RFC: Disabling MinGW support in Wine packages (e.g. Debian)

[Francois Gouget]

Debian's Wine packages build with --without-mingw.

See debian/rules in:
http://deb.debian.org/debian/pool/main/w/wine/wine_5.0.3-3.debian.tar.xz

If I remember correctly this increases the chance of not being able to 
run applications that expect specific binary layouts for Windows 
function prologues, i.e. mostly those with copy-prevention / anti-cheat 
schemes.

If so packages that needlessly disable MinGW are doing their users a 
disservice and are not presenting Wine in its best light.

Do I remember correctly?
Should packagers be advised not to compile without MinGW?

[Alexandre Julliard]

Francois Gouget  writes:
> Debian's Wine packages build with --without-mingw.
>
> See debian/rules in:
> http://deb.debian.org/debian/pool/main/w/wine/wine_5.0.3-3.debian.tar.xz
>
> If I remember correctly this increases the chance of not being able to 
> run applications that expect specific binary layouts for Windows 
> function prologues, i.e. mostly those with copy-prevention / anti-cheat 
> schemes.
>
> If so packages that needlessly disable MinGW are doing their users a 
> disservice and are not presenting Wine in its best light.
>
> Do I remember correctly?
> Should packagers be advised not to compile without MinGW?

It's less important for an older release like 5.0, but yes, in general
disabling MinGW is a bad idea.

[Sveinar Søpler]

The upcoming packages for debian currently in unstable:
https://packages.debian.org/unstable/wine-development
Seems to be built with mingw on x86/amd64

# build with mingw on intel architectures
ifeq ($(DEB_HOST_ARCH), i386)
CONFLAGS+=--with-mingw
endif
ifeq ($(DEB_HOST_ARCH), amd64)
CONFLAGS+=--with-mingw
endif

Sveinar

On 28.07.2021 21:50, Alexandre Julliard wrote:
> Francois Gouget  writes:
> > Debian's Wine packages build with --without-mingw.
> >
> > See debian/rules in:
> > http://deb.debian.org/debian/pool/main/w/wine/wine_5.0.3-3.debian.tar.xz
> >
> > If I remember correctly this increases the chance of not being able to
> > run applications that expect specific binary layouts for Windows
> > function prologues, i.e. mostly those with copy-prevention / anti-cheat
> > schemes.
> >
> > If so packages that needlessly disable MinGW are doing their users a
> > disservice and are not presenting Wine in its best light.
> >
> > Do I remember correctly?
> > Should packagers be advised not to compile without MinGW?
>
> It's less important for an older release like 5.0, but yes, in general
> disabling MinGW is a bad idea.