Re: [Nano-devel] [PATCH] add rebinding for Ctrl+Arrow-keys combinations

[Benno Schulenberg]

On Wed, Aug 3, 2016, at 16:12, Rishabh Dave wrote:
> On Wed, Aug 3, 2016 at 1:36 PM, Benno Schulenberg
> <address@hidden> wrote:
> > You might have wanted to discuss the bug first.  Maybe the fix is not
> > to make the arrow keys rebindable but to properly warn that they are
> > unrebindable (instead of making it seem as if they were rebound but
> > not actually having done anything, as it is now).
> 
> Unrebindable... how? This patch does add the ability of reassigning
> the functions to ^Arrow keys. (which is rebinding, if I am correct)

In the current code they /are/ unrebindable, but make it seem as if
they were rebound.  Maybe the solution is to no longer make them
/seem/ rebound, instead of trying to allow them to be rebound.  The
plain arrow keys, and Home and End and such, cannot be rebound either.
That's why I said: discuss first.  I reported that thing a year ago;
opinions change.

> > Ah, but /that/ you should do.  A patch against old code is not useful.
> 
> Ohkay. Updated patch is attached.

Patch is useless -- you didn't address /any/ of the other issues I
mentioned: literal values, breaking of iTerm, and breaking of any
other terminal that doesn't produce 544 for Ctrl+Left.  /And/ you
dumped the first block of code just somewhere in the middle, breaking
an if-else chain.

> About defines, assigning CONTROL_LEFT (or respective for
> other arrow keys) doesn't help.

No, you'll have to make corresponding changes everywhere else.


> > What are those lines doing there?  Are you saying that the rebinding
> > of toggles doesn't work right now?
> 
> Before adding this code, valgrind reported that jump depends on
> uninitialized value at the body of the loop below in src/help.c.

Ah.  Interesting.  So it doesn't have anything to do with rebinding
^Arrow keys, so it doesn't belong in your patch -- it is a separate
issue.  Report a bug.  Include the relevant info from valgrind, and
include the minimal nanorc that makes the conditional jump appear,
and the minimum chain of keystrokes to trigger it.


> Can I make a request? I would like to try searchable help
> feature-request (http://savannah.gnu.org/bugs/?28994). Can it be
> kind-of reserved?

Then post a comment on the bug, saying that you are working on it.
And post a comment again when you've figured out a way to address it.

Benno

-- 
http://www.fastmail.com - Same, same, but different...