In the process of writing unit tests I find my code generates seg faults. OK, sort of good, thats what the tests are for. The tests however suppress all cout output from the code they are testing, and will not cout themselves either. GDB also will not run over the unit test executable. Valgrind will, which is helpful, but does not always give line numbers. By and large the lack of couts make it extremely difficult to track the cause of the seg faults (is the test environment causing the problem, is the main code flawed, if so where, etc). Can anything be done about this? Thanks.