Re: Problem with gnus-article-hide-citation

[David Engster <deng@randomsample.de>]

Katsumi Yamaoka <yamaoka@jpl.org> writes:
> I doubted the behavior of C-a (i.e. move-beginning-of-line) is
> really useful.  It moves the point to the next visible field,
> but shouldn't it be the end of the previous visible field (or
> the beginning of the buffer)?  Anyway, how about trying the
> following for a while?  It is just what Emacs 21 does:
>
> (global-set-key "\C-a" 'beginning-of-line)

The problem is that the hidden text practically shadows the first
character of the visible one. You may try this with the following test
code:

(save-excursion
  (set-buffer (get-buffer-create "invtest"))
  (erase-buffer)
  (insert "\naaaaa\nbbbbb\n")
  (goto-char (point-min))
  (forward-line 1)
  (set-text-properties (point) (1+ (point-at-eol)) '(invisible t intangible t)))

Now switch to the buffer "invtest", and put the cursor on the first
"b". Doing C-x = here will show the hidden "a" on position 2. If you go
one step to the right, C-x = will (correctly) show the second "b" at
position 9. In fact, it seems you cannot really access the first "b" at
position 8 through cursor movements. Therefore, EmacsSpeak would not
only have to ignore invisible text, but skip it until it reaches the
first visible character. This seems rather odd to me, and I also didn't
expect that behavior after reading the section on invisible text in the
ELisp reference.

-David