[Lldb-commits] AdvDataFormatterTestCase relies on zeroed memory to pass
emaste at freebsd.org
Tue Jun 25 11:08:38 PDT 2013
On 25 June 2013 13:53, <jingham at apple.com> wrote:
> I'm not sure I agree with this. It doesn't seem like a good idea to have tests with undefined behavior ("on some systems there might be junk there that trips us up" seems a unreliable way to test handling bad data.) If we want to set up that kind of test we should either initialize values with random junk on purpose (though this seems a bit brute force) or test with patterns that we know have caused us problems or suspect might be tricky to handle.
Ok, you and Daniel have convinced me. We should craft an explicit
test for this behaviour if we think it's valuable to test. Let's go
with Daniel's change.
More information about the lldb-commits