Re: Edit - wants all keys when multiline

[Dmitry Timoshkov]

"Vitaliy Margolen"  wrote:
> @@ -761,6 +761,12 @@
>
> case WM_GETDLGCODE:
>    result = DLGC_HASSETSEL | DLGC_WANTCHARS | DLGC_WANTARROWS;
> -
> - if (GetWindowLongW( hwnd, GWL_STYLE ) & ES_MULTILINE)
> - {
> -
> result |= DLGC_WANTALLKEYS;

EDITSTATE structure already has the 'style' member for the optimization
sake, so please use it instead of GetWindowLong(GWL_STYLE).
> - r = SendMessage(hwEdit, WM_GETDLGCODE, 0, (LPARAM) &msMessage);
> - ok(r==0x89, "Expected DLGC_WANTCHARS | DLGC_HASSETSEL | DLGC_WANTARROWS got %x\n", r);

Do not use 0x89 and similar cryptic values in the test, use appropriate
symbolic constants instead.

And please merge your test into existing win.c tests.

Edit - wants all keys when multiline

[Alexandre Julliard]

"Dmitry Timoshkov"  writes:
> And please merge your test into existing win.c tests.

Actually I think it's fine to have separate test files for controls,
we already have a listbox.c so we can have an edit.c. The guideline is
that the layout of the test directory should be close to the layout of
the source directory being tested.

Re[2]: Edit - wants all keys when multiline

[Vitaliy Margolen]

Should I then go for take 3 then? I have added few more tests in take 2. Mainly
for styles being consistent with native.

Monday, November 1, 2004, 12:33:38 PM, you wrote:
> "Dmitry Timoshkov"  writes:
> > And please merge your test into existing win.c tests.
> Actually I think it's fine to have separate test files for controls,
> we already have a listbox.c so we can have an edit.c. The guideline is
> that the layout of the test directory should be close to the layout of
> the source directory being tested.