Re: [PATCH] prompt: support generic non-visible regions

[Felipe Contreras <felipe.contreras@gmail.com>]

On Thu, Aug 11, 2022 at 2:35 PM Roman Perepelitsa
<roman.perepelitsa@gmail.com> wrote:
>
> On Thu, Aug 11, 2022 at 9:24 PM Felipe Contreras
> <felipe.contreras@gmail.com> wrote:
> >
> > On Thu, Aug 11, 2022 at 12:24 PM Roman Perepelitsa
> > <roman.perepelitsa@gmail.com> wrote:
> > >
> > > On Thu, Aug 11, 2022 at 4:37 PM Roman Perepelitsa
> > > <roman.perepelitsa@gmail.com> wrote:
> > > >
> > > > I like this patch [...]
> > >
> > > I should've said that I like the idea behind it. The patch implements
> > > one part of it. To complete the implementation percent expansion needs
> > > to output \001 and \002.
> >
> > That's easy. The current code in prompt.c uses Inpar and Outpar, which
> > are 0x88 and 0x8A respectively. If instead we do:
> >
> > #define invis_start ((char) 0x01)
> > #define invis_end   ((char) 0x02)
> >
> > And replace Inpar/Outpar with invis_start/invis_end the code should
> > behave identically (since those characters are removed).
> >
> > Then in promptexpand() we simply not remove them.
>
> Let me clarify. Ideally we want (I think) for these two tests to pass:
>
> 1.
>
>     [[ $(print -r -- ${(%):-'%{hello%}'}) == hello ]]
>
> 2.
>
>     [[ $(print -r -- ${(m)#${(%):-'%F{1}❎%f'}}) == 2 ]]
>
> The first test currently passes, the second fails. In my previous
> comment I tried to say that making the second test pass will cause the
> first test to fail.
>
> Do you see how to make both of these tests pass?

We could add a mode to the % flag which is passed to promptexpand() in
order to tell it to not remove the invisible start/end markers.

However, in my tests this simply adds two characters to $((m)#...}. I
don't see how this would return the correct amount of visible
characters.

-- 
Felipe Contreras