bug#8856: 24.0.50; regression: special-display-frame is no longer dedicated

[Drew Adams]

> 3. I do not see the read-only-error problem I reported 
> earlier today from using my full setup, so this pared-down 
> test is no longer sufficient to get that.  I also do not see 
> that error if I use the original recipe, with just hexrgb.el 
> and oneonone.el.  So I'll have to start again from my full 
> setup and pare down to something smaller.  I cannot do that 
> right away, however.

OK, I've now pared down from my full setup again.  Took quite a while.

Use the attached file, starting Emacs, as before, this way, where the first load
file is your latest window.el:

runemacs.exe -Q --debug-init  -l "window-2011-06-21a-MARTIN.el" -l
"throw-three.el" -f "1on1-emacs"

Then, as before (but no need to exit to top level before typing more chars):

M-x f TAB o

When you hit TAB, *Completions* is shown in its frame, as expected. As soon as
you hit the second char (`o'), you get the read-only error.


Note: In the file, both of these two lines are necessary, to cause the error:

(add-hook 'post-command-hook '1on1-fit-minibuffer-frame)
(setq w32-grab-focus-on-raise    nil)

The second is of course Windows-specific.  If I comment that out then the error
is not raised, but instead when I hit `o' the focus just changes to the
minibuffer (no `o' char appears in the minibuffer at this point).

I can tell that the focus changes because the frame border shows that (the
active frame border is a diff color from inactive frame borders - this is a
window mgr thing, not an Emacs thing).

Then, if I hit `o' a second time it (a single `o' is inserted in the minibuffer.

Thx/HTH.

throw-three.el
Description: Binary data