Re: [Nano-devel] [PATCH] Window resize handling

[Benno Schulenberg]

Hi Mahyar,

On Tue, Apr 21, 2015, at 23:10, Mahyar Abbaspour wrote:
> As it's mentioned in the TODO list, Nano does not handle window resizes
> properly, so I decided to write a patch to handle window resizes properly
> (hopefully).

Thanks for the patch.  After just some short testing, it appears
to work fine.  Nice!

However, compilation failed at first:
proto.h:31: error: expected ‘=’, ‘,’, ‘;’, ‘asm’ or ‘__attribute__’ before 
‘sigwinch_flag’

After inlcuding <signal.h> there, it compiled fine.

Some nitpicks:

> +     if (kbinput == KEY_WINCH) {
> +         browser_refresh();
> +             old_selected = selected;

This last one seems superfluous, as it was done just seven lines
earlier.

(Also, please use four spaces for the first, third, fifth level indent;
it's not all tabs everywhere.  Use M-P to see the difference.)

>  #ifndef NANO_TINY
> +/* An imaginary key we return to caller when we get SIGWINCH */
> +#define KEY_WINCH -2
> +#endif
> +
> +#ifndef NANO_TINY

There's no need for the #endif and second #ifndef.

> --- src/text.c        (revision 5213)
> +++ src/text.c        (working copy)
> @@ -2207,15 +2207,6 @@
>  
>      edit_refresh();
>  
> -#ifndef NANO_TINY
> -    /* We're going to set jump_buf so that we return here after a
> -     * SIGWINCH instead of to main().  Indicate this. */
> -    jump_buf_main = FALSE;
> -
> -    /* Return here after a SIGWINCH. */
> -    sigsetjmp(jump_buf, 1);
> -#endif

Well... removing this breaks the special case for Unjustify:
right after a ^J it is possible to do an unjustify with ^U.
With your patch it is no longer possible to unjustify when
a window resize took place right after the ^J.  With current
nano this works fine.  (True, a highly unlikely scenario,
but...).

>      /* Read in a character and interpret it.  Continue doing this until
>       * we get a recognized value or sequence. */
> -    while ((kbinput = parse_kbinput(win)) == ERR)
> -     ;
> +     do {
> +             kbinput = parse_kbinput(win);
> +     } while (kbinput == ERR
> +#ifndef NANO_TINY
> +                     && kbinput != KEY_WINCH
> +#endif
> +                     );

Hm?  I don't see what the "&& kbinput != KEY_WINCH" brings.
When kbinput == ERR it necessarily means it isn't anything else.

Benno

-- 
http://www.fastmail.com - A fast, anti-spam email service.