Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=55978

Starcraft Remastered is looking up for the GPU physical memory size to
check for minimum requirements to enable RealTime Lighting.

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=46624Signed-off-by: Jon Doron <arilou@gmail.com>

If explicit_handle is defined in the idl file, c/s use explicit handles,
then an explicit handle must be passed in to the server-side  interface

wg_transform_get_status() returns bool, not HRESULT.

video_format can be leaked when encountering an invalid WMV format.

Signed-off-by: Gabriel Ivăncescu <gabrielopcode@gmail.com>

These methods are not called from any of our objects that hold ref to the
inner window only, and since we return outer windows to external callers
now, they *must* have a ref to the outer window, which effectively means
these cannot be NULL. And some other places didn't check for NULL either
(e.g. get_parent).
Signed-off-by: Gabriel Ivăncescu <gabrielopcode@gmail.com>

Signed-off-by: Gabriel Ivăncescu <gabrielopcode@gmail.com>

This would have crashed on detached windows, and checking the navigation_start
time has the same effect except it's decoupled from the outer window.
Signed-off-by: Gabriel Ivăncescu <gabrielopcode@gmail.com>

We often use "native" to talk about Windows "native" code, which is
confusing here.

pEntry could be freed after cache_container_unlock_index(pContainer, pHeader).

process_entry->szExeFile has length of MAX_PATH but info->image_name only has 32.

*pszTmpFileName will be freed with free() but it's allocated with HeapAlloc().

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=55591

Fixes: 370b8299

As with the previous commit, this is clearer and avoids a state ordering assumption.

wined3d: Use context->stream_info.position_transformed instead of context->last_was_rhw in state handlers.

Use of last_was_rhw in these state handlers is, firstly, confusing to read.
The variable is actually supposed to be set for *this* draw, not the last one,
and state handlers indeed expect that, but that fact isn't obvious. Reading from
context->stream_info instead is clearer.

Secondly, since last_was_rhw is set by the STATE_VDECL handler, any other state
depending on it (including some, though not all, of the sites changed here) thus
has an implicit state handler ordering assumption. This is, needless to say,
fragile. context->stream_info is computed before state handlers, so that
dependency is a bit less fragile.

The comment was originally added in 4b831a5d.
At that point, it referred only to non-FFP shader constants. Subsequently
38934fe7 moved bumpenv constant loading to the
misc pipeline, apparently out of concerns that bumpenv constants might interact
badly with shader constants.

It's not clear whether those concerns were ever justified, and normal shader
constants are now sufficiently untangled from each other anyway. However,
bumpenv constants, like e.g. color keys, affect both FFP and non-FFP pipelines,
so it makes the most sense for them to remain in the misc state table anyway.

Signed-off-by: Nikolay Sivov <nsivov@codeweavers.com>

Signed-off-by: Nikolay Sivov <nsivov@codeweavers.com>

Signed-off-by: Nikolay Sivov <nsivov@codeweavers.com>

Signed-off-by: Nikolay Sivov <nsivov@codeweavers.com>