On the other hand, the windshield of my volvo has a serious curvature
error (you can feel it by hand) which I assume stresses it as it's
constrained into the correct shape, and I've been watching a crack
slowly propagate from the top to the bottom for the past 15 years. Just
about there.
I have seen this happen a lot. I got a small rock crack in my Toyota
windshield 10 years ago. I immediately flowed cyanoacralayte into it
using a vacuum. The crack hasn't moved at all in that time. I've heard
it explained that water molecules stress the crack, expanding it. If it
went through a freeze thaw cycle that would do it.
Richard