Re: [vile] mini-buffer question

[Thomas Dickey]

On Tue, Mar 10, 2015 at 11:14:29PM -0400, Paul Fox wrote:
> thomas wrote:
>  > On Wed, Mar 04, 2015 at 08:44:58PM -0500, Thomas Dickey wrote:
>  > > On Wed, Mar 04, 2015 at 04:30:29PM -0500, Wayne Cuddy wrote:
>  > > > According the help doc ^G toggles the mini-buffer insert mode. If I
>  > > > use ^G then 'h' to move back a few characters followed by (^G or i) to
>  > > > insert a few characters, if the last character that I insert is '\'
>  > > > when I hit ^G to exit insert mode an actual control-G character is
>  > > > inserted into the mini-buffer. As if I'd entered it similar to ^V^G. 
>  > > 
>  > > I see (had not noticed).  I'll take a look for the cause.
>  > 
>  > hmm - it's inserting ^G because in general vile is using '\' to escape
>  > things such as double-quote (though I see the point that ^V is treated
>  > specially in this regard).
>  > 
>  > so... should '\' not escape control-characters, or should there be some
>  > other special rule?
> 
> well, it traditionally (in vi) didn't -- you have to use ^V to insert
> ^M, or ESC, for instance.  '\' was normally just for quoting, or other
> higher level parsing issues.  at least that was my mental model.  i
> don't think we've ever been good at enforcing that distinction in
> vile, since the actions of ^V and '\' take place in so many places. 
> with something like ^G in the minibuffer, it's not like there's
> great precedent to look at, but i think wayne was right to be surprised.

I agree that it's surprising.  But then once I realized what the cause
was, I saw that I should get some input from others before stirring the
muck around a little. I _think_ that by special-casing control characters
for this bit of logic, I can make the behavior less surprising, but was
unsure if there are other aspects to consider.

-- 
Thomas E. Dickey <address@hidden>
http://invisible-island.net
ftp://invisible-island.net

signature.asc
Description: Digital signature