cor3ntin added inline comments.
================ Comment at: clang/lib/Sema/SemaConcept.cpp:722 + const_cast<CXXMethodDecl *>(cast<CXXMethodDecl>(FD))); + // Captures are not checked from within the lambda. + LSI->AfterParameterList = false; ---------------- shafik wrote: > This comment does not really explain to me the effect of > `LSI->AfterParameterList = false;` `AfterParameterList` affects in which context we look for captured variables. I will try to think of a better comment. ================ Comment at: clang/lib/Sema/SemaExpr.cpp:19754 + + // When evaluating some attributes (like enable_if) we might refer to a + // function parameter appertaining to the same declaration as that ---------------- shafik wrote: > Why move this block of code for? `isVariableAlreadyCapturedInScopeInfo` is kinda badly name because it will adjust `DeclRefType` by adding `const` as necessary. if we capture a parameter, and then use it in a concept - which are not checked from within the scope of the lambda, we need to add const to it, which we can only do by reordering these paths. It's a bit subtle, and could do with some improvement as I'm not sure parameters will always be const correct in attributes currently. But I tried to do a minimal fix Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D158433/new/ https://reviews.llvm.org/D158433 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
