[PATCH] d3dx11_43: Implement D3DX11GetImageInfoFromMemory

[Matteo Bruni]

On Fri, Jul 30, 2021 at 10:15 AM Alistair Leslie-Hughes
<leslie_alistair at hotmail.com> wrote:
>
> Wine-bug: https://bugs.winehq.org/show_bug.cgi?id=50210
>
> Signed-off-by: Alistair Leslie-Hughes <leslie_alistair at hotmail.com>
> ---
>  dlls/d3dx11_42/Makefile.in    |   1 +
>  dlls/d3dx11_43/Makefile.in    |   1 +
>  dlls/d3dx11_43/main.c         |   9 -
>  dlls/d3dx11_43/tests/d3dx11.c | 668 ++++++++++++++++++++++++++++++++++
>  dlls/d3dx11_43/texture.c      | 176 +++++++++
>  5 files changed, 846 insertions(+), 9 deletions(-)

This looks like the perfect occasion for reworking the d3dx DLLs to
avoid duplicating the whole thing, along the lines of
https://www.winehq.org/pipermail/wine-devel/2021-June/189228.html and
followups.
Specifically I'd introduce a wined3dx DLL, then move the bulk of
D3DX10GetImageInfoFromMemory() over there, keeping only the entrypoint
(and any d3dx10-specific part, if they exist) in d3dx10. Then this
patch would just introduce a wined3dx import to d3dx11 and the
D3DX11GetImageInfoFromMemory() implementation will mostly call into
wined3dx for the actual functionality. The tests should probably stay
duplicated as in the current version of this patch. Finally, I'd wait
and see what happens :P

I'm sorry your patch has the honor of being the first to come after I
made up my mind on the d3dx DLLs rework (and I do recall that this is
the second time you get the lucky pick, after the forwards ->
PARENTSRC transition from a few years back). I'd be glad if you want
to try your hand at this, otherwise I'll do it myself, hopefully in
the near future.