So, I'm running some hardware tests using a GTEST binary deployed in a semihosted mode to my target hardware platform. Everything works great, except...
....when I get a failure with my external watchdog enabled, the external watchdog expires and resets my device before GTEST has a chance to dump any of its test data back to the host, even if I disable the watchdog on the next line after an EXPECT_xxx call. I've worked my way around it by doing simple logical tests and then disabling the watchdog before calling FAIL, but it feels like a little bit of a hack. Has anyone else run into this problem? Is there any way to maybe insert a cleanup callback into the assert_/expect_ calls?