Merge pull request libgit2#1652 from jcansdale/fixes/git-diff-indent-heuristic

Add support for git diff --indent-heuristic option

Author: ethomson

LibGit2Sharp.Tests/DiffBlobToBlobFixture.cs

@@ -1,4 +1,5 @@
 using System.IO;
+using System.Linq;
 using System.Text;
 using LibGit2Sharp.Tests.TestHelpers;
 using Xunit;
@@ -126,5 +127,85 @@ public void ComparingTwoNullBlobsReturnsAnEmptyContentChanges()
                 Assert.Equal(0, changes.LinesDeleted);
             }
         }
+
+        [Fact]
+        public void ComparingBlobsWithNoSpacesAndIndentHeuristicOptionMakesADifference()
+        {
+            var path = SandboxStandardTestRepoGitDir();
+            using (var repo = new Repository(path))
+            {
+                // Based on test diff indent heuristic from:
+                // https://github.com/git/git/blob/433860f3d0beb0c6f205290bd16cda413148f098/t/t4061-diff-indent.sh#L17
+                var oldContent =
+@"	1
+	2
+	a
+
+	b
+	3
+	4";
+                var newContent =
+@"	1
+	2
+	a
+
+	b
+	a
+
+	b
+	3
+	4";
+                var oldBlob = repo.ObjectDatabase.CreateBlob(new MemoryStream(Encoding.UTF8.GetBytes(oldContent)));
+                var newBlob = repo.ObjectDatabase.CreateBlob(new MemoryStream(Encoding.UTF8.GetBytes(newContent)));
+                var noIndentHeuristicOption = new CompareOptions { IndentHeuristic = false };
+                var indentHeuristicOption = new CompareOptions { IndentHeuristic = true };
+
+                ContentChanges changes0 = repo.Diff.Compare(oldBlob, newBlob, noIndentHeuristicOption);
+                ContentChanges changes1 = repo.Diff.Compare(oldBlob, newBlob, indentHeuristicOption);
+
+                Assert.NotEqual(changes0.Patch, changes1.Patch);
+                Assert.Equal(CanonicalChangedLines(changes0), CanonicalChangedLines(changes1));
+            }
+        }
+
+        [Fact]
+        public void ComparingBlobsWithNoSpacesIndentHeuristicOptionMakesNoDifference()
+        {
+            var path = SandboxStandardTestRepoGitDir();
+            using (var repo = new Repository(path))
+            {
+                var oldContent =
+@"	1
+	2
+	a
+	b
+	3
+	4";
+                var newContent =
+@"	1
+	2
+	a
+	b
+	a
+	b
+	3
+	4";
+                var oldBlob = repo.ObjectDatabase.CreateBlob(new MemoryStream(Encoding.UTF8.GetBytes(oldContent)));
+                var newBlob = repo.ObjectDatabase.CreateBlob(new MemoryStream(Encoding.UTF8.GetBytes(newContent)));
+                var noIndentHeuristicOption = new CompareOptions { IndentHeuristic = false };
+                var indentHeuristicOption = new CompareOptions { IndentHeuristic = true };
+
+                ContentChanges changes0 = repo.Diff.Compare(oldBlob, newBlob, noIndentHeuristicOption);
+                ContentChanges changes1 = repo.Diff.Compare(oldBlob, newBlob, indentHeuristicOption);
+
+                Assert.Equal(changes0.Patch, changes1.Patch);
+            }
+        }
+
+        static string CanonicalChangedLines(ContentChanges changes)
+        {
+            // Create an ordered representation of lines that have been added or removed
+            return string.Join("\n", changes.Patch.Split('\n').Where(l => l.StartsWith("+") || l.StartsWith("-")).OrderBy(l => l));
+        }
     }
 }

LibGit2Sharp/CompareOptions.cs

@@ -44,5 +44,11 @@ public CompareOptions()
         /// By default, <see cref="DiffAlgorithm.Myers"/> will be used.
         /// </summary>
         public DiffAlgorithm Algorithm { get; set; }
+
+        /// <summary>
+        /// Enable --indent-heuristic Diff option, that attempts to produce more aesthetically pleasing diffs.
+        /// By default, this option will be false.
+        /// </summary>
+        public bool IndentHeuristic { get; set; }
     }
 }

LibGit2Sharp/Core/GitDiff.cs

@@ -133,6 +133,13 @@ internal enum GitDiffOptionFlags
          * Options controlling how output will be generated
          */
 
+        /// <summary>
+        /// Use a heuristic that takes indentation and whitespace into account
+        /// which generally can produce better diffs when dealing with ambiguous
+        /// diff hunks.
+        /// </summary>
+        GIT_DIFF_INDENT_HEURISTIC = (1 << 18),
+
         /// <summary>
         /// Treat all files as text, disabling binary attributes and detection
         /// </summary>
@@ -304,11 +311,11 @@ enum GitDiffLineOrigin : byte
 
     enum GitDiffFormat
     {
-        GIT_DIFF_FORMAT_PATCH        = 1, // < full git diff
+        GIT_DIFF_FORMAT_PATCH = 1, // < full git diff
         GIT_DIFF_FORMAT_PATCH_HEADER = 2, // < just the file headers of patch
-        GIT_DIFF_FORMAT_RAW          = 3, // < like git diff --raw
-        GIT_DIFF_FORMAT_NAME_ONLY    = 4, // < like git diff --name-only
-        GIT_DIFF_FORMAT_NAME_STATUS  = 5, // < like git diff --name-status
+        GIT_DIFF_FORMAT_RAW = 3, // < like git diff --raw
+        GIT_DIFF_FORMAT_NAME_ONLY = 4, // < like git diff --name-only
+        GIT_DIFF_FORMAT_NAME_STATUS = 5, // < like git diff --name-status
     }
 
     [Flags]

LibGit2Sharp/Diff.cs

@@ -63,6 +63,11 @@ private static GitDiffOptions BuildOptions(DiffModifiers diffOptions, FilePath[]
                 options.Flags |= GitDiffOptionFlags.GIT_DIFF_DISABLE_PATHSPEC_MATCH;
             }
 
+            if (compareOptions.IndentHeuristic)
+            {
+                options.Flags |= GitDiffOptionFlags.GIT_DIFF_INDENT_HEURISTIC;
+            }
+
             if (matchedPathsAggregator != null)
             {
                 options.NotifyCallback = matchedPathsAggregator.OnGitDiffNotify;
@@ -351,7 +356,7 @@ public virtual T Compare<T>(Tree oldTree, DiffTargets diffTargets, IEnumerable<s
             }
 
             DiffHandle diff = BuildDiffList(oldTreeId, null, comparer, diffOptions, paths, explicitPathsOptions, compareOptions);
-            
+
             try
             {
                 return BuildDiffResult<T>(diff);