> We shouldn't mark the operation as used if the generic lambda itself appears > in a dependent context. I would strongly prefer that we set the CurContext to > the context surrounding the generic lambda before processing an implicit > capture to handle that case.
But I thought this patch won't mark it as used if the enclosing context is dependent - since I explicitly check for that - what am I missing? http://reviews.llvm.org/D6171 _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
