[PATCH 6/7] kernel32: Resolve an indirect string in time zone names.

[Akihiro Sagawa]

Signed-off-by: Akihiro Sagawa 
---
 dlls/kernel32/kernel_private.h |  2 +
 dlls/kernel32/resource.c       | 96 ++++++++++++++++++++++++++++++++++++++++++
 dlls/kernel32/time.c           | 28 ++++++++++++
 3 files changed, 126 insertions(+)

[Alexandre Julliard]

Akihiro Sagawa  writes:
> Signed-off-by: Akihiro Sagawa
>
> dlls/kernel32/kernel_private.h |  2 +
>  dlls/kernel32/resource.c       | 96 ++++++++++++++++++++++++++++++++++++++++++
>  dlls/kernel32/time.c           | 28 ++++++++++++
>  3 files changed, 126 insertions(+)
>
> diff --git a/dlls/kernel32/kernel_private.h b/dlls/kernel32/kernel_private.h
> index 09a2cad..15a4ab6 100644
> --- a/dlls/kernel32/kernel_private.h
> +++ b/dlls/kernel32/kernel_private.h
> @@ -84,4 +84,6 @@ extern void TIMEZONE_InitRegistry(void) DECLSPEC_HIDDEN;
>  /* oldconfig.c */
>  extern void convert_old_config(void) DECLSPEC_HIDDEN;
>
> +/* resource.c */
> +extern BOOL load_indirect_string(LPCWSTR src, LPWSTR dst, UINT dst_len) DECLSPEC_HIDDEN;
>  #endif
> diff --git a/dlls/kernel32/resource.c b/dlls/kernel32/resource.c
> index a04fe62..49d87a0 100644
> --- a/dlls/kernel32/resource.c
> +++ b/dlls/kernel32/resource.c
> @@ -1807,3 +1807,99 @@ BOOL WINAPI UpdateResourceA( HANDLE hUpdate, LPCSTR lpType, LPCSTR lpName,
>      if(!IS_INTRESOURCE(lpName)) RtlFreeUnicodeString(&NameW);
>      return ret;
>  }
> -
>
> +/******************************************************************************
> -
> - load_string [Internal]
> -
> -
> -
> - This is basically a copy of advapi32/reg.c's load_string. Necessary to
> -
> - avoid importing user32, which is higher level than kernel32. Helper for
> -
> - following load_indirect_string.
> - */
>
> +static int load_string(HINSTANCE hModule, UINT resId, LPWSTR pwszBuffer, INT cMaxChars)
> +{

You should probably use RegLoadMUIStringW instead of duplicating this
code.

[Akihiro Sagawa]

On Wed, 05 Dec 2018 19:45:20 +0100, Alexandre Julliard wrote:
> > @@ -1807,3 +1807,99 @@ BOOL WINAPI UpdateResourceA( HANDLE hUpdate, LPCSTR lpType, LPCSTR lpName,
> >      if(!IS_INTRESOURCE(lpName)) RtlFreeUnicodeString(&NameW);
> >      return ret;
> >  }
> > -
> >
> > +/******************************************************************************
> > -
> > - load_string [Internal]
> > -
> > -
> > -
> > - This is basically a copy of advapi32/reg.c's load_string. Necessary to
> > -
> > - avoid importing user32, which is higher level than kernel32. Helper for
> > -
> > - following load_indirect_string.
> > - */
> >
> > +static int load_string(HINSTANCE hModule, UINT resId, LPWSTR pwszBuffer, INT cMaxChars)
> > +{
>
> You should probably use RegLoadMUIStringW instead of duplicating this
> code.

Is it safe to call advapi32 functions from kernel32?
From my point of view, kernel32 doesn't import advapi32 functions at
this point.

Akihiro Sagawa

[Alexandre Julliard]

Akihiro Sagawa  writes:
> On Wed, 05 Dec 2018 19:45:20 +0100, Alexandre Julliard wrote:
> > > @@ -1807,3 +1807,99 @@ BOOL WINAPI UpdateResourceA( HANDLE hUpdate, LPCSTR lpType, LPCSTR lpName,
> > >      if(!IS_INTRESOURCE(lpName)) RtlFreeUnicodeString(&NameW);
> > >      return ret;
> > >  }
> > > -
> > >
> > > +/******************************************************************************
> > > -
> > > - load_string [Internal]
> > > -
> > > -
> > > -
> > > - This is basically a copy of advapi32/reg.c's load_string. Necessary to
> > > -
> > > - avoid importing user32, which is higher level than kernel32. Helper for
> > > -
> > > - following load_indirect_string.
> > > - */
> > >
> > > +static int load_string(HINSTANCE hModule, UINT resId, LPWSTR pwszBuffer, INT cMaxChars)
> > > +{
> >
> > You should probably use RegLoadMUIStringW instead of duplicating this
> > code.
>
> Is it safe to call advapi32 functions from kernel32?
> From my point of view, kernel32 doesn't import advapi32 functions at
> this point.

It has been moved to kernelbase.dll in recent versions, we should do the
same, then it could be used from anywhere. In the meantime I think
loading it explicitly from advapi32 would be acceptable.

[Akihiro Sagawa]

On Thu, 06 Dec 2018 13:07:12 +0100, Alexandre Julliard wrote:
> Akihiro Sagawa  writes:
> > On Wed, 05 Dec 2018 19:45:20 +0100, Alexandre Julliard wrote:
> > > > @@ -1807,3 +1807,99 @@ BOOL WINAPI UpdateResourceA( HANDLE hUpdate, LPCSTR lpType, LPCSTR lpName,
> > > >      if(!IS_INTRESOURCE(lpName)) RtlFreeUnicodeString(&NameW);
> > > >      return ret;
> > > >  }
> > > > -
> > > >
> > > > +/******************************************************************************
> > > > -
> > > > - load_string [Internal]
> > > > -
> > > > -
> > > > -
> > > > - This is basically a copy of advapi32/reg.c's load_string. Necessary to
> > > > -
> > > > - avoid importing user32, which is higher level than kernel32. Helper for
> > > > -
> > > > - following load_indirect_string.
> > > > - */
> > > >
> > > > +static int load_string(HINSTANCE hModule, UINT resId, LPWSTR pwszBuffer, INT cMaxChars)
> > > > +{
> > >
> > > You should probably use RegLoadMUIStringW instead of duplicating this
> > > code.
> >
> > Is it safe to call advapi32 functions from kernel32?
> > From my point of view, kernel32 doesn't import advapi32 functions at
> > this point.
>
> It has been moved to kernelbase.dll in recent versions, we should do the
> same, then it could be used from anywhere. In the meantime I think
> loading it explicitly from advapi32 would be acceptable.

That's sounds good. I'll prepare revised patch.

Akihiro Sagawa