Re: [PATCH v14 0/6] MR1445: kernelbase: Some refactoring of the open_key() and create_key() functions.

[Sven Baars (＠sbaars)]

On Thu Mar  9 09:30:49 2023 +0000, Alexandre Julliard wrote:
> The last patch is introducing a warning:
> i686-w64-mingw32-gcc -c -o dlls/kernelbase/i386-windows/registry.o
> dlls/kernelbase/registry.c -Idlls/kernelbase -Iinclude \
>   -Iinclude/msvcrt -D_MSVCR_VER=0 -D__WINESRC__ -DWINBASEAPI=
> -DWINADVAPI= -DWINUSERAPI= \
>   -DWINSHLWAPI= -D__WINE_PE_BUILD -Wall -fno-strict-aliasing
> -Wdeclaration-after-statement \
>   -Wempty-body -Wignored-qualifiers -Winit-self -Wno-packed-not-aligned
> -Wshift-overflow=2 \
>   -Wstrict-prototypes -Wtype-limits -Wunused-but-set-parameter -Wvla
> -Wwrite-strings -Wpointer-arith \
>   -Wlogical-op -Wabsolute-value -fno-omit-frame-pointer -gdwarf-4
> -Werror -g -O2 -fno-diagnostics-show-caret
> dlls/kernelbase/registry.c: In function ‘create_key’:
> dlls/kernelbase/registry.c:262:26: error: storing the address of local
> variable ‘str’ in ‘*attr.ObjectName’ [-Werror=dangling-pointer=]
> dlls/kernelbase/registry.c:259:24: note: ‘str’ declared here
> dlls/kernelbase/registry.c:259:24: note: ‘attr’ declared here
> cc1: all warnings being treated as errors
> make: *** [Makefile:68674: dlls/kernelbase/i386-windows/registry.o]
> Error 1

That's very strange since I didn't modify these lines of code, so I'd expect the warning to also be there before any of my patches. I also don't see the warning locally.

I do get rid of this code later in my patch series. Fixing it earlier will create a mess since attr is currently used everywhere, but I guess it could be done. Since I don't see the warning it may be difficult to do without introducing it again in intermediate commits.