It wasn't working, I'm not really sure why.
Since #2997 we rely on video common to mark efb copies 'written'
during recording, and for old dffs we just ignore the bad texture
while playing back in the texture cache.
We actually discovered a bug while combining the two functions with
FifoRecordAnalzyer's vertex array loading code. If per-vertex
postion or texture matrices were enabled and vertex arrays in use
then the wrong data would be used to calculate the minimum/maxmium
indices, which would result in either too much or too little vertex
data being included in the dff.
So this commit also increments the dff version number, so we can
identify old broken dffs later.
- Change the path of the Sys folder to the executable's location
- Add LINUX_LOCAL_DEV flag to use relocatable version on Linux
- Add CMake definition for relocatable build
It was only implemented in OpenGL, though the option was visible in both
backends, leading to memory leaks if you enabled it in DirectX.
And it wasn't particularly useful as a debug feature as it only showed
where in the EFB the copies were taken from, not what format it was, or
what the copy was used for, or what content was in the EFB at that point
in time.
Also, it stretched the copy regions relative to the window, so the
on-screen regions don't even line up with the window unless the game used
the full EFB (some pal games) and you game image stretched to the full
window.
This 'absolutely not' a read of uninitialized memory is '100% not'
the cause of our non-deterministic behavior on the Intel NUC.
If there was a such an error, it would show up on all FifoCi
backends equally. This is 'probably' unrelated to the fact that
the Intel NUC is the only fifoci runner not running under virtualization.
If there were two commands in the buffer at once, it would only run the
first because of an error in UpdateInput.
If you sent the command "SET C" it would segfault because of a logic
issue in ParseCommand.
