From 8a2f019d69324f47331afb87d6f7cd85fc1cfe68 Mon Sep 17 00:00:00 2001
From: Zettat123 <zettat123@gmail.com>
Date: Fri, 4 Aug 2023 10:53:15 +0800
Subject: [PATCH] Support getting changed files when commit ID is `EmptySHA`
 (#26290)

Fixes #26270.

Co-Author: @wxiaoguang

Thanks @lunny for providing this solution

As
https://github.com/go-gitea/gitea/issues/26270#issuecomment-1661695151
said, at present we cannot get the names of changed files correctly when
the `OldCommitID` is `EmptySHA`. In this PR, the `GetCommitFilesChanged`
method is added and will be used to get the changed files by commit ID.

References:
- https://stackoverflow.com/a/424142

---------

Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
---
 modules/git/repo_compare.go      | 10 +++++++-
 modules/git/repo_compare_test.go | 39 ++++++++++++++++++++++++++++++++
 modules/git/sha1.go              |  4 ++--
 3 files changed, 50 insertions(+), 3 deletions(-)

diff --git a/modules/git/repo_compare.go b/modules/git/repo_compare.go
index e70627585..aad725fa9 100644
--- a/modules/git/repo_compare.go
+++ b/modules/git/repo_compare.go
@@ -280,8 +280,16 @@ func (repo *Repository) GetPatch(base, head string, w io.Writer) error {
 }
 
 // GetFilesChangedBetween returns a list of all files that have been changed between the given commits
+// If base is undefined empty SHA (zeros), it only returns the files changed in the head commit
+// If base is the SHA of an empty tree (EmptyTreeSHA), it returns the files changes from the initial commit to the head commit
 func (repo *Repository) GetFilesChangedBetween(base, head string) ([]string, error) {
-	stdout, _, err := NewCommand(repo.Ctx, "diff", "--name-only", "-z").AddDynamicArguments(base + ".." + head).RunStdString(&RunOpts{Dir: repo.Path})
+	cmd := NewCommand(repo.Ctx, "diff-tree", "--name-only", "--root", "--no-commit-id", "-r", "-z")
+	if base == EmptySHA {
+		cmd.AddDynamicArguments(head)
+	} else {
+		cmd.AddDynamicArguments(base, head)
+	}
+	stdout, _, err := cmd.RunStdString(&RunOpts{Dir: repo.Path})
 	if err != nil {
 		return nil, err
 	}
diff --git a/modules/git/repo_compare_test.go b/modules/git/repo_compare_test.go
index 5b50bc82a..603aabde4 100644
--- a/modules/git/repo_compare_test.go
+++ b/modules/git/repo_compare_test.go
@@ -119,3 +119,42 @@ func TestReadWritePullHead(t *testing.T) {
 	err = repo.RemoveReference(PullPrefix + "1/head")
 	assert.NoError(t, err)
 }
+
+func TestGetCommitFilesChanged(t *testing.T) {
+	bareRepo1Path := filepath.Join(testReposDir, "repo1_bare")
+	repo, err := openRepositoryWithDefaultContext(bareRepo1Path)
+	assert.NoError(t, err)
+	defer repo.Close()
+
+	testCases := []struct {
+		base, head string
+		files      []string
+	}{
+		{
+			EmptySHA,
+			"95bb4d39648ee7e325106df01a621c530863a653",
+			[]string{"file1.txt"},
+		},
+		{
+			EmptySHA,
+			"8d92fc957a4d7cfd98bc375f0b7bb189a0d6c9f2",
+			[]string{"file2.txt"},
+		},
+		{
+			"95bb4d39648ee7e325106df01a621c530863a653",
+			"8d92fc957a4d7cfd98bc375f0b7bb189a0d6c9f2",
+			[]string{"file2.txt"},
+		},
+		{
+			EmptyTreeSHA,
+			"8d92fc957a4d7cfd98bc375f0b7bb189a0d6c9f2",
+			[]string{"file1.txt", "file2.txt"},
+		},
+	}
+
+	for _, tc := range testCases {
+		changedFiles, err := repo.GetFilesChangedBetween(tc.base, tc.head)
+		assert.NoError(t, err)
+		assert.ElementsMatch(t, tc.files, changedFiles)
+	}
+}
diff --git a/modules/git/sha1.go b/modules/git/sha1.go
index 7d9d9776d..8d6403e86 100644
--- a/modules/git/sha1.go
+++ b/modules/git/sha1.go
@@ -11,10 +11,10 @@ import (
 	"strings"
 )
 
-// EmptySHA defines empty git SHA
+// EmptySHA defines empty git SHA (undefined, non-existent)
 const EmptySHA = "0000000000000000000000000000000000000000"
 
-// EmptyTreeSHA is the SHA of an empty tree
+// EmptyTreeSHA is the SHA of an empty tree, the root of all git repositories
 const EmptyTreeSHA = "4b825dc642cb6eb9a060e54bf8d69288fbee4904"
 
 // SHAFullLength is the full length of a git SHA