================ @@ -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( ---------------- martinboehme 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. > > Even then, a `SmallVector` isn't a bad choice -- and potentially a better > choice than a `std::vector`. (The "small" in `SmallVector` doesn't mean "this > only performs well if the contents are small" -- it just means "this performs > particularly well if the contents are small".) > > The main cost of a `SmallVector` is that it occupies more memory than a > `std::vector`. This isn't an issue here since we're allocating only one of > these vectors at a time, on the stack. In terms of runtime, my understanding > is that `SmallVector` is typically _faster_ than `std::vector` (for small > _and_ large sizes). From the [LLVM Programmers' > Manual](https://llvm.org/docs/ProgrammersManual.html#llvm-adt-smallvector-h): > > > SmallVector has grown a few other minor advantages over std::vector, > > causing `SmallVector<Type, 0>` to be preferred over `std::vector<Type>`." > > [...] > > SmallVector understands std::is_trivially_copyable and uses realloc > > aggressively > > This talks about `SmallVector<Type, 0>`, but this is just because for this > general statement they're thinking about the size of the object. In our case, > that isn't a concern, and we want to take advantage of the "small number of > elements" optimization, so `SmallVector<Type>` seems like the right choice. So bottom line, I don't expect anyone to ever want to use `std::vector` here. 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
