Re: [PATCH v6 0/6] MR9977: wineandroid: Fix build issues and early runtime regressions on modern Android toolchains

[Twaik Yont (＠twaik)]

On Thu Feb  5 07:05:30 2026 +0000, Alexandre Julliard wrote:
> I merged most of these, the remaining ones will need some more thought.
> Thanks for looking at these patches.

About the remaining commits:

1) `ntdll: Include <strings.h> for ffs() declaration`

This change should be harmless. ffs() is already implicitly pulled in through other headers, which is why this issue usually does not surface. However, with stricter C99+ toolchains (e.g. clang used in Termux / Android environments), implicit function declarations are rejected and this results in build failures like:

```
/home/twaik/.termux-build/wine-stable/src/dlls/ntdll/unix/unix_private.h:660:26: error: call to undeclared function 'ffs'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
  660 |         idx = idx * 32 + ffs( ~ldt_bitmap[idx] ) - 1;
      |                          ^
1 error generated.
make: *** [Makefile:250029: dlls/ntdll/unix/loadorder.o] Error 1
```

Including <strings.h> only makes the dependency explicit and does not change behavior on platforms where it was already indirectly included.

---

2) `winepulse: Do not use pthread mutex attributes on Android`

This change is Android-only and does not affect other platforms.

Android only provides pthread_mutexattr_setprotocol() starting from API 28 and does not support robust mutexes at all. Attempting to use these attributes prevents winepulse from building in Android environments.

Falling back to a default mutex on Android keeps existing behavior elsewhere unchanged. This is particularly useful for downstream Android-based projects (compiling Wine with Android NDK: Termux, Winlator-bionic, MiceWine, etc.) which rely on PulseAudio for sound output and currently have no alternative backend.

---

3) `wineandroid: force --rosegment for wine-preloader`

This change is inside the Android-specific configure section and should not affect other platforms.

Without --rosegment the linker may generate a wine-preloader ELF roughly 2GB in size due to segment layout. The patch only appends the flag to WINEPRELOADER_LDFLAGS and only when the toolchain supports it, so it should be safe and fully isolated to Android builds.

---

Should I make separate merge request for these changes?

-- 
https://gitlab.winehq.org/wine/wine/-/merge_requests/9977#note_128846