Re: [Nano-devel] Better tabstospaces support

[markweston]

On 2018-05-11 18:40, Benno Schulenberg wrote:
> Op 10-05-18 om 12:42 schreef address@hidden:
> > First of all, these new behaviours should only apply when:
> > 1. ISSET(TABS_TO_SPACES)
> > 2. Cursor is on a multiple of tabsize (0 is a multiple too)
> > 3. All characters before cursor are spaces
> >
> > Here's the list:
> > 1. If there are >= tabsize spaces behind cursor,
> >    Left should move cursor left by tabsize spaces
> > 2. If there are >= tabsize spaces ahead of cursor,
> >    Right should move cursor right by tabsize spaces.
>
> What's the use case?  When do you want to go somewhere in the
> middle of contiguous whitespace?  What do you want to do there?

It's annoying. When tabstospaces is on, blocks of tabsize spaces should 
be treated as tabs.

> > 3. If there are >= tabsize spaces behind cursor,
> >    Backspace should remove tabsize spaces instead of just one.
> > 4. If there are >= tabsize spaces ahead of cursor,
> >    Delete should remove tabsize spaces instead of just one.
>
> Use Shift+Tab instead.

It's annoying as well, for the same reason.

Would you merge the patch if I made one?

Because otherwise I have another patch idea, similar to that
"pipe selection to external program" feature you're guys working on.
On save, I want to be able to pipe all text to an external program
(without any prompt) and pass the filename and some nanorc values as 
arguments.
The external program could:
1. Strip trailing whitespace.
2. Convert tabs to spaces or vice versa depending if 
ISSET(TABS_TO_SPACES) and tabsize.
3. Remove trailing semicolons if it's a Lua/Python script.
4. Run gcc -fsyntax-only filename if it's a C file, add gcc diagnostic 
comments (something
   like //gcc: main.c:2:5: error: 'what' undeclared (first use in this 
function))
5. If //gcc: comments are already there then remove them