Re: [Nano-devel] [PATCH/RFC] fix crash when using certain keybindings

[Benno Schulenberg]

On Sun, Apr 6, 2014, at 4:50, Mike Frysinger wrote:
> if you enable debug output, do you see the shortcut_init func getting called 
> multiple times ?  it didn't seem correct for that to be reinitialized for 
> some 
> seemingly unrelated things (like loading into a new buffer).

With debugging, doing the 'src/nano -I' and then  ^R  ^M  ^X,
I see this in the debugging output:

parse_kbinput(): kbinput = 18, meta_key = FALSE, func_key = FALSE, escapes = 0, 
byte_digits = 0, retval = 18
get_shortcut(): kbinput = 18, meta_key = FALSE, func_key = FALSE
matched seq "^R", and btw meta was 0 (menus 1 = 1)

parse_kbinput(): kbinput = 13, meta_key = FALSE, func_key = FALSE, escapes = 0, 
byte_digits = 0, retval = 13
get_shortcut(): kbinput = 13, meta_key = FALSE, func_key = FALSE
matched seq "^M", and btw meta was 0 (menus 40 = 5fff)
get_shortcut(): kbinput = 13, meta_key = FALSE, func_key = FALSE
matched seq "^M", and btw meta was 0 (menus 40 = 5fff)
answer = ""

(Why does it match ^M twice?  That doesn't seem right.)

[and then it reinitializes the shortcuts]

parse_kbinput(): kbinput = 24, meta_key = FALSE, func_key = FALSE, escapes = 0, 
byte_digits = 0, retval = 24
get_shortcut(): kbinput = 24, meta_key = FALSE, func_key = FALSE
matched seq "^X", and btw meta was 0 (menus 1 = 501)

(This is without the sanitizing switch, and both with and without your patch.)

The shortcut list getting reinitialized seems to be intentional, though:
a full justify might have been done (M-J is present in all menus -- mistakenly)
which would have resulted in Uncut being temporarily replaced by Unjustify.

Benno

-- 
http://www.fastmail.fm - Faster than the air-speed velocity of an
                          unladen european swallow