Re: [PATCH 2/4] shell32: Implement IShellImageData::Decode() when created from file path

[Alexandre Julliard]

Nikolay Sivov <nsivov(a)codeweavers.com> writes:
> @@ -29,9 +29,17 @@ cpp_quote("#endif")
>
> cpp_quote("#ifndef _GDIPLUSHEADERS_H")
>  typedef BYTE EncoderParameters;
> +cpp_quote("#endif")
> +
> +/* Wine does not define GDI+ class wrappers */
> +cpp_quote("#if defined(__WINESRC__) || !defined(_GDIPLUSHEADERS_H)")
>  typedef BYTE Image;
>  cpp_quote("#endif")

We don't have C++ code in the Wine source, so I don't see why you'd need
that.

-- 
Alexandre Julliard
julliard(a)winehq.org

[PATCH 2/4] shell32: Implement IShellImageData::Decode() when created from file path

[Nikolay Sivov]

On 07.02.2017 17:44, Alexandre Julliard wrote:
> Nikolay Sivov <nsivov(a)codeweavers.com> writes:
> > @@ -29,9 +29,17 @@ cpp_quote("#endif")
> >
> > cpp_quote("#ifndef _GDIPLUSHEADERS_H")
> >  typedef BYTE EncoderParameters;
> > +cpp_quote("#endif")
> > +
> > +/* Wine does not define GDI+ class wrappers */
> > +cpp_quote("#if defined(__WINESRC__) || !defined(_GDIPLUSHEADERS_H)")
> >  typedef BYTE Image;
> >  cpp_quote("#endif")
>
> We don't have C++ code in the Wine source, so I don't see why you'd need
> that.

We do have some class definitions in gdiplus headers, like in
gdiplusimaging.h. My motivation was to have a header compatible with SDK
one in terms of these include guards. I understand that it won't be
possible to use Image class in its C++ form from shell32, luckily I had
no need for that to fix the bug.

Do you mean it's enough to typedef Image to BYTE uncoditionaly?

[PATCH 2/4] shell32: Implement IShellImageData::Decode() when created from file path

[Alexandre Julliard]

Nikolay Sivov <bunglehead(a)gmail.com> writes:
> On 07.02.2017 17:44, Alexandre Julliard wrote:
> > Nikolay Sivov <nsivov(a)codeweavers.com> writes:
> > > @@ -29,9 +29,17 @@ cpp_quote("#endif")
> > >
> > > cpp_quote("#ifndef _GDIPLUSHEADERS_H")
> > >  typedef BYTE EncoderParameters;
> > > +cpp_quote("#endif")
> > > +
> > > +/* Wine does not define GDI+ class wrappers */
> > > +cpp_quote("#if defined(__WINESRC__) || !defined(_GDIPLUSHEADERS_H)")
> > >  typedef BYTE Image;
> > >  cpp_quote("#endif")
> >
> > We don't have C++ code in the Wine source, so I don't see why you'd need
> > that.
>
> We do have some class definitions in gdiplus headers, like in
> gdiplusimaging.h. My motivation was to have a header compatible with SDK
> one in terms of these include guards. I understand that it won't be
> possible to use Image class in its C++ form from shell32, luckily I had
> no need for that to fix the bug.
>
> Do you mean it's enough to typedef Image to BYTE uncoditionaly?

I mean the existing ifndef _GDIPLUSHEADERS_H should be enough, because
gdiplusheaders.h is a C++ header that can't be used from C code.

-- 
Alexandre Julliard
julliard(a)winehq.org

[PATCH 2/4] shell32: Implement IShellImageData::Decode() when created from file path

[Nikolay Sivov]

On 07.02.2017 18:20, Alexandre Julliard wrote:
> Nikolay Sivov <bunglehead(a)gmail.com> writes:
> > On 07.02.2017 17:44, Alexandre Julliard wrote:
> > > Nikolay Sivov <nsivov(a)codeweavers.com> writes:
> > > > @@ -29,9 +29,17 @@ cpp_quote("#endif")
> > > >
> > > > cpp_quote("#ifndef _GDIPLUSHEADERS_H")
> > > >  typedef BYTE EncoderParameters;
> > > > +cpp_quote("#endif")
> > > > +
> > > > +/* Wine does not define GDI+ class wrappers */
> > > > +cpp_quote("#if defined(__WINESRC__) || !defined(_GDIPLUSHEADERS_H)")
> > > >  typedef BYTE Image;
> > > >  cpp_quote("#endif")
> > >
> > > We don't have C++ code in the Wine source, so I don't see why you'd need
> > > that.
> >
> > We do have some class definitions in gdiplus headers, like in
> > gdiplusimaging.h. My motivation was to have a header compatible with SDK
> > one in terms of these include guards. I understand that it won't be
> > possible to use Image class in its C++ form from shell32, luckily I had
> > no need for that to fix the bug.
> >
> > Do you mean it's enough to typedef Image to BYTE uncoditionaly?
>
> I mean the existing ifndef _GDIPLUSHEADERS_H should be enough, because
> gdiplusheaders.h is a C++ header that can't be used from C code.

This won't work for EncoderParameters, that we define for C code too as
a bare structure, while SDK only has it as a class. So it's probably
already not very compatible. I guess I can change first guard to
_GDIPLUSIMAGING_H, that will work for both SDK and Wine, and keep second
as is.

[PATCH 2/4] shell32: Implement IShellImageData::Decode() when created from file path

[Alexandre Julliard]

Nikolay Sivov <bunglehead(a)gmail.com> writes:
> On 07.02.2017 18:20, Alexandre Julliard wrote:
> > Nikolay Sivov <bunglehead(a)gmail.com> writes:
> > > On 07.02.2017 17:44, Alexandre Julliard wrote:
> > > > Nikolay Sivov <nsivov(a)codeweavers.com> writes:
> > > > > @@ -29,9 +29,17 @@ cpp_quote("#endif")
> > > > >
> > > > > cpp_quote("#ifndef _GDIPLUSHEADERS_H")
> > > > >  typedef BYTE EncoderParameters;
> > > > > +cpp_quote("#endif")
> > > > > +
> > > > > +/* Wine does not define GDI+ class wrappers */
> > > > > +cpp_quote("#if defined(__WINESRC__) || !defined(_GDIPLUSHEADERS_H)")
> > > > >  typedef BYTE Image;
> > > > >  cpp_quote("#endif")
> > > >
> > > > We don't have C++ code in the Wine source, so I don't see why you'd need
> > > > that.
> > >
> > > We do have some class definitions in gdiplus headers, like in
> > > gdiplusimaging.h. My motivation was to have a header compatible with SDK
> > > one in terms of these include guards. I understand that it won't be
> > > possible to use Image class in its C++ form from shell32, luckily I had
> > > no need for that to fix the bug.
> > >
> > > Do you mean it's enough to typedef Image to BYTE uncoditionaly?
> >
> > I mean the existing ifndef _GDIPLUSHEADERS_H should be enough, because
> > gdiplusheaders.h is a C++ header that can't be used from C code.
>
> This won't work for EncoderParameters, that we define for C code too as
> a bare structure, while SDK only has it as a class. So it's probably
> already not very compatible. I guess I can change first guard to
> _GDIPLUSIMAGING_H, that will work for both SDK and Wine, and keep second
> as is.

That sounds more correct. We can't be 100% compatible, but I don't think
we should be using __WINESRC__ to fix such issues; the workarounds we
apply should also be available to Winelib apps.

-- 
Alexandre Julliard
julliard(a)winehq.org