That said, the unfortunate thing about this is that it means that all checkers that implement VisitCallExpr() will need to do this checking. That seems really suboptimal. It seems to me that only a few checkers will care about operator methods.
On Nov 1, 2010, at 4:46 PM, Ted Kremenek wrote: > Hi Jim, > > I agree that VisitCallExpr() needs to handle explicit calls to operatorXX(), > but for the regular cases we should handle that in a separate visitor > callback. > > Ted > > On Nov 1, 2010, at 3:32 PM, Jim Goodnow II wrote: > >> At 03:25 PM 11/1/2010, Ted Kremenek wrote: >> >>> On Nov 1, 2010, at 2:45 PM, Jim Goodnow II wrote: >>> >>>> Well, those calls would actually be MemberCallExpr's, so it wouldn't >>>> occur in VisitCallExpr. The problem is really just an artifact of >>>> CXXOperatorCalls being grouped together with Calls. Eventually, they >>>> will probably be separated and this check can go away. >>> >>> Hi Jim, >>> >>> Why don't we do the right fix now? (i.e., do the refactoring you are >>> suggesting). >>> >>> Ted >> >> >> Well, Doug has a point. I hadn't seen that syntax before, but it does come >> through as a CallExpr and needs to be handled properly. My revised patch >> handles it. I'll look at the refactor as well. >> >> - jim >> > > > _______________________________________________ > cfe-commits mailing list > [email protected] > http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
