bug#31204: 25.3; Make word motion more customizable

[Lars Ingebrigtsen]

Lars Ingebrigtsen <larsi@gnus.org> writes:

>> As an illustration (where ‘|’ specifies word motion stops when going
>> left to right):
>>
>>     foo| ***| +++| (|bar|)|

[...]

> I think this sounds like it could be useful.  If we added such a hook to
> `forward-word', what would the rest of the code look like to make
> `C-<right>' work this way?

I've played a bit at the patch below, but I tend to think that this is
going about things the wrong way.  That is, for something like this to
work meaningfully, it would require a lot of setup (because
Vfind_word_boundary_function_table) would also have to be altered in
conjunction with this.

I.e., it's really about changing the definition of what a "word" is, and
in that case, I think it would be easier to just do that in a syntax
table, and then everything would work automatically.

(Or by advising the functions here.)

So I don't think it'd be worth it to proceed with something like the
below, and I'm therefore closing this bug report.

diff --git a/src/syntax.c b/src/syntax.c
index f9022d18d2..02d4dd4b9a 100644
--- a/src/syntax.c
+++ b/src/syntax.c
@@ -1462,20 +1462,33 @@ scan_words (ptrdiff_t from, EMACS_INT count)
 
   while (count > 0)
     {
-      while (true)
+      if (!NILP (Vfind_word_start_function))
        {
-         if (from == end)
+         Lisp_Object np = call2 (Vfind_word_start_function,
+                                 make_fixnum (from), make_fixnum (end));
+         if (!FIXNUMP (np))
            return 0;
-         UPDATE_SYNTAX_TABLE_FORWARD (from);
+         from = XFIXNUM (np);
+         from_byte = CHAR_TO_BYTE (from);
          ch0 = FETCH_CHAR_AS_MULTIBYTE (from_byte);
-         code = SYNTAX (ch0);
-         inc_both (&from, &from_byte);
-         if (words_include_escapes
-             && (code == Sescape || code == Scharquote))
-           break;
-         if (code == Sword)
-           break;
-         rarely_quit (from);
+       }
+      else
+       {
+         while (true)
+           {
+             if (from == end)
+               return 0;
+             UPDATE_SYNTAX_TABLE_FORWARD (from);
+             ch0 = FETCH_CHAR_AS_MULTIBYTE (from_byte);
+             code = SYNTAX (ch0);
+             inc_both (&from, &from_byte);
+             if (words_include_escapes
+                 && (code == Sescape || code == Scharquote))
+               break;
+             if (code == Sword)
+               break;
+             rarely_quit (from);
+           }
        }
       /* Now CH0 is a character which begins a word and FROM is the
          position of the next character.  */
@@ -3792,6 +3805,12 @@ syms_of_syntax (void)
 In both cases, LIMIT bounds the search. */);
   Vfind_word_boundary_function_table = Fmake_char_table (Qnil, Qnil);
 
+  DEFVAR_LISP ("find-word-start-function",
+              Vfind_word_start_function,
+              doc: /* Function called to find the start of a word.
+It's called with two parameters, POS and LIMIT.  */);
+  Vfind_word_start_function = Qnil;
+
   DEFVAR_BOOL ("comment-end-can-be-escaped", comment_end_can_be_escaped,
                doc: /* Non-nil means an escaped ender inside a comment doesn't 
end the comment.  */);
   comment_end_can_be_escaped = false;


-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no