[PATCH 1/7] d3d8/tests: Fix possible test failures.

[Zhiyi Zhang]

expect_messages points to a local variable. When it go out of scope, it
reads into random memory, causing failures.

Signed-off-by: Zhiyi Zhang 
---
 dlls/d3d8/tests/device.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/dlls/d3d8/tests/device.c b/dlls/d3d8/tests/device.c
index d87832b716..2b7a768f65 100644
--- a/dlls/d3d8/tests/device.c
+++ b/dlls/d3d8/tests/device.c
@@ -3200,6 +3200,7 @@ static void test_wndproc(void)
     flush_events();
     ok(!expect_messages->message, "Expected message %#x for window %#x, but didn't receive it, i=%u.\n",
             expect_messages->message, expect_messages->window, i);
+    expect_messages = NULL;

/* World of Warplanes hides the window by removing WS_VISIBLE and expects Reset() to show it again. */
     device_style = GetWindowLongA(device_window, GWL_STYLE);
@@ -3219,6 +3220,7 @@ static void test_wndproc(void)
     flush_events();
     ok(!expect_messages->message, "Expected message %#x for window %#x, but didn't receive it.\n",
             expect_messages->message, expect_messages->window);
+    expect_messages = NULL;

ok(windowpos.hwnd == device_window && !windowpos.hwndInsertAfter
             && !windowpos.x && !windowpos.y && !windowpos.cx && !windowpos.cy

[Marvin]

Hi,

While running your changed tests, I think I found new failures.
Being a bot and all I'm not very good at pattern recognition, so I might be
wrong, but could you please double-check?

Full results can be found at:
https://testbot.winehq.org/JobDetails.pl?Key=68431

Your paranoid android.

=== w1064v1809_2scr (32 bit report) ===

d3d8:
device.c:1002: Test failed: The cursor handle is 00010005

[Henri Verbeet]

On Mon, 30 Mar 2020 at 11:48, Zhiyi Zhang  wrote:
> @@ -3200,6 +3200,7 @@ static void test_wndproc(void)
>      flush_events();
>      ok(!expect_messages->message, "Expected message %#x for window %#x, but didn't receive it, i=%u.\n",
>              expect_messages->message, expect_messages->window, i);
> -
>
> expect_messages = NULL;
>
> /* World of Warplanes hides the window by removing WS_VISIBLE and expects Reset() to show it again. */
> device_style = GetWindowLongA(device_window, GWL_STYLE);
>
> @@ -3219,6 +3220,7 @@ static void test_wndproc(void)
>      flush_events();
>      ok(!expect_messages->message, "Expected message %#x for window %#x, but didn't receive it.\n",
>              expect_messages->message, expect_messages->window);
> -
>
> expect_messages = NULL;
>
> ok(windowpos.hwnd == device_window && !windowpos.hwndInsertAfter
>         && !windowpos.x && !windowpos.y && !windowpos.cx && !windowpos.cy

We should probably clear "expect_messages" before creating windows
that use test_proc(), instead of relying on the preceding tests to
clean things up properly. That's not to say we shouldn't clear
"expect_messages" at the "done:" label in test_wndproc() like we do
for "filter_messages" as well though. Also, mode_change_messages[] and
mode_change_messages_hidden[] should be static const. (The reason they
aren't already is probably "windowpos", but there is no particular
reason that couldn't be static as well.)