harfbuzz/src/hb-repacker.hh

/*
 * Copyright © 2020  Google, Inc.
 *
 *  This is part of HarfBuzz, a text shaping library.
 *
 * Permission is hereby granted, without written agreement and without
 * license or royalty fees, to use, copy, modify, and distribute this
 * software and its documentation for any purpose, provided that the
 * above copyright notice and the following two paragraphs appear in
 * all copies of this software.
 *
 * IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
 * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
 * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
 * IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
 * DAMAGE.
 *
 * THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
 * BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
 * FITNESS FOR A PARTICULAR PURPOSE.  THE SOFTWARE PROVIDED HEREUNDER IS
 * ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
 * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
 *
 * Google Author(s): Garret Rieger
 */

#ifndef HB_REPACKER_HH
#define HB_REPACKER_HH

#include "hb-open-type.hh"
#include "hb-map.hh"
#include "hb-priority-queue.hh"
#include "hb-serialize.hh"
#include "hb-vector.hh"


struct graph_t
{
  // TODO(garretrieger): add an error tracking system similar to what serialize_context_t
  //                     does.

  struct overflow_record_t
  {
    unsigned parent;
    const hb_serialize_context_t::object_t::link_t* link;
  };

  struct clone_buffer_t
  {
    clone_buffer_t () : head (nullptr), tail (nullptr) {}

    void copy (const hb_serialize_context_t::object_t& object)
    {
      fini ();
      unsigned size = object.tail - object.head;
      head = (char*) malloc (size);
      memcpy (head, object.head, size);
      tail = head + size;
    }

    char* head;
    char* tail;

    void fini ()
    {
      if (!head) return;
      free (head);
      head = nullptr;
    }
  };

  /*
   * A topological sorting of an object graph. Ordered
   * in reverse serialization order (first object in the
   * serialization is at the end of the list). This matches
   * the 'packed' object stack used internally in the
   * serializer
   */
  graph_t (const hb_vector_t<hb_serialize_context_t::object_t *>& objects)
  {
    bool removed_nil = false;
    for (unsigned i = 0; i < objects.length; i++)
    {
      // TODO(grieger): check all links point to valid objects.

      // If this graph came from a serialization buffer object 0 is the
      // nil object. We don't need it for our purposes here so drop it.
      if (i == 0 && !objects[i])
      {
        removed_nil = true;
        continue;
      }

      auto* copy = objects_.push (*objects[i]);
      if (!removed_nil) continue;
      for (unsigned i = 0; i < copy->links.length; i++)
        // Fix indices to account for removed nil object.
        copy->links[i].objidx--;
    }
  }

  ~graph_t ()
  {
    objects_.fini_deep ();
    clone_buffers_.fini_deep ();
  }

  /*
   * serialize graph into the provided serialization buffer.
   */
  void serialize (hb_serialize_context_t* c)
  {
    c->start_serialize<void> ();
    for (unsigned i = 0; i < objects_.length; i++) {
      c->push ();

      size_t size = objects_[i].tail - objects_[i].head;
      char* start = c->allocate_size <char> (size);
      if (!start) return;

      memcpy (start, objects_[i].head, size);

      for (const auto& link : objects_[i].links)
        serialize_link (link, start, c);

      c->pop_pack (false);
    }
    c->end_serialize ();
  }

  /*
   * Generates a new topological sorting of graph using Kahn's
   * algorithm: https://en.wikipedia.org/wiki/Topological_sorting#Algorithms
   */
  void sort_kahn ()
  {
    if (objects_.length <= 1) {
      // Graph of 1 or less doesn't need sorting.
      return;
    }

    hb_vector_t<unsigned> queue;
    hb_vector_t<hb_serialize_context_t::object_t> sorted_graph;
    hb_vector_t<unsigned> id_map;
    id_map.resize (objects_.length);
    hb_vector_t<unsigned> edge_count;
    incoming_edge_count (&edge_count);

    // Object graphs are in reverse order, the first object is at the end
    // of the vector. Since the graph is topologically sorted it's safe to
    // assume the first object has no incoming edges.
    queue.push (objects_.length - 1);
    int new_id = objects_.length - 1;

    while (queue.length)
    {
      unsigned next_id = queue[0];
      queue.remove(0);

      hb_serialize_context_t::object_t& next = objects_[next_id];
      sorted_graph.push (next);
      id_map[next_id] = new_id--;

      for (const auto& link : next.links) {
        // TODO(garretrieger): sort children from smallest to largest
        edge_count[link.objidx] -= 1;
        if (!edge_count[link.objidx])
          queue.push (link.objidx);
      }
    }

    if (new_id != -1)
    {
      // Graph is not fully connected, there are unsorted objects.
      // TODO(garretrieger): handle this.
      assert (false);
    }

    remap_obj_indices (id_map, &sorted_graph);

    sorted_graph.as_array ().reverse ();
    objects_ = sorted_graph;
    sorted_graph.fini_deep ();
  }

  /*
   * Generates a new topological sorting of graph ordered by the shortest
   * distance to each node.
   */
  void sort_shortest_distance ()
  {
    if (objects_.length <= 1) {
      // Graph of 1 or less doesn't need sorting.
      return;
    }

    hb_vector_t<int64_t> distance_to;
    compute_distances (&distance_to);

    hb_priority_queue_t queue;
    hb_vector_t<hb_serialize_context_t::object_t> sorted_graph;
    hb_vector_t<unsigned> id_map;
    id_map.resize (objects_.length);
    hb_vector_t<unsigned> edge_count;
    incoming_edge_count (&edge_count);

    // Object graphs are in reverse order, the first object is at the end
    // of the vector. Since the graph is topologically sorted it's safe to
    // assume the first object has no incoming edges.
    queue.insert (objects_.length - 1, add_order(distance_to[objects_.length - 1], 0));
    int new_id = objects_.length - 1;
    unsigned order = 1;
    while (!queue.is_empty ())
    {
      unsigned next_id = queue.extract_minimum().first;

      hb_serialize_context_t::object_t& next = objects_[next_id];
      sorted_graph.push (next);
      id_map[next_id] = new_id--;

      for (const auto& link : next.links) {
        edge_count[link.objidx] -= 1;
        if (!edge_count[link.objidx])
          // Add the order that the links were encountered to the priority.
          // This ensures that ties between priorities objects are broken in a consistent
          // way. More specifically this is set up so that if a set of objects have the same
          // distance they'll be added to the topolical order in the order that they are
          // referenced from the parent object.
          queue.insert (link.objidx, add_order(distance_to[link.objidx], order++));
      }
    }

    if (new_id != -1)
    {
      // Graph is not fully connected, there are unsorted objects.
      // TODO(garretrieger): handle this.
      assert (false);
    }

    remap_obj_indices (id_map, &sorted_graph);

    sorted_graph.as_array ().reverse ();
    objects_ = sorted_graph;
    sorted_graph.fini_deep ();
  }

  /*
   * Creates a copy of child and re-assigns the link from
   * parent to the clone. The copy is a shallow copy, objects
   * linked from child are not duplicated.
   */
  void duplicate (unsigned parent_idx, unsigned child_idx)
  {
    const auto& child = objects_[child_idx];
    clone_buffer_t* buffer = clone_buffers_.push ();
    buffer->copy (child);

    auto* clone = objects_.push ();
    clone->head = buffer->head;
    clone->tail = buffer->tail;
    for (const auto& l : child.links)
      clone->links.push (l);

    auto& parent = objects_[parent_idx];
    unsigned clone_idx = objects_.length - 2;
    for (unsigned i = 0; i < parent.links.length; i++)
    {
      auto& l = parent.links[i];
      if (l.objidx == child_idx) l.objidx = clone_idx;
    }

    // The last object is the root of the graph, so swap back the root to the end.
    // The root's obj idx does change, however since it's root nothing else refers to it.
    // all other obj idx's will be unaffected.
    hb_serialize_context_t::object_t root = objects_[objects_.length - 2];
    objects_[objects_.length - 2] = *clone;
    objects_[objects_.length - 1] = root;
  }

  /*
   * Will any offsets overflow on graph when it's serialized?
   */
  bool will_overflow (hb_vector_t<overflow_record_t>* overflows)
  {
    if (overflows) overflows->resize (0);
    hb_vector_t<unsigned> start_positions;
    start_positions.resize (objects_.length);
    hb_vector_t<unsigned> end_positions;
    end_positions.resize (objects_.length);

    unsigned current_pos = 0;
    for (int i = objects_.length - 1; i >= 0; i--)
    {
      start_positions[i] = current_pos;
      current_pos += objects_[i].tail - objects_[i].head;
      end_positions[i] = current_pos;
    }


    for (int parent_idx = objects_.length - 1; parent_idx >= 0; parent_idx--)
    {
      for (const auto& link : objects_[parent_idx].links)
      {
        int64_t offset = compute_offset (parent_idx,
                                         link,
                                         start_positions,
                                         end_positions);

        if (is_valid_offset (offset, link))
          continue;

        if (!overflows) return true;

        overflow_record_t r;
        r.parent = parent_idx;
        r.link = &link;
        overflows->push (r);
      }
    }

    if (!overflows) return false;
    return overflows->length;
  }

  /*
   * Creates a map from objid to # of incoming edges.
   */
  void incoming_edge_count (hb_vector_t<unsigned>* out) const
  {
    out->resize (0);
    out->resize (objects_.length);
    for (const auto& o : objects_)
    {
      for (const auto& l : o.links)
      {
        (*out)[l.objidx] += 1;
      }
    }
  }

  void print_overflows (const hb_vector_t<overflow_record_t>& overflows) const
  {
    if (!DEBUG_ENABLED(SUBSET_REPACK)) return;

    hb_vector_t<unsigned> edge_count;
    incoming_edge_count (&edge_count);
    for (const auto& o : overflows)
    {
      DEBUG_MSG (SUBSET_REPACK, nullptr, "overflow from %d => %d (%d incoming , %d outgoing)",
                 o.parent,
                 o.link->objidx,
                 edge_count[o.link->objidx],
                 objects_[o.link->objidx].links.length);
    }
  }

 private:

  int64_t add_order (int64_t distance, unsigned order)
  {
    return (distance << 24) | (0x00FFFFFF & order);
  }

  /*
   * Finds the distance too each object in the graph
   * from the initial node.
   */
  void compute_distances (hb_vector_t<int64_t>* distance_to)
  {
    // Uses Dijkstra's algorithm to find all of the shortest distances.
    // https://en.wikipedia.org/wiki/Dijkstra%27s_algorithm
    //
    // Implementation Note:
    // Since our priority queue doesn't support fast priority decreases
    // we instead just add new entries into the queue when a priority changes.
    // Redundant ones are filtered out later on by the visited set.
    // According to https://www3.cs.stonybrook.edu/~rezaul/papers/TR-07-54.pdf
    // for practical performance this is faster then using a more advanced queue
    // (such as a fibonaacci queue) with a fast decrease priority.
    distance_to->resize (0);
    distance_to->resize (objects_.length);
    for (unsigned i = 0; i < objects_.length; i++)
    {
      if (i == objects_.length - 1)
        (*distance_to)[i] = 0;
      else
        (*distance_to)[i] = hb_int_max (int64_t);
    }

    hb_priority_queue_t queue;
    queue.insert (objects_.length - 1, 0);

    hb_set_t visited;

    while (!queue.is_empty ())
    {
      unsigned next_idx = queue.extract_minimum ().first;
      if (visited.has (next_idx)) continue;
      const auto& next = objects_[next_idx];
      int64_t next_distance = (*distance_to)[next_idx];
      visited.add (next_idx);

      for (const auto& link : next.links)
      {
        if (visited.has (link.objidx)) continue;

        const auto& child = objects_[link.objidx];
        int64_t child_weight = child.tail - child.head +
                               (!link.is_wide ? (1 << 16) : ((int64_t) 1 << 32));
        int64_t child_distance = next_distance + child_weight;

        if (child_distance < (*distance_to)[link.objidx])
        {
          (*distance_to)[link.objidx] = child_distance;
          queue.insert (link.objidx, child_distance);
        }
      }
    }
    // TODO(garretrieger): Handle this. If anything is left, part of the graph is disconnected.
    assert (queue.is_empty ());
  }

  int64_t compute_offset (
      unsigned parent_idx,
      const hb_serialize_context_t::object_t::link_t& link,
      const hb_vector_t<unsigned>& start_positions,
      const hb_vector_t<unsigned>& end_positions)
  {
    unsigned child_idx = link.objidx;
    int64_t offset = 0;
    switch ((hb_serialize_context_t::whence_t) link.whence) {
      case hb_serialize_context_t::whence_t::Head:
        offset = start_positions[child_idx] - start_positions[parent_idx]; break;
      case hb_serialize_context_t::whence_t::Tail:
        offset = start_positions[child_idx] - end_positions[parent_idx]; break;
      case hb_serialize_context_t::whence_t::Absolute:
        offset = start_positions[child_idx]; break;
    }

    assert (offset >= link.bias);
    offset -= link.bias;
    return offset;
  }

  bool is_valid_offset (int64_t offset,
                        const hb_serialize_context_t::object_t::link_t& link)
  {
    if (link.is_signed)
    {
      if (link.is_wide)
        return offset >= -((int64_t) 1 << 31) && offset < ((int64_t) 1 << 31);
      else
        return offset >= -(1 << 15) && offset < (1 << 15);
    }
    else
    {
      if (link.is_wide)
        return offset >= 0 && offset < ((int64_t) 1 << 32);
      else
        return offset >= 0 && offset < (1 << 16);
    }
  }

  /*
   * Updates all objidx's in all links using the provided mapping.
   */
  void remap_obj_indices (const hb_vector_t<unsigned>& id_map,
                          hb_vector_t<hb_serialize_context_t::object_t>* sorted_graph)
  {
    for (unsigned i = 0; i < sorted_graph->length; i++)
    {
      for (unsigned j = 0; j < (*sorted_graph)[i].links.length; j++)
      {
        auto& link = (*sorted_graph)[i].links[j];
        link.objidx = id_map[link.objidx];
      }
    }
  }

  template <typename O> void
  serialize_link_of_type (const hb_serialize_context_t::object_t::link_t& link,
                          char* head,
                          hb_serialize_context_t* c)
  {
    OT::Offset<O>* offset = reinterpret_cast<OT::Offset<O>*> (head + link.position);
    *offset = 0;
    c->add_link (*offset,
                 // serializer has an extra nil object at the start of the
                 // object array. So all id's are +1 of what our id's are.
                 link.objidx + 1,
                 (hb_serialize_context_t::whence_t) link.whence,
                 link.bias);
  }

  void serialize_link (const hb_serialize_context_t::object_t::link_t& link,
                 char* head,
                 hb_serialize_context_t* c)
  {
    if (link.is_wide)
    {
      if (link.is_signed)
      {
        serialize_link_of_type<OT::HBINT32> (link, head, c);
      } else {
        serialize_link_of_type<OT::HBUINT32> (link, head, c);
      }
    } else {
      if (link.is_signed)
      {
        serialize_link_of_type<OT::HBINT16> (link, head, c);
      } else {
        serialize_link_of_type<OT::HBUINT16> (link, head, c);
      }
    }
  }

 public:
  hb_vector_t<hb_serialize_context_t::object_t> objects_;
  hb_vector_t<clone_buffer_t> clone_buffers_;
};


/*
 * Re-serialize the provided object graph into the serialization context
 * using BFS (Breadth First Search) to produce the topological ordering.
 */
inline void
hb_resolve_overflows (const hb_vector_t<hb_serialize_context_t::object_t *>& packed,
                      hb_serialize_context_t* c) {
  graph_t sorted_graph (packed);
  sorted_graph.sort_kahn ();
  if (!sorted_graph.will_overflow (nullptr)) return;

  sorted_graph.sort_shortest_distance ();

  unsigned round = 0;
  hb_vector_t<graph_t::overflow_record_t> overflows;
  // TODO(garretrieger): select a good limit for max rounds.
  while (sorted_graph.will_overflow (&overflows) && round++ < 10) {
    DEBUG_MSG (SUBSET_REPACK, nullptr, "Over flow resolution round %d", round);
    sorted_graph.print_overflows (overflows);

    // TODO(garretrieger): cache ege count in the graph object . Will need to be invalidated
    //                     by graph modifications.
    hb_vector_t<unsigned> edge_count;
    sorted_graph.incoming_edge_count (&edge_count);

    // Try resolving the furthest overflow first.
    bool resolution_attempted = false;
    for (int i = overflows.length - 1; i >= 0; i--)
    {
      const graph_t::overflow_record_t& r = overflows[i];
      if (edge_count[r.link->objidx] > 1)
      {
        DEBUG_MSG (SUBSET_REPACK, nullptr, "Duplicating %d => %d",
                   r.parent, r.link->objidx);
        // The child object is shared, we may be able to eliminate the overflow
        // by duplicating it.
        sorted_graph.duplicate (r.parent, r.link->objidx);
        sorted_graph.sort_shortest_distance ();
        resolution_attempted = true;
        break;
      }

      // TODO(garretrieger): add additional offset resolution strategies
      // - Promotion to extension lookups.
      // - Table splitting.
    }

    if (!resolution_attempted)
    {
      DEBUG_MSG (SUBSET_REPACK, nullptr, "No resolution available :(");
      break;
    }
  }

  sorted_graph.serialize (c);
}


#endif /* HB_REPACKER_HH */