Re: bug in forward-visible-line: Patch

[Stefan Monnier]

>    BTW, how about a `invisible-p' function that does above ?
> 
> I believe visiblep would be slightly more useful (does not make any
> real difference though).  I believe the usual convention is to only
> use -p if the function name is already multi-word, although this
> convention is not universally followed.
> 
> What about the following function?  I should still double-check it
> more carefully, but on first testing in a reasonably complex
> situation, it seems to work OK.  I believe that I could use the
> function not only to make forward-visible-line behave correctly (even
> after my patch, a problem remains, as I pointed out earlier), but
> probably in other situations as well.
> 
> ===File ~/invp.el===========================================
> (defun visiblep (&optional pos)
>   "Return t if character at POS is currently visible.
> POS defaults to point."
>   (unless pos (setq pos (point)))
>   (let ((prop (get-text-property pos 'invisible)))
>     (cond
>      ((null prop))
>      ((eq buffer-invisibility-spec t) nil)
>      ((memq prop buffer-invisibility-spec) nil)
>      ((assq prop buffer-invisibility-spec) nil)
>      ((listp prop)
>       (catch 'found
>       (dolist (var prop)
>         (if (or (memq var buffer-invisibility-spec)
>                 (assq var buffer-invisibility-spec))
>             (throw 'found nil)))))
>      (t))))

Don't reinvent the wheel: implement it in C where it's
already available (and so you're sure it really works the
same).  The C version return 0 1 or 2 so you can also tell
whether it's got an ellipsis or not.


        Stefan