Memory leak reporting is typically done at the end of execution. If the main
program had dynamically loaded some modules, and memory leaks occurred in those
modules, and the modules are unloaded before program termination, then the leak
reports will not have symbol name info for the addresses in the stack trace.

If the program loads and unloads modules repeatedly over the course of
execution, and the same address space is reused multiple times (does this
happen?) then the address/symbol info for the stack trace is unreliable; the
symbols found may not be relevant to the code that was actually executing at the
time.

It would be a good idea to add a time stamp or a sequence number to all the
malloc/free events, and remember the span of time or sequences in which a
particular module occupied a particular address space, so that the leaks can be
unambiguously tied to the specific module that caused them.