Re: [Nano-devel] More good news

[David Lawrence Ramsey]

Jordi Mallach:

<snip>

>Do we have progress on the FreeBSD backspace stuff, btw?

Yes, but it's too much for the stable branch.  The main problem is that
the wrong behavior that crops up without -K (Backspace == KEY_DC, Delete
== Esc [ 3 ~) doesn't occur consistently at all.  If it always occurred
on FreeBSD boxes under all circumstances, then the best solution of
having a configure check for FreeBSD and hence having nano Just Work
would be possible.  However, according to Wouter's input, this isn't the
case:

                Backspace     Delete
                ------------- ---------
console:        KEY_BACKSPACE KEY_DC
aterm:          KEY_BACKSPACE Esc [ 3 ~
eterm:          KEY_BACKSPACE KEY_DC
gnome-terminal: KEY_DC        Esc [ 3 ~
xterm:          KEY_DC        Esc [ 3 ~

Given this mess, there doesn't appear to be any fix that would make nano
Just Work under all circumstances (short of directly checking the value
of $TERM, which is ugly and would cause breakage every time one of the
associated terminfo entries changed).  Therefore, we'd have to fall back
on using -d (or something similar) as a workaround, but since there's no
space left in the int for new flags, we'd have to

(a) add another flags variable and hack all the code depending on flags
to support it, which is too intrusive for the stable branch,

(b) make flags a 'long long', which isn't ANSI C,

or

(c) take out one flag and replace it with the one for -d, which breaks
compatibility with the stable branch.

I have a patch against current CVS which adds my new lowest-level input
routines and handles possibility c.  This should fix the input problems,
but of course it's only appropriate for the unstable branch.  In case
anyone wants to try it, it's at:

http://pooka_regent.tripod.com/patches/nano/nanokbinput-patch.txt


_____________________________________________________________
Sluggy.Net: The Sluggy Freelance Community!