Merge branch 'dev' into s3-no-md5

TileDB-Inc · Aug 15, 2024 · 9623cc7 · 9623cc7
2 parents 9479a81 + 2edb451
commit 9623cc7
Show file tree

Hide file tree

Showing 5 changed files with 27 additions and 16 deletions.
diff --git a/.github/workflows/build-windows.yml b/.github/workflows/build-windows.yml
@@ -136,6 +136,8 @@ jobs:
         uses: seanmiddleditch/gha-setup-ninja@v4
       - name: Prevent vcpkg from building debug variants
         run: python $env:GITHUB_WORKSPACE/scripts/ci/patch_vcpkg_triplets.py
+      - name: Install dependencies from pip
+        run: python -m pip install pybind11[global]
 
       - name: Configure TileDB
         shell: pwsh

diff --git a/test/CMakeLists.txt b/test/CMakeLists.txt
@@ -58,9 +58,9 @@ if (${TILEDB_ARROW_TESTS})
   if (NOT ${pybind11_FOUND})
     # Get the include arguments from the python executable (has "-I" compiler option)
     execute_process(COMMAND  ${Python_EXECUTABLE} -m pybind11 --includes
-                    OUTPUT_VARIABLE CMD_PYBIND11_INCLUDE
-                    RESULT_VARIABLE CMD_PYBIND11_RESULT
-                    OUTPUT_STRIP_TRAILING_WHITESPACE)
+            OUTPUT_VARIABLE CMD_PYBIND11_INCLUDE
+            RESULT_VARIABLE CMD_PYBIND11_RESULT
+            OUTPUT_STRIP_TRAILING_WHITESPACE)
     if (${CMD_PYBIND11_RESULT})
       message(FATAL_ERROR "Unable to find pybind11 via cmake or 'python3 -m pybind11 --includes'")
     endif()

diff --git a/tiledb/sm/fragment/fragment_metadata.cc b/tiledb/sm/fragment/fragment_metadata.cc
@@ -1938,22 +1938,19 @@ void FragmentMetadata::expand_non_empty_domain(const NDRange& mbr) {
 
 // ===== FORMAT =====
 //  bounding_coords_num (uint64_t)
-//  bounding_coords_#1 (void*) bounding_coords_#2 (void*) ...
+//  bounding_coords_#1 (dim_type[2][n_dims])
+//  bounding_coords_#2 (dim_type[2][n_dims])
+//  ...
+// Note: This version supports only dimensions domains with the same type
 void FragmentMetadata::load_bounding_coords(Deserializer& deserializer) {
   // Get number of bounding coordinates
-  uint64_t bounding_coords_num = 0;
-  bounding_coords_num = deserializer.read<uint64_t>();
+  uint64_t bounding_coords_num = deserializer.read<uint64_t>();
 
-  // Get bounding coordinates
-  // Note: This version supports only dimensions domains with the same type
   auto coord_size{array_schema_->domain().dimension_ptr(0)->coord_size()};
   auto dim_num = array_schema_->domain().dim_num();
   uint64_t bounding_coords_size = 2 * dim_num * coord_size;
-  bounding_coords_.resize(bounding_coords_num);
-  for (uint64_t i = 0; i < bounding_coords_num; ++i) {
-    bounding_coords_[i].resize(bounding_coords_size);
-    deserializer.read(&bounding_coords_[i][0], bounding_coords_size);
-  }
+  // Skip bounding coordinates. They are not actually being used.
+  deserializer.skip(bounding_coords_num * bounding_coords_size);
 }
 
 void FragmentMetadata::load_file_sizes(Deserializer& deserializer) {

diff --git a/tiledb/sm/fragment/fragment_metadata.h b/tiledb/sm/fragment/fragment_metadata.h
@@ -853,9 +853,6 @@ class FragmentMetadata {
    */
   std::unordered_map<std::string, unsigned> idx_map_;
 
-  /** A vector storing the first and last coordinates of each tile. */
-  std::vector<std::vector<uint8_t>> bounding_coords_;
-
   /** True if the fragment is dense, and false if it is sparse. */
   bool dense_;
 

diff --git a/tiledb/storage_format/serialization/serializers.h b/tiledb/storage_format/serialization/serializers.h
@@ -184,6 +184,21 @@ class Deserializer {
     size_ -= size;
   }
 
+  /**
+   * Advances the deserializer's pointer by a specified number of bytes, without
+   * reading them.
+   *
+   * @param size number of bytes to skip.
+   */
+  void skip(storage_size_t size) {
+    if (size > size_) {
+      throw std::logic_error("Reading data past end of serialized data size.");
+    }
+
+    ptr_ += size;
+    size_ -= size;
+  }
+
   /**
    * Return remaining number of bytes to deserialize.
    *