Re: Coverity doing scans of Wine codebase!

[James Hawkins]

On 4/6/06, Troy Rollo  wrote:
> Some of the "bugs" it picks up are cases of defensive programming, such as
> checking for a NULL pointer even though the NULL pointer is impossible, given
> the functioning of routines that are being called. What should the policy be
> on such bugs - should we remove the unnecessary check or keep the
> extra-defensive code?

In most internal functions we should not have the check so we crash if
we ever send in a NULL ptr (in cases where we should never send in
NULL.)  Mosts of the win32 API should have checks, but the test suite
verifies these for sure.

--
James Hawkins