The app does run properly in 2k3sp2.
When I run it under ros, I currently get with 100% chance a crash due to unhandled exception.
The log first contains a run without breaking into debugger on first chance and a stacktrace with that.
Then I activated break on first chance, and let it crash another time, captured the stacktrace and translated it with log2lines.
This is a regression, because the same binary had at least a small probability (maybe 10%?) to run properly in the past in ros (in case it did not BSOD like in
CORE-13907 with ~90%).
This ticket here therefore currently blocks the analysis of