[PATCH] D25475: [analyzer] Add a new SVal to support pointer-to-member operations.

[Kirill Romanenkov via Phabricator via cfe-commits]

kromanenkov added inline comments.


================
Comment at: lib/StaticAnalyzer/Core/ExprEngineC.cpp:899
+    case UO_AddrOf: {
+      // Process pointer-to-member address operation
+      const Expr *Ex = U->getSubExpr()->IgnoreParens();
----------------
kromanenkov wrote:
> dcoughlin wrote:
> > Just sticking this in the middle of a fallthrough cascade seems quite brittle. For example, it relies on the sub expression of a unary deref never being a DeclRefExpr to a field. This may be guaranteed by Sema (?) but if so it is quite a non-obvious invariant to rely upon.
> > 
> > I think it would be better the restructure this so that the AddrOf case doesn't fall in the the middle of a fallthrough cascade. You could either factor out the default behavior into a function or use a goto.
> It seems that UO_Extension case is the default behavior for this case cascade (just below UO_AddrOf). AFAIU it would be better to explicitly call the default behavior function following by break for each case preceding UO_Extension (UO_Plus,UO_Deref and UO_AddrOf). Or i missed something?
Or maybe just move UO_AddrOf to the beginning of the case cascade and proceed with the default behavior if it falls (no need to change UO_Plus and UO_Deref in that case)?


https://reviews.llvm.org/D25475