erichkeane added inline comments.
================ Comment at: clang/lib/Sema/Sema.cpp:460-462 + Scope S(TUScope, Scope::DeclScope, getDiagnostics()); + PushDeclContext(&S, DC); + PushOnScopeChains(ND, &S); ---------------- mizvekov wrote: > aaron.ballman wrote: > > Is it valid to use a local variable as a `Scope` object? I would assume > > that scope goes away at the end of this compound statement? I didn't spot > > other cases where we did this -- usually we call `Parser::EnterScope()` > > which allocates a new `Scope` that `Sema` then uses. > I don't see anything problematic about allocating it on the stack like this, > implementation-wise. > > I thought it would be a bit counter-intuitive to use parser functions here, > since these declarations are entirely synthetic. I dont know much about "Scope", as it is particularly used in Parsing, but I DO see the comment says: "A scope is a transient data structure that is used while parsing the program. It assists with resolving identifiers to the appropriate declaration." Which means to me it doesn't need to out-live its parsing time. I THINK it gets replaced with a CXXScope at one point, but I dont know when/where that happens. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D136886/new/ https://reviews.llvm.org/D136886 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
