Closed jamiebullock closed 3 years ago
Fixed in hg commit 45c248b6bab8! Only a little over three years, not bad. Ahem.
Of course this could happen only in debug mode (the profiler class is eliminated in release builds). It was caused by a map being accessed without a mutex when the profiler saved its findings on exiting scope, and of course this can happen more than once simultaneously in multi-channel multi-threaded mode.
This happened to me once, and I certainly can't reproduce it. It's possible my program is at fault and not
RubberBand
, however I'm posting this here in case there's some obviously fixable issue with RB from looking at the call stack.Basically, I was stress testing my app under the debugger by "clicking around wildly" and this happened:
Thread 110: EXC_BAD_ACCESS (code=1, address=0x18)