sense ..

I had never issued a CTRL-C -- either accidentally or with quitting in
mind and it does what one would expect. 

> > Is there any way I could deactivate this annoying "Ctrl-4" key combo? 
> 
> On a POSIX system, you can disable the key with the stty(1) utility:
> 
>    stty quit ''
> 
> ELinks could catch the signal, but that would defeat the purpose of
> having the key, don't you think?

:-)

> > I have checked the "Keybinding manager" and the screen configuration
> > file but I haven't found anything.
> > 
> > Also it only seems to affect ELinks:
> > 
> > I checked a few other ncurses applications as well as a bare bash shell
> > and when I hit Ctrl-4 nothing happened.
> 
> How odd. Apparantly, everybody else thinks that it is fine to ignore the
> signal.

Fooled around a little more and it turns out mutt does the same.  

Not sure why bash doesn't, though..  I thought for a second that the
shell in my xterm bombed and that xterm automatically launched a new
one but this is apparently not the case.

I did a "before and after" ..

$ echo $$

.. and the pid of the shell didn't change.

> > Well..  hopefully someone will have run into this and knows the whys and
> > the wherefores of this peculiar phenomenon and perhaps have found a
> > workaround .. 
> > 
> > I would much appreciate any pointers in the right direction since it
> > always happens under the worse circumstances and it's both a time waster
> > and source of aggravation to lose a session with a dozen tabs open right
> > when you're in the middle of something.
> 
> See Setup -> Options manager -> User interface -> Periodic snapshotting.

Kewl.

For the record, in my version of ELinks () it is under ..
  
Setup -> Options manager -> User interface -> Sessions -> Periodic snapshotting.
  
.. shouldn't "snapshooting" be more correct..?

;-)

> HTH,

As always.

Thanks much and have a great day.

cga