Re: user32:msg: Test for SetFocus

[Dmitry Timoshkov]

"Anatoly Lyutin"  wrote:
> I try to make this test (in test_ShowWindow : "/* 32 */ { SW_HIDE, TRUE, 
> 0, WmHide_3, TRUE }, ") passes clear and after some changes I observe 
> that SetFocus() is working strange for some case.
> After this I wrote a test case for SetFocus(). SetFocus() have one 
> incorrect behavior for working of combination ( visible-invisible parent 
> or child).

The test actually doesn't do anything useful. It doesn't check
any focus windows and many vital message parameters. Besides we
already have test_SetFocus() in win.c.

user32:msg: Test for SetFocus

[Anatoly Lyutin]

Dmitry Timoshkov wrote:
> The test actually doesn't do anything useful. It doesn't check
> any focus windows

Hmm, I can to fix it.
> and many vital message parameters.

Example?
> Besides we
> already have test_SetFocus() in win.c.

Yes, I have seen this but this test does not show issue that I have found.
I have written this test for one purpose : I can not understand in which 
cases sets hooks HCBT_SETFOCUS and HCBT_ACTIVATE for there combinations 
of visible/invisible windows.

Also it will be nice if somebody says how I can make this test well and 
useful.

user32:msg: Test for SetFocus

[Dmitry Timoshkov]

"Anatoly Lyutin"  wrote:
> > The test actually doesn't do anything useful. It doesn't check
> > any focus windows
>
> Hmm, I can to fix it.
> > and many vital message parameters.
>
> Example?

WM_ACTIVATEAPP, WM_NCACTIVATE, WM_ACTIVATE.
> > Besides we
> > already have test_SetFocus() in win.c.
>
> Yes, I have seen this but this test does not show issue that I have found.
> I have written this test for one purpose : I can not understand in which 
> cases sets hooks HCBT_SETFOCUS and HCBT_ACTIVATE for there combinations 
> of visible/invisible windows.

It doesn't matter for a hook if a window is visible or not, so the behaviour
you are trying to figure out is not useful for a test IMO.

user32:msg: Test for SetFocus

[Anatoly Lyutin]

Dmitry Timoshkov wrote:
> > Example?
>
> WM_ACTIVATEAPP, WM_NCACTIVATE, WM_ACTIVATE.

I shall see this..
> It doesn't matter for a hook if a window is visible or not,

How I can see from this test - it is matter:

/* SetFocus( hwnd ) Invisible parent, visible child */
static const struct message SetFocusSeq_5[] =
{
    { HCBT_SETFOCUS, hook },
    { HCBT_ACTIVATE, hook },
   { WM_WINDOWPOSCHANGING, sent|wparam, SWP_NOSIZE|SWP_NOMOVE },

...

and

/* SetFocus( hwnd ) Invisible parent, invisible child */
static const struct message SetFocusSeq_8[] =
{
    { HCBT_SETFOCUS, hook },
    { WM_IME_SETCONTEXT, sent|wparam|optional, 1 },

Am I wrong?

I know a little about hooks :(.
> so the behaviour
> you are trying to figure out is not useful for a test IMO.

It may to be true, but I have seen a  difference from message sequences  
in Windows and Wine. Only this have motivated me to post this test in 
wine-patches list.

user32:msg: Test for SetFocus

[Dmitry Timoshkov]

"Anatoly Lyutin"  wrote:
> > It doesn't matter for a hook if a window is visible or not,
>
> How I can see from this test - it is matter:
>
> /* SetFocus( hwnd ) Invisible parent, visible child */
> static const struct message SetFocusSeq_5[] =
> {
>    { HCBT_SETFOCUS, hook },
>    { HCBT_ACTIVATE, hook },
>   { WM_WINDOWPOSCHANGING, sent|wparam, SWP_NOSIZE|SWP_NOMOVE },
>
> ...
>
> and
>
> /* SetFocus( hwnd ) Invisible parent, invisible child */
> static const struct message SetFocusSeq_8[] =
> {
>    { HCBT_SETFOCUS, hook },
>    { WM_IME_SETCONTEXT, sent|wparam|optional, 1 },
>
> Am I wrong?
>
> I know a little about hooks :(.

Looks like you need a test for window activation, not focus or hooks.
You may try to add some missing bits to win.c,test_SetActiveWindow().
> > so the behaviour
> > you are trying to figure out is not useful for a test IMO.
>
> It may to be true, but I have seen a  difference from message sequences  
> in Windows and Wine. Only this have motivated me to post this test in 
> wine-patches list.

Are you trying to fix a bug?

user32:msg: Test for SetFocus

[Anatoly Lyutin]

Dmitry Timoshkov wrote:
> Looks like you need a test for window activation, not focus or hooks.
> You may try to add some missing bits to win.c,test_SetActiveWindow().

May be.. I decide to test SetFocus() because in this function calls 
set_active_window() and another case set_active_window() not calls. I 
try to understand how SetFocus() works..
> Are you trying to fix a bug?

No. I have written in the first letter with test: I try to make
"test_ShowWindow : "/* 32 */ { SW_HIDE, TRUE, 0, WmHide_3, TRUE },

passes clear. And I found that it falls because HCBT_SETFOCUS not set 
for this window after  WM_WINDOWPOSCHANGING and if I put a 
SetFocus(hwnd) in static BOOL show_window( HWND hwnd, INT cmd )
{
...
if (cmd == SW_HIDE)
    {
...
          /* Revert focus to parent */
        hFocus = GetFocus();
        if (hwnd == hFocus || IsChild(hwnd, hFocus))
        {
            HWND parent = GetAncestor(hwnd, GA_PARENT);
            if (parent == GetDesktopWindow()) parent = 0;
            SetFocus(parent);
        }

to this -> SetFocus(hwnd); // naturally this is a very bad hack, 
but I only search working of show_window()
       or
       SetFocus(GetAncestor(hwnd, GA_ROOT));
        this hook will be set and broke working of another test, but in 
another cases such us :
       SetFocus( GetAncestor(hwnd, GA_PARENT));
        or
       SetFocus(0);
       this hook not set. // nothing to strange of cause..

..
}

I try to catch issue why in another test hook sets well and why in this 
test hook is not sets.