musl itself expects to be configured to compile with either
-ffloat-store or -fexcess-precision=standard - but when imported
into Wine, those flags aren't used.
This seems to be essential for getting reasonable precision
from some math functions such as exp2() - without the expected
precision truncation, the output value of exp2() can be off by
as much as 0.2% in some cases.
As Wine doesn't build the musl sources with those flags, use
volatile to force storing/reloading floats in order to limit
their intermediate precision, as musl expects. Only do this on
i386 where this is known be required.
This fixes https://bugs.winehq.org/show_bug.cgi?id=56372.
Apparently this issue only appears when compiled with GCC; with
Clang, this already works as expected.
Signed-off-by: Martin Storsjö <martin(a)martin.st>
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/5276
The problem is that registry save is a very heavy operation (scheduled each 30sec in wineserver) during which server doesn't process any requests and the whole prefix is stalled for the duration of the operation.
For some reference, the process takes from 50-100ms here up to 1-1.5sec with default initial registry (after some registry modifications which trigger actual registry flush), depending on the filesystem type and state (as huge time may be spent in file close / rename). With the same registry after this patchset the server part (flush_key returning the whole registry data) is taking ~4-5mcs, measured from the client side so that already includes data transfer.
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/3124
Some cleanup for expressions, sanitizing the structure a bit especially for abstract operations with multiple args.
I also found out about named bison variables, implicit (as `$token` referring to `token`), or explicit (as `$name` referring to `token[name]`). I think it would make the rules much more readable so I started using them here. My plan is to add them progressively to the rules, unless there's some good reason not to (like idk how portable they are, if that's something we should be worried about).
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/4201
I'm not completely sure about the mechanism, but I think it's simple enough to consider having that upstream now. This shows at least how we can leverage win32u surface changes to decide to switch surfaces on/off-screen and fallback to manual blitting.
Having the surfaces on-screen makes sure they are presented as efficiently as possible, having them off-screen we use GDI blit to indirectly call XCopyArea and this will be suboptimal, probably with visible tearing, but hopefully not too common.
--
v4: win32u: Use GDI blit to implement partial or other process presentation.
winex11: Return an offscreen HDC from vulkan_surface_detach.
win32u: Pass a HDC parameter to vulkan_surface_detach.
winex11: Create a window surface even when there is client window.
winex11: Also attach child client windows to their toplevel window.
win32u: Make sure vulkan windows have a pixel format selected.
win32u: Detach vulkan surfaces that aren't fully visible.
win32u: Detach offscreen, child or vulkan surfaces for another process.
win32u: Move vulkan surfaces to their new parent when reparenting.
win32u: Introduce a new vulkan offscreen surfaces list.
https://gitlab.winehq.org/wine/wine/-/merge_requests/5573