I have a reservation about one of the comments. I will fix the rest of the issues in a new version of the patch (probably tomorrow).
================ Comment at: lib/StaticAnalyzer/Core/ExprEngineC.cpp:534 @@ +533,3 @@ + SValBuilder &SVB = State->getStateManager().getSValBuilder(); + return State->BindExpr(E, LC, SVB.evalCast(X, B->getType(), XType)); +} ---------------- Jordan Rose wrote: > Rather than have this bind anything, why not just return the properly-casted > value? For this patch, I could indeed just return the correct SVal. However, here I am also preparing the ground for the second patch. For that, I need the value to be stored in the state, so that I can later query them when I run through the temporary destructor branches. So, in theory I could do the binding only when cfg-temporary-dtors is true, but I'm not sure if that's worth the increased code complexity. http://llvm-reviews.chandlerc.com/D1340 _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
