[PATCH v2 0/2] MR8031: winex11: Ensure initialized thread data in X11DRV_get_ic.

[Etaash Mathamsetty (＠etaash.mathamsetty)]

Intends to replace: https://gitlab.winehq.org/wine/wine-staging/-/blob/master/patches/winex11-im...

This patch has been in proton for almost 2 years now, never realized the same code existed upstream, lol

--
  v2: win32u/tests: Add test for multi-thread NotifyIME.
      winex11: Ensure initialized thread data in X11DRV_get_ic.
 
https://gitlab.winehq.org/wine/wine/-/merge_requests/8031

[PATCH v2 1/2] winex11: Ensure initialized thread data in X11DRV_get_ic.

[Etaash Mathamsetty]

From: Etaash Mathamsetty
 <45927311+Etaash-mathamsetty(a)users.noreply.github.com>

---
 dlls/winex11.drv/xim.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/dlls/winex11.drv/xim.c b/dlls/winex11.drv/xim.c
index ef09356fbb0..4c93c73cc23 100644
--- a/dlls/winex11.drv/xim.c
+++ b/dlls/winex11.drv/xim.c
@@ -509,7 +509,7 @@ XIC X11DRV_get_ic( HWND hwnd )
     XIC ret;
 
     if (!(data = get_win_data( hwnd ))) return 0;
-    x11drv_thread_data()->last_xic_hwnd = hwnd;
+    x11drv_init_thread_data()->last_xic_hwnd = hwnd;
     if (!(ret = data->xic) && (xim = x11drv_thread_data()->xim))
         ret = data->xic = xic_create( xim, hwnd, data->whole_window );
     release_win_data( data );
-- 
GitLab


https://gitlab.winehq.org/wine/wine/-/merge_requests/8031

[PATCH v2 2/2] win32u/tests: Add test for multi-thread NotifyIME.

[Etaash Mathamsetty]

From: Etaash Mathamsetty
 <45927311+Etaash-mathamsetty(a)users.noreply.github.com>

---
 dlls/win32u/tests/win32u.c | 28 ++++++++++++++++++++++++++++
 1 file changed, 28 insertions(+)

diff --git a/dlls/win32u/tests/win32u.c b/dlls/win32u/tests/win32u.c
index acbc5bc3034..41b39f03714 100644
--- a/dlls/win32u/tests/win32u.c
+++ b/dlls/win32u/tests/win32u.c
@@ -2862,6 +2862,33 @@ static void test_NtUserRegisterWindowMessage(void)
     ok( !wcscmp( buf, L"#0xabc" ), "buf = %s\n", debugstr_w(buf) );
 }
 
+static CALLBACK DWORD test_NtUserNotifyIMEStatus_thread(void *user_data)
+{
+    HWND hwnd = user_data;
+
+    /* validate that the user driver does not crash */
+    NtUserNotifyIMEStatus(hwnd, 0);
+
+    return 0;
+}
+
+static void test_NtUserNotifyIMEStatus(void)
+{
+    HWND hwnd;
+    HANDLE thread;
+
+    winetest_push_context("notifyime");
+
+    hwnd = CreateWindowW( L"static", L"static", WS_OVERLAPPEDWINDOW | WS_HSCROLL | WS_VSCROLL | WS_CLIPSIBLINGS, CW_USEDEFAULT, 0, CW_USEDEFAULT, 0, 0, 0, NULL, 0 );
+    thread = CreateThread( NULL, 0, test_NtUserNotifyIMEStatus_thread, hwnd, 0, NULL );
+
+    WaitForSingleObject(thread, INFINITE);
+    CloseHandle(thread);
+    DestroyWindow(hwnd);
+
+    winetest_pop_context();
+}
+
 START_TEST(win32u)
 {
     char **argv;
@@ -2891,6 +2918,7 @@ START_TEST(win32u)
         return;
     }
 
+    test_NtUserNotifyIMEStatus();
     test_NtUserEnumDisplayDevices();
     test_window_props();
     test_class();
-- 
GitLab

https://gitlab.winehq.org/wine/wine/-/merge_requests/8031

[Etaash Mathamsetty (＠etaash.mathamsetty)]

On Tue Jun 17 01:44:39 2025 +0000, Etaash Mathamsetty wrote:
> Yes, will do
> added test

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/8031#note_106803

[Elizabeth Figura (＠zfigura)]

Is that really what the application is doing?

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/8031#note_106812

[Etaash Mathamsetty (＠etaash.mathamsetty)]

On Tue Jun 17 04:15:27 2025 +0000, Elizabeth Figura wrote:
> Is that really what the application is doing?
> Probably not, but it was the easiest way that i could come up with that crashes winex11 but won't crash windows. Maybe the application is doing this exact thing though, I'll have to check

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/8031#note_106813

[Elizabeth Figura (＠zfigura)]

On Tue Jun 17 04:16:51 2025 +0000, Etaash Mathamsetty wrote:
> Probably not, but it was the easiest way that i could come up with that
> crashes winex11 but won't crash windows. Maybe the application is doing
> this exact thing though, I'll have to check
> In fact I'm pretty sure I remember this bug existing even before we had NtUserNotifyIMEStatus().

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/8031#note_107674