bug#39517: [PATCH] Add new option 'word-wrap-boundary'

[Eli Zaretskii <eliz@gnu.org>]

> From: Jaehwang Jerry Jung <tomtomjhj@gmail.com>
> Date: Sun,  9 Feb 2020 16:43:34 +0900
> Cc: Jaehwang Jerry Jung <tomtomjhj@gmail.com>
> 
> * src/buffer.c (syms_of_buffer): Define buffer-local variable
> word-wrap-boundary.
> * src/buffer.h (struct buffer): Add word_wrap_boundary_.
> * src/xdisp.c (IT_DISPLAYING_WORD_WRAP_BOUNDARY): replaces
> IT_DISPLAYING_WHITESPACE.

Thank you for your interest in Emacs.

When proposing a new feature, please always describe the rationale and
relevant practical use cases.  In this case, it is not clear to me
when it would make sense to allow arbitrary characters as wrap
boundaries.  Maybe you had only white-space characters in mind, but
then why not use [:space:] or some Unicode category specification
instead of allowing any characters?

> +  /* Characters that may cause word wrapping.  */
> +  Lisp_Object word_wrap_boundary_;

Any reason this is a Lisp object and not a C string?

> -#define IT_DISPLAYING_WHITESPACE(it)					\
> -  ((it->what == IT_CHARACTER && (it->c == ' ' || it->c == '\t'))	\
> +#define IT_DISPLAYING_WORD_WRAP_BOUNDARY(it)				\
> +  ((it->what == IT_CHARACTER						\
> +    && strchr ((char *) SDATA (BVAR (current_buffer, word_wrap_boundary)), \
> +		it->c))							\

This cannot be right: characters are stored in Lisp strings in a
multibyte encoding that is superset of UTF-8, so the above will only
support pure-ASCII boundary characters, which is probably not what you
had in mind.

This feature, if we decide to accept it, will also need to be
described in the Emacs user manual, as that is why you exposed this to
Lisp.  We would also want a defcustom form for it, probably in
cus-start.el.

Last, but not least, for a contribution this large, we will need you
to assign the copyright to the FSF.  If you agree, I will send you the
form to fill and the instructions to send it.