Re: /* FIXME: Call signal_after_change! */ in callproc.c. Well, why not?

[Eli Zaretskii <eliz@gnu.org>]

> Date: Sat, 21 Dec 2019 21:47:52 +0000
> Cc: emacs-devel@gnu.org
> From: Alan Mackenzie <acm@muc.de>
> 
>  	  else if (NILP (BVAR (current_buffer, enable_multibyte_characters))
>  		   && ! CODING_MAY_REQUIRE_DECODING (&process_coding))
> -	    insert_1_both (buf, nread, nread, 0, 1, 0);
> +            {
> +              beg = PT;
> +              insert_1_both (buf, nread, nread, 0, prepared_position < PT, 0);
> +              if (prepared_position < PT)
> +                prepared_position = PT;
> +              signal_after_change (beg, 0, PT - beg);
                                              ^^^^^^^^
'PT - beg' is just 'nread' in this case, since we are inserting
'nread' characters.  Right?

And I don't think you need the prepared_position stuff here, since PT
doesn't move in signal_after_change.

> @@ -788,7 +796,11 @@ call_process (ptrdiff_t nargs, Lisp_Object *args, int filefd,
>  
>  	      XSETBUFFER (curbuf, current_buffer);
>  	      /* FIXME: Call signal_after_change!  */
> -	      prepare_to_modify_buffer (PT, PT, NULL);
> +              if (prepared_position < PT)
> +                {
> +                  prepare_to_modify_buffer (PT, PT, NULL);
> +                  prepared_position = PT;
> +                }
>  	      /* We cannot allow after-change-functions be run
>  		 during decoding, because that might modify the
>  		 buffer, while we rely on process_coding.produced to
> @@ -822,6 +834,7 @@ call_process (ptrdiff_t nargs, Lisp_Object *args, int filefd,
>  		  continue;
>  		}
>  
> +              signal_after_change (PT, 0, process_coding.produced_char);
>  	      TEMP_SET_PT_BOTH (PT + process_coding.produced_char,
>  				PT_BYTE + process_coding.produced);
>  	      carryover = process_coding.carryover_bytes;

This really ugly, IMO.  And the code logic is very hard to follow and
verify its correctness, given the various use cases.

I think you should simply call signal_after_change after the call to
del_range_2 (telling the after-change hooks that actually nothing was
inserted or deleted).  Then you won't need the prepared_position
thingy.

Thanks.