Has anyone seen any anomalies running older OpenGLES1.1 apps (compiied w/ ios7 SDK) on
the newest 6S models? (presumably running iOS9.x since that was factory installed on them)
We are seeing some odd behavior but havent traced it yet (E.g. some textured polys not being drawn).
Everything fine on older devices running ios9.x.
FYI in case anyone sees this:
this seemed to be due to uninitialized (ludicrous size) Z values in a couple of glTranslate calls (even tho our app is 2D)
looks like the A9 chip might be clipping Z when it is an extreme value (no z-buffer in our app though), when in 2D mode
zeroing out the Z values in glTranslate calls seemed to have fixed the problem
Do you have any idea what caused the unitialised but ludicrously sized Z values?
If so... you might have inadvertantly found the cause of some of the probelms in Sprite Kit.
Our data was in a C++ class variable that was a struct which had .x, .y, .z ...we were not init'ing the .z of the struct but
then passing the .x .y .z values to glTranslate. So it would translate to "outer space" and get Z-clipped (my guess).
The uninit'd variable checker in the compiler may not be catching this case.
Once I fixed it, I didnt have much time to trace the why, etc... but I suspect the root cause is that class variables are not
guaranteed to be zeroed in C++. I think this is Not like ObjC/NS, where the default init] method does set class variables to zero (by default) in an NSObject.
