* It was using the last bound render target, which could be empty if
state was reset before presentation. Instead we use the copy target
listed in the present drawcall as the actual swap buffer presented to.
* DDS will always save the format entirely literally, and the float
formats do not support alpha processing currently.
* TGA and PNG will either discard the alpha and write entirely opaque,
or they'll include it in the file. Blending to colour or checkerboard
is not supported.
* Using OS-specific append writing functions instead of simple fopen/
fwrite/fclose we can guarantee that multiple processes writing to the
same log file won't trash or interleave their log output.
* To make the logs readable the PID is now included along with the
project/timestamp.
* After that, we make sure that when we launch a program we set the
output log to the same file as the parent UI. This keeps all the logs
together and removes user confusion over which logfile is needed.
* We also move all of RenderDoc's temporary files under a RenderDoc
folder in the temp folder, instead of scattering the files in the root
* This should at least support the swapping of the current pipeline
state but it probably won't support all the weird 'emulate different
D3D versions' stuff.
* Without this, a cast from WrappedID3D11DeviceContext to
ID3D11DeviceContext3 will not properly adjust the vtables so the
returned pointer will be invalid.
* During the frame we don't update our internal descriptor tracking, we
just record the updates on the CPU timeline.
* However, it's possible that one of those copies or writes references a
resource that is persistently mapped, so we need to know about that so
we can flush it at the next execute.
* If we're capturing, the frame references will keep the resource alive
and the application has to at least keep resources alive until after
GPU execution is finished.
baldurk