[PATCH] D43254: Use LLVM's DenseSet insteead of unordered_set.

[David Blaikie via llvm-commits]

On Tue, Feb 13, 2018 at 1:49 PM Rui Ueyama via Phabricator via llvm-commits
<llvm-commits at lists.llvm.org> wrote:

> ruiu added inline comments.
>
>
> ================
> Comment at: lld/wasm/SymbolTable.cpp:37
>  void SymbolTable::reportRemainingUndefines() {
> -  std::unordered_set<Symbol *> Undefs;
> +  DenseSet<Symbol *> Undefs;
>    for (Symbol *Sym : SymVector) {
> ----------------
> efriedma wrote:
> > ruiu wrote:
> > > efriedma wrote:
> > > > Don't you need a SetVector here?
> > > Forgive my ignorance, but what's the difference between DenseSet and
> SetVector?
> > Iterating over a SetVector produces values in the order they were
> inserted; iterating over a DenseSet uses an arbitrary ordering.  See
> http://llvm.org/docs/ProgrammersManual.html#llvm-adt-setvector-h .
> Ah, thanks. This code is to produce error messages, so the order of the
> output of this code doesn't have to be deterministic, but it is still nice
> if it's deterministic. I'll make a change to use SetVector.
>

I'm pretty sure in other cases across the LLVM project determinism of
output, even for diagnostics, is considered necessary. (usually this comes
up simply enough in test cases - if diagnostic order isn't deterministic,
test cases fail on one platform and succeed on another, or fail randomly
depending on hash of memory layout, etc)


>
>
> https://reviews.llvm.org/D43254
>
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20180220/9da8030d/attachment.html>