[rescue] Mainframe on eBay

[Charles Shannon Hendrix]

der Mouse wrote:
>>>Not as far as bash is concerned; whenever I type a ^P to bash, I
>>>want it to be previous-history.  ^P as `suspend' I want only when I
>>>type it to something else.
>>
>>readline is your problem.  That's what bash uses to interface with
>>ttys.
> 
> 
> Then readline needs fixing, or bash needs to stop using it.  Really,
> this is *not* rocket science; I managed to do it on 4.3BSD when hacking
> line editing into csh.  tcsh manages to do it today.  bash could do it
> too - if its maintainers cared to bother.

readline uses CBREAK mode, and until someone rewrites it to use raw 
mode, it will never see some characters.

I don't know if changing readline would break anything or not.  Seems 
like it shouldn't.  Probably most people just aren't going to care about 
changing it though.

Chet Ramey is the maintainer of readline, and his email is chet at po.cwru.edu.

Maybe you could ask him.

The readline homepage is

http://cnswww.cns.cwru.edu/~chet/readline/rltop.html

However, keep in mind that when Chet was told about a bug that caused 
bash to suspend even when it was the only thing running, he said two 
things (summarized):

	- don't do that
	- you can always use another window to send the sleeping bash
	  a SIGCONT

Fortunately, someone else patched bash to not suspend itself in the 
offending situations, which is what tcsh does.

> If I were going to be using Linux enough to make it worth my while, I'd
> port over my shell (a rather mutant csh with, among other things,
> command-line editing that doesn't get upset over c_cc[VSUSP]=0x10).

Personally, I like using tcsh, but I have an overall problem with csh 
shells in general because they just aren't very good.  It was a bad 
idea, implemented poorly.

I am waiting to see what AT&T's new Korn Shell release will be like.

> Of course, ranting about it here is of dubious value, since it's
> clearly not a case like the ls sort order where there is a fix and I
> just don't yet know about it.  (I don't suppose you know of any way to
> rebind ^X so that it's not a prefix key, by any chance?)

I don't think so.  I mean, you can probably bind ^X in your ~/.inputrc 
file, but I don't see a prefix function to bind another key to in case
you still want a prefix key.

If you don't, then just bind ^X to whatever you want and it should be
fine.

You can do it in .bashrc probably, but it might be better to do it in
.inputrc if you want other readline programs to behave the same.

Do you (or anyone) know how to get a dump of the current readline
bindings, which might yield information not in the man pages?