- Dec 06, 2024
-
-
Inspired by a patch by Rémi Bernon. This allows us to clear the buffer a bit earlier than waiting for an underrun does, and avoids the spurious FIXME message in that case. It also unifies code paths a bit. This change also correctly clears the buffer for 8-bit formats.
-
As tests show, the DirectSound renderer will return immediately from Receive() and EndOfStream(), holding the sample to be rendered arbitrarily later. This is also necessary to properly clear the buffer if the renderer is starved for a long period of time.
-
-
-
-
Ignoring their start time. This matches native behaviour.
-
-
-
This might happen if the window manager is restarted or changed while a Wine window is opened. We should support that use case properly.
-
-
-
Native Windows robustly handles unexpected shutdown of a media source used in a session, and always sends MESessionStarted with an error status, but if the native source is contained within a wrapper source object, MESessionStarted is never sent even if BeginGetEvent() fails with MF_E_SHUTDOWN, which should be easily handled. This implies that Windows uses a private interface for shutdown notification, and its proper functioning is disrupted when the source is wrapped.
-
-
-
-
Windows robustly handles asynchronous shutdown of a media source used in a session.
-
-
-
Also works around a Rosetta 2 bug, where code execution can remove VM_PROT_WRITE permissions from a page.
-
-
-
-
-
-
-
- Dec 05, 2024
-
-
-
-
-
-
-
This avoids the server piling up closed_fd objects when several delete are called on the same path. Signed-off-by:
Eric Pouech <epouech@codeweavers.com>
-
-
-
ProcessMessage() can return MF_E_TRANSFORM_TYPE_NOT_SET if necessary.
-
Other methods for the video object return DMO_E_TYPE_NOT_SET if wg_transform is null. A race condition in Nioh can result in Flush() being called when the transform is not set.
-
-
-
-
-
-