[PATCH] Store pointers to seen formatting states.

[Manuel Klimek]

test


On Mon, Feb 18, 2013 at 8:21 AM, Daniel Jasper <djasper at google.com> wrote:

> Hi klimek,
>
> As the memory ownership is handled by the SpecificBumpPtrAllocator anyway,
> there is no need to duplicate states when inserting them into the Seen-set.
> This leads to an improvement of ~10% on the benchmark formatting file. Not
> sure whether this is worth it, or whether we should straight away look into
> implementing this as a hash map.
>
> http://llvm-reviews.chandlerc.com/D418
>
> Files:
>   lib/Format/Format.cpp
>
> Index: lib/Format/Format.cpp
> ===================================================================
> --- lib/Format/Format.cpp
> +++ lib/Format/Format.cpp
> @@ -676,6 +676,12 @@
>    typedef std::priority_queue<QueueItem, std::vector<QueueItem>,
>                                std::greater<QueueItem> > QueueType;
>
> +  struct Compare {
> +    bool operator()(LineState *obj1, LineState *obj2) const {
> +      return *obj1 < *obj2;
> +    }
> +  };
> +
>    /// \brief Analyze the entire solution space starting from \p
> InitialState.
>    ///
>    /// This implements a variant of Dijkstra's algorithm on the graph that
> spans
> @@ -683,7 +689,7 @@
>    /// find the shortest path (the one with lowest penalty) from \p
> InitialState
>    /// to a state where all tokens are placed.
>    unsigned analyzeSolutionSpace(LineState &InitialState) {
> -    std::set<LineState> Seen;
> +    std::set<LineState *, Compare> Seen;
>
>      // Insert start element into queue.
>      StateNode *Node =
> @@ -701,7 +707,7 @@
>        }
>        Queue.pop();
>
> -      if (!Seen.insert(Node->State).second)
> +      if (!Seen.insert(&Node->State).second)
>          // State already examined with lower penalty.
>          continue;
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20130708/efbf8ae2/attachment.html>