On 9/3/21 12:36 AM, Matteo Bruni wrote:
On Wed, Sep 1, 2021 at 7:13 PM Zebediah Figura zfigura@codeweavers.com wrote:
Signed-off-by: Zebediah Figura zfigura@codeweavers.com
dlls/d3dcompiler_43/Makefile.in | 1 + dlls/d3dcompiler_43/tests/Makefile.in | 2 +- dlls/d3dcompiler_43/tests/asm.c | 80 +++++++------------------- dlls/d3dcompiler_43/tests/hlsl_d3d11.c | 32 ++--------- dlls/d3dcompiler_43/tests/hlsl_d3d9.c | 78 ++++++++----------------- dlls/d3dcompiler_43/tests/reflection.c | 52 ++++------------- dlls/d3dcompiler_47/tests/Makefile.in | 2 +- 7 files changed, 62 insertions(+), 185 deletions(-)
This is effectively a revert of a8496091931c65d7a2437281195b388498fb4533. For reference, the original emails for that patch series from Jacek are https://www.winehq.org/pipermail/wine-devel/2018-July/129711.html, https://www.winehq.org/pipermail/wine-devel/2018-July/129712.html.
I don't think we care about winelib dxvk at this point and, regardless, we probably don't have any reason for loading d3dcompiler_xx.dll dynamically anymore. AFAIU with this we're basically reintroducing a (versioned) importlib for d3dcompiler_43.dll, which seems fine to me.
So, I guess I'm asking for anybody to confirm that this seems all correct, or let me know if not. I'm probably going to sign-off on this in a day or so if I don't hear anything.
Now that we use PARENTSRC for tests, dynamic loading indeed doesn't make much sense. We still want to provide an importlib without a version in its name, but as long as d3dcompiler_47 uses a non-versioned name, that's fine.
Jacek