Fix for computing blob size without loading entire blob content in memory - Fix libgit2#873

ravindp · ravindp · commit ae654785c22f · 2014-12-04T10:08:22.000+05:30
diff --git a/LibGit2Sharp.Tests/ObjectDatabaseFixture.cs b/LibGit2Sharp.Tests/ObjectDatabaseFixture.cs
@@ -48,6 +48,25 @@ public void CanCreateABlobFromAFileInTheWorkingDirectory()
             }
         }
 
+        [Fact]
+        public void RetrieveObjectMetadataReturnsCorrectSizeAndTypeForBlob()
+        {
+            string path = InitNewRepository();
+
+            using (var repo = new Repository(path))
+            {
+                Blob blob = CreateBlob(repo, "I'm a new file\n");
+                Assert.NotNull(blob);
+
+                GitObjectMetadata blobMetadata = repo.ObjectDatabase.RetrieveObjectMetadata(blob.Id);
+                Assert.Equal(blobMetadata.Size, blob.Size);
+                Assert.Equal(blobMetadata.Type, ObjectType.Blob);
+
+                Blob fetchedBlob = repo.Lookup<Blob>(blob.Id);
+                Assert.Equal(blobMetadata.Size, fetchedBlob.Size);
+            }
+        }
+
         [Fact]
         public void CanCreateABlobIntoTheDatabaseOfABareRepository()
         {
diff --git a/LibGit2Sharp/Blob.cs b/LibGit2Sharp/Blob.cs
@@ -27,6 +27,10 @@ internal Blob(Repository repo, ObjectId id)
 
         /// <summary>
         /// Gets the size in bytes of the raw content of a blob.
+        /// <para> Please note that this would load entire blob content in the memory to compute the Size.
+        /// In order to read blob size from header, Repository.ObjectMetadata.RetrieveObjectMetadata(Blob.Id).Size
+        /// can be used.
+        /// </para>
         /// </summary>
         public virtual int Size { get { return (int)lazySize.Value; } }
 
diff --git a/LibGit2Sharp/Core/NativeMethods.cs b/LibGit2Sharp/Core/NativeMethods.cs
@@ -754,6 +754,9 @@ internal static extern int git_odb_foreach(
         [DllImport(libgit2)]
         internal static extern void git_odb_free(IntPtr odb);
 
+        [DllImport(libgit2)]
+        internal static extern int git_odb_read_header(out UIntPtr len_out, out GitObjectType type, ObjectDatabaseSafeHandle odb, ref GitOid id);
+
         [DllImport(libgit2)]
         internal static extern void git_object_free(IntPtr obj);
 
diff --git a/LibGit2Sharp/Core/Proxy.cs b/LibGit2Sharp/Core/Proxy.cs
@@ -1378,6 +1378,21 @@ public static bool git_odb_exists(ObjectDatabaseSafeHandle odb, ObjectId id)
             }
         }
 
+        public static GitObjectMetadata git_odb_read_header(ObjectDatabaseSafeHandle odb, ObjectId id)
+        {
+            using (ThreadAffinity())
+            {
+                GitOid oid = id.Oid;
+                UIntPtr length;
+                GitObjectType objectType;
+
+                int res = NativeMethods.git_odb_read_header(out length, out objectType, odb, ref oid);
+                Ensure.ZeroResult(res);
+
+                return new GitObjectMetadata((long)length, objectType);
+            }
+        }
+
         public static ICollection<TResult> git_odb_foreach<TResult>(
             ObjectDatabaseSafeHandle odb,
             Func<IntPtr, TResult> resultSelector)
diff --git a/LibGit2Sharp/GitObjectMetadata.cs b/LibGit2Sharp/GitObjectMetadata.cs
@@ -0,0 +1,35 @@
+﻿using LibGit2Sharp.Core;
+using System;
+
+namespace LibGit2Sharp
+{
+    /// <summary>
+    /// Exposes low level Git object metadata
+    /// </summary>
+    public sealed class GitObjectMetadata
+    {
+        private readonly GitObjectType type;
+
+        /// <summary>
+        /// Size of the Object 
+        /// </summary>
+        public long Size { get; private set; }
+
+        /// <summary>
+        /// Object Type
+        /// </summary>
+        public ObjectType Type
+        {
+            get
+            {
+                return type.ToObjectType();
+            }
+        }
+
+        internal GitObjectMetadata(long size, GitObjectType type)
+        {
+            this.Size = size;
+            this.type = type;
+        }
+    }
+}
diff --git a/LibGit2Sharp/LibGit2Sharp.csproj b/LibGit2Sharp/LibGit2Sharp.csproj
@@ -95,6 +95,7 @@
     <Compile Include="MergeOptions.cs" />
     <Compile Include="MergeResult.cs" />
     <Compile Include="NotFoundException.cs" />
+    <Compile Include="GitObjectMetadata.cs" />
     <Compile Include="PatchEntryChanges.cs" />
     <Compile Include="PatchStats.cs" />
     <Compile Include="PullOptions.cs" />
diff --git a/LibGit2Sharp/ObjectDatabase.cs b/LibGit2Sharp/ObjectDatabase.cs
@@ -72,6 +72,20 @@ public virtual bool Contains(ObjectId objectId)
             return Proxy.git_odb_exists(handle, objectId);
         }
 
+        /// <summary>
+        /// Retrieves the header of a GitObject from the object database. The header contains the Size
+        /// and Type of the object. Note that most backends do not support reading only the header
+        /// of an object, so the whole object will be read and then size would be returned.  
+        /// </summary>
+        /// <param name="objectId">Object Id of the queried object</param>
+        /// <returns>GitObjectMetadata object instance containg object header information</returns>
+        public virtual GitObjectMetadata RetrieveObjectMetadata(ObjectId objectId)
+        {
+            Ensure.ArgumentNotNull(objectId, "objectId");
+
+            return Proxy.git_odb_read_header(handle, objectId);
+        }
+
         /// <summary>
         /// Inserts a <see cref="Blob"/> into the object database, created from the content of a file.
         /// </summary>
