Refactor idlist

src/CMakeLists.txt

@@ -19,6 +19,7 @@ target_sources(osm2pgsql_lib PRIVATE
   geom-from-osm.cpp
   geom-functions.cpp
   geom-pole-of-inaccessibility.cpp
+  idlist.cpp
   input.cpp
   logging.cpp
   middle.cpp

src/dependency-manager.cpp

@@ -15,17 +15,17 @@
 
 void full_dependency_manager_t::node_changed(osmid_t id)
 {
-    m_changed_nodes.set(id);
+    m_changed_nodes.push_back(id);
 }
 
 void full_dependency_manager_t::way_changed(osmid_t id)
 {
-    m_changed_ways.set(id);
+    m_changed_ways.push_back(id);
 }
 
 void full_dependency_manager_t::relation_changed(osmid_t id)
 {
-    m_changed_relations.set(id);
+    m_changed_relations.push_back(id);
 }
 
 void full_dependency_manager_t::after_nodes()
@@ -39,29 +39,13 @@ void full_dependency_manager_t::after_nodes()
     m_changed_nodes.clear();
 }
 
-static osmium::index::IdSetSmall<osmid_t>
-set_diff(osmium::index::IdSetSmall<osmid_t> const &set,
-         osmium::index::IdSetSmall<osmid_t> const &to_be_removed)
-{
-    osmium::index::IdSetSmall<osmid_t> new_set;
-
-    for (auto const id : set) {
-        if (!to_be_removed.get_binary_search(id)) {
-            new_set.set(id);
-        }
-    }
-
-    return new_set;
-}
-
 void full_dependency_manager_t::after_ways()
 {
     if (!m_changed_ways.empty()) {
         if (!m_ways_pending_tracker.empty()) {
             // Remove ids from changed ways in the input data from
             // m_ways_pending_tracker, because they have already been processed.
-            m_ways_pending_tracker =
-                set_diff(m_ways_pending_tracker, m_changed_ways);
+            m_ways_pending_tracker.remove_ids_if_in(m_changed_ways);
 
             // Add the list of pending way ids to the list of changed ways,
             // because we need the parents for them, too.
@@ -85,14 +69,13 @@ void full_dependency_manager_t::after_relations()
 {
     // Remove ids from changed relations in the input data from
     // m_rels_pending_tracker, because they have already been processed.
-    m_rels_pending_tracker =
-        set_diff(m_rels_pending_tracker, m_changed_relations);
+    m_rels_pending_tracker.remove_ids_if_in(m_changed_relations);
 
     m_changed_relations.clear();
 }
 
 void full_dependency_manager_t::mark_parent_relations_as_pending(
-    osmium::index::IdSetSmall<osmid_t> const &way_ids)
+    idlist_t const &way_ids)
 {
     assert(m_rels_pending_tracker.empty());
     m_object_store->get_way_parents(way_ids, &m_rels_pending_tracker);
@@ -102,18 +85,3 @@ bool full_dependency_manager_t::has_pending() const noexcept
 {
     return !m_ways_pending_tracker.empty() || !m_rels_pending_tracker.empty();
 }
-
-idlist_t
-full_dependency_manager_t::get_ids(osmium::index::IdSetSmall<osmid_t> *tracker)
-{
-    tracker->sort_unique();
-
-    idlist_t list;
-    list.reserve(tracker->size());
-
-    std::copy(tracker->cbegin(), tracker->cend(), std::back_inserter(list));
-
-    tracker->clear();
-
-    return list;
-}

src/dependency-manager.hpp

@@ -10,10 +10,9 @@
  * For a full list of authors see the git log.
  */
 
+#include "idlist.hpp"
 #include "osmtypes.hpp"
 
-#include <osmium/index/id_set.hpp>
-
 #include <cassert>
 #include <memory>
 #include <utility>
@@ -58,8 +57,7 @@ class dependency_manager_t
     virtual void after_ways() {}
     virtual void after_relations() {}
 
-    virtual void mark_parent_relations_as_pending(
-        osmium::index::IdSetSmall<osmid_t> const & /*way_ids*/)
+    virtual void mark_parent_relations_as_pending(idlist_t const & /*way_ids*/)
     {
     }
 
@@ -111,24 +109,29 @@ class full_dependency_manager_t : public dependency_manager_t
     void after_ways() override;
     void after_relations() override;
 
-    void mark_parent_relations_as_pending(
-        osmium::index::IdSetSmall<osmid_t> const &ids) override;
+    void mark_parent_relations_as_pending(idlist_t const &ids) override;
 
     bool has_pending() const noexcept override;
 
     idlist_t get_pending_way_ids() override
     {
-        return get_ids(&m_ways_pending_tracker);
+        idlist_t list;
+        using std::swap;
+        swap(list, m_ways_pending_tracker);
+        list.sort_unique();
+        return list;
     }
 
     idlist_t get_pending_relation_ids() override
     {
-        return get_ids(&m_rels_pending_tracker);
+        idlist_t list;
+        using std::swap;
+        swap(list, m_rels_pending_tracker);
+        list.sort_unique();
+        return list;
     }
 
 private:
-    static idlist_t get_ids(osmium::index::IdSetSmall<osmid_t> *tracker);
-
     std::shared_ptr<middle_t> m_object_store;
 
     /**
@@ -139,7 +142,7 @@ class full_dependency_manager_t : public dependency_manager_t
      * the change file, too, and so we don't have to find out which ones they
      * are.
      */
-    osmium::index::IdSetSmall<osmid_t> m_changed_nodes;
+    idlist_t m_changed_nodes;
 
     /**
      * In append mode all new and changed ways will be added to this. After
@@ -148,17 +151,17 @@ class full_dependency_manager_t : public dependency_manager_t
      * relations that referenced deleted ways must be in the change file, too,
      * and so we don't have to find out which ones they are.
      */
-    osmium::index::IdSetSmall<osmid_t> m_changed_ways;
+    idlist_t m_changed_ways;
 
     /**
      * In append mode all new and changed relations will be added to this.
      * This is then used to remove already processed relations from the
      * pending list.
      */
-    osmium::index::IdSetSmall<osmid_t> m_changed_relations;
+    idlist_t m_changed_relations;
 
-    osmium::index::IdSetSmall<osmid_t> m_ways_pending_tracker;
-    osmium::index::IdSetSmall<osmid_t> m_rels_pending_tracker;
+    idlist_t m_ways_pending_tracker;
+    idlist_t m_rels_pending_tracker;
 };
 
 #endif // OSM2PGSQL_DEPENDENCY_MANAGER_HPP

src/idlist.cpp

@@ -0,0 +1,54 @@
+/**
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ *
+ * This file is part of osm2pgsql (https://osm2pgsql.org/).
+ *
+ * Copyright (C) 2006-2023 by the osm2pgsql developer community.
+ * For a full list of authors see the git log.
+ */
+
+#include "idlist.hpp"
+
+#include <algorithm>
+#include <cassert>
+#include <iterator>
+#include <utility>
+
+osmid_t idlist_t::pop_id()
+{
+    assert(!m_list.empty());
+    auto const id = m_list.back();
+    m_list.pop_back();
+    return id;
+}
+
+void idlist_t::sort_unique()
+{
+    std::sort(m_list.begin(), m_list.end());
+    const auto last = std::unique(m_list.begin(), m_list.end());
+    m_list.erase(last, m_list.end());
+}
+
+void idlist_t::merge_sorted(idlist_t const &other)
+{
+    std::vector<osmid_t> new_list;
+
+    new_list.reserve(m_list.size() + other.m_list.size());
+    std::set_union(m_list.cbegin(), m_list.cend(), other.m_list.cbegin(),
+                   other.m_list.cend(), std::back_inserter(new_list));
+
+    using std::swap;
+    swap(new_list, m_list);
+}
+
+void idlist_t::remove_ids_if_in(idlist_t const &other)
+{
+    std::vector<osmid_t> new_list;
+
+    new_list.reserve(m_list.size());
+    std::set_difference(m_list.cbegin(), m_list.cend(), other.m_list.cbegin(),
+                        other.m_list.cend(), std::back_inserter(new_list));
+
+    using std::swap;
+    swap(new_list, m_list);
+}

src/idlist.hpp

@@ -0,0 +1,107 @@
+#ifndef OSM2PGSQL_IDLIST_HPP
+#define OSM2PGSQL_IDLIST_HPP
+
+/**
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ *
+ * This file is part of osm2pgsql (https://osm2pgsql.org/).
+ *
+ * Copyright (C) 2006-2023 by the osm2pgsql developer community.
+ * For a full list of authors see the git log.
+ */
+
+/**
+ * \file
+ *
+ * This file contains the definition of the idlist_t class.
+ */
+
+#include "osmtypes.hpp"
+
+#include <vector>
+
+/**
+ * A list of OSM object ids. Internally this is a vector of ids.
+ *
+ * Some operations are only allowed when the list of ids is sorted and
+ * without duplicates. Call sort_unique() to achieve this.
+ */
+class idlist_t
+{
+public:
+    using value_type = osmid_t;
+
+    idlist_t() = default;
+    ~idlist_t() noexcept = default;
+
+    idlist_t(std::initializer_list<osmid_t> ids) : m_list(ids) {}
+
+    idlist_t(idlist_t const &) = delete;
+    idlist_t &operator=(idlist_t const &) = delete;
+
+    idlist_t(idlist_t &&) = default;
+    idlist_t &operator=(idlist_t &&) = default;
+
+    bool empty() const noexcept { return m_list.empty(); }
+
+    std::size_t size() const noexcept { return m_list.size(); }
+
+    auto begin() const noexcept { return m_list.begin(); }
+
+    auto end() const noexcept { return m_list.end(); }
+
+    auto cbegin() const noexcept { return m_list.cbegin(); }
+
+    auto cend() const noexcept { return m_list.cend(); }
+
+    osmid_t operator[](std::size_t n) const noexcept { return m_list[n]; }
+
+    void clear() noexcept { m_list.clear(); }
+
+    void push_back(osmid_t id) { m_list.push_back(id); }
+
+    void reserve(std::size_t size) { m_list.reserve(size); }
+
+    /**
+     * Remove id at the end of the list and return it.
+     *
+     * \pre \code !m_list.empty()) \endcode
+     */
+    osmid_t pop_id();
+
+    /// List are equal if they contain the same ids in the same order.
+    friend bool operator==(idlist_t const &lhs, idlist_t const &rhs) noexcept
+    {
+        return lhs.m_list == rhs.m_list;
+    }
+
+    friend bool operator!=(idlist_t const &lhs, idlist_t const &rhs) noexcept
+    {
+        return !(lhs == rhs);
+    }
+
+    /**
+     * Sort this list and remove duplicates.
+     */
+    void sort_unique();
+
+    /**
+     * Merge other list into this one.
+     *
+     * \pre Both lists must be sorted and without duplicates.
+     */
+    void merge_sorted(idlist_t const &other);
+
+    /**
+     * Remove all ids in this list that are also in the other list.
+     *
+     * \pre Both lists must be sorted and without duplicates.
+     */
+    void remove_ids_if_in(idlist_t const &other);
+
+private:
+    std::vector<osmid_t> m_list;
+
+}; // class idlist_t
+
+#endif // OSM2PGSQL_IDLIST_HPP