Clearly, this is not an issue valgrind is meant to detect (like invalidmemory access, etc.) but an application issue. I've tried reproducing itwithout valgrind, but it only ever happens with valgrind - my theory isit's some sort of race condition, and valgrind changes the timing in away ...
* being logically empty, because we don't attempt to detect pfree'ing the * last active chunk. */typedefstructAllocSetContext{MemoryContextData header;/* Standard memory-context fields *//* Info about storage allocated in this context: */AllocBlock blocks;/* head of list of blocks in this...