On 05/17/2017 09:06 AM, Eli Zaretskii wrote: > I think we might be talking about 2 different reports. I meant the > one here: > > http://lists.gnu.org/archive/html/emacs-devel/2017-05/msg00246.html Ah, that came after his original report , which is the one I was referring to. I reproduced the later problem, and as near as I can make it out it is due to an incompatibility between vfork and -fsanitize=address. My guess is that the vforked child corrupts the parent's shadow memory. I worked around the problem by installing the attached patch into Emacs.