user32: Use top level window as dialog parent for modal dialogs.

[Dmitry Timoshkov]

Jacek Caban <jacek at codeweavers.com> wrote:

> > The source of the regression that your patch is supposedly tries to fix is
> > that the dialog code enables wrong window at the end of dialog message loop.
> No, it fixes issue I found while looking at this, but it's a separated 
> and isolated issue. I don't touch the code that you ask me to write a 
> test for. You fail to understand that.

I perfectly understand what you are trying to fix, and you actually mentioned
it yourself already: "This is part of the fix for bug 40282". And that bug is
a regression caused by your previous changes not backed up by the tests, and
I already explained that the source of the regression is in wrong handling
of enable/disable of the dialog owner, and once again I'd like to point out
that fixing this without any tests simply can't work.

> BTW, your tests don't test what you ask me to test and they were enough 
> for you to write a wrong fix. Otherwise you'd know that this code in the 
> end of the loop should be removed. Then again, it's not related to this 
> particular patch.

Well, at lease I have written some tests before I tried to fix anything,
and I'm not claiming that my fix is perfect/correct/useful at all. You are
more than welcome to show your own exhaustive tests that check every detail
of the dialog owner enable/disable logic.

P.S.
I'm not the person who commits patches to Wine, why are you arguing with me
if it's pretty clear that you are not going to write any tests at all in order
to demonstrate that your patch is correct?

-- 
Dmitry.