Re: address@hidden: C-k must be repeated in shell to kill a line that has a prompt]

[Miles Bader]

Karl Fogel <address@hidden> writes:
> > That seems like one possible solution.
> > But what do people think of my proposal
> > that beginning-of-line and end-of-line should do what
> > line-move does, and ignore the fields used in M-x shell?
> 
> I like it.

I think we should be more precise about what's being suggested.

`beginning-of-line' and `end-of-line' should continue to obey fields
when invoked within an input field (though this only matters for
`beginning-of-line', since input fields always end at the end of a line
in comint anyway).  To suggest that they should simply ignore fields
entirely is quite clearly wrong.

The question is -- what should they do when invoked within an _output_
field (and more specifically, within the prompt, which is an output field).

Currently `end-of-line' (and consequently `kill-line'), when invoked
inside a prompt, stops at the end of the prompt, not a the end of the
line, and this can be confusing.  Changing things so that `end-of-line'
only stops when it hits the end of a `real' field (e.g., the comint
input field) seems to be a reasonable fix for this problem.

-Miles
-- 
80% of success is just showing up.  --Woody Allen