Looks good to me. Please file a PR about checking for null derefs with the output parameters.
Thanks! On Mar 2, 2012, at 1:18 AM, Erik Verbruggen <[email protected]> wrote: > New patch attached. > > And in my question I mixed input and output parameters up. The input > parameters are checked, but it looks like the output parameters are not. See > the test in the patch to the nullptr.cpp. > > -- Erik. > > On 1-3-12 20:35, Ted Kremenek wrote: >> Interesting. We aren't actually simulating a load, which is done by >> 'evalLoad', which triggers the dereference checking. We could do that after >> your cleanup. >> >> On Mar 1, 2012, at 11:29 AM, Erik Verbruggen<[email protected]> wrote: >> >>> Hi Ted, >>> >>> I'll Hove it a spin. Do you by chance have any idea why clang (with and >>> without patch) does not warn on the de-ref of the input parameter? >>> >>> -- Erik. >>> > <0001-Remove-a-recursive-visitiation-in-ExprEngine-that-is.patch> _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
