I just reproed with 130. It's a race condition, doesn't repro every time. Seems related to how long file loading takes --- i.e. once the pictures are in the file cache, the problem doesn't repro with the same pictures (I think, this is anecdotal, of course).
Seriously, if you can share the symbols (pdb) (no source), I can probably tell you the line numbers, or at least a meaningful stack trace.