================ @@ -245,10 +245,10 @@ runDataflowAnalysis( /// iterations. /// - This limit is still low enough to keep runtimes acceptable (on typical /// machines) in cases where we hit the limit. -template <typename AnalysisT, typename Diagnostic> -llvm::Expected<std::vector<Diagnostic>> diagnoseFunction( +template <typename AnalysisT, typename DiagnosticList> +llvm::Expected<DiagnosticList> diagnoseFunction( ---------------- ymand wrote:
That's my intuition as well (that `llvm::SmallVector` is likely better). But, who knows? An analysis could expect to have a non-small number of diagnostics per line. I really can't predict. So, I figured better to leave it open since the implementation doesn't care. I'm fine either way. The conservative approach is to go with `SmallVector` and open it up to other containers in the future if there's demand. https://github.com/llvm/llvm-project/pull/66014 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
