[PATCH 0/1] MR10933: kernelbase: Add null check for str in OutputDebugStringW

[Matthias Gorzellik (＠gotzl)]

The following code crashes wine but seems to be OK in Windows:
```
#include <windows.h>
int main() {
    OutputDebugStringW(NULL);
    return 0;
}
```
The MR aligns behavior of `OutputDebugStringW` to Windows by adding a null check and bailing out of `OutputDebugStringW`.

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/10933

[PATCH 1/1] kernelbase: Add null check for str in OutputDebugStringW

[Matthias Gorzellik]

From: Matthias Gorzellik <[e-mail]>

---
 dlls/kernelbase/debug.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/dlls/kernelbase/debug.c b/dlls/kernelbase/debug.c
index a600b6b59c9..8656859b7bb 100644
--- a/dlls/kernelbase/debug.c
+++ b/dlls/kernelbase/debug.c
@@ -281,6 +281,7 @@ void WINAPI DECLSPEC_HOTPATCH OutputDebugStringW( LPCWSTR str )
     STRING strA;
 
     WARN( "%s\n", debugstr_w(str) );
+    if (!str) return;
 
     RtlInitUnicodeString( &strW, str );
     if (!RtlUnicodeStringToAnsiString( &strA, &strW, TRUE ))
-- 
GitLab

https://gitlab.winehq.org/wine/wine/-/merge_requests/10933

[Jinoh Kang (＠iamahuman)]

Jinoh Kang (@iamahuman) commented about dlls/kernelbase/debug.c:
> STRING strA;
>
> WARN( "%s\n", debugstr_w(str) );
> +    if (!str) return;

This would need tests. Is NULL string skipped entirely? Or coerced to empty (but valud) string and passed to the debugger regardless.

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/10933#note_140358