From 5bc7bdd2b5238db543c5eb8550a5cd52542d4df3 Mon Sep 17 00:00:00 2001
From: RerankerGuo <121015044+RerankerGuo@users.noreply.github.com>
Date: Wed, 12 Aug 2026 16:09:33 +0800
Subject: [PATCH] fix(tool): preserve special paths in file_find

Use NUL-delimited Git output so Unicode and special-character paths are not C-quoted or split during file discovery. Add workspace and commit-mode regression coverage.

Test: make check && make test && make build && make coverage
---
 internal/tool/file_find.go      | 12 ++---
 internal/tool/file_find_test.go | 83 +++++++++++++++++++++++++++++++++
 2 files changed, 89 insertions(+), 6 deletions(-)

diff --git a/internal/tool/file_find.go b/internal/tool/file_find.go
index 44c2b955d..51bb72811 100644
--- a/internal/tool/file_find.go
+++ b/internal/tool/file_find.go
@@ -79,9 +79,9 @@ func (p *FileFindProvider) listGitFiles(parentCtx context.Context) ([]string, er
 
 	var args []string
 	if ref := p.FileReader.Ref; ref != "" {
-		args = []string{"ls-tree", "-r", "--name-only", "--end-of-options", ref}
+		args = []string{"ls-tree", "-r", "--name-only", "-z", "--end-of-options", ref}
 	} else {
-		args = []string{"ls-files", "--cached", "--others", "--exclude-standard"}
+		args = []string{"ls-files", "-z", "--cached", "--others", "--exclude-standard"}
 	}
 
 	if p.FileReader.Runner != nil {
@@ -107,10 +107,10 @@ func (p *FileFindProvider) listGitFiles(parentCtx context.Context) ([]string, er
 	}
 
 	var files []string
-	lines := bytes.Split(bytes.TrimRight(output, "\n"), []byte{'\n'})
-	for _, line := range lines {
-		if len(line) > 0 {
-			s := string(line)
+	paths := bytes.Split(output, []byte{0})
+	for _, path := range paths {
+		if len(path) > 0 {
+			s := string(path)
 			// Skip binary-like files that lack meaningful extensions patterns
 			// and filter out paths in common generated/artifact directories.
 			if shouldSkipFile(s) {
diff --git a/internal/tool/file_find_test.go b/internal/tool/file_find_test.go
index 518b259ab..f33fc8416 100644
--- a/internal/tool/file_find_test.go
+++ b/internal/tool/file_find_test.go
@@ -8,6 +8,7 @@ import (
 	"os"
 	"os/exec"
 	"path/filepath"
+	"runtime"
 	"strings"
 	"testing"
 
@@ -159,6 +160,88 @@ func TestFileFind_GitRepo_CommitMode(t *testing.T) {
 	}
 }
 
+func TestFileFind_GitRepo_PreservesSpecialPaths(t *testing.T) {
+	if runtime.GOOS == "windows" {
+		t.Skip("test paths contain characters unsupported by Windows")
+	}
+
+	dir := setupFileFindRepo(t)
+	specialPaths := []string{
+		"unicode-测试.go",
+		"tab\tquote\"file.go",
+		"line\nbreak.go",
+	}
+	for _, path := range specialPaths {
+		if err := os.WriteFile(filepath.Join(dir, path), []byte("package special\n"), 0o644); err != nil {
+			t.Fatal(err)
+		}
+	}
+
+	git := func(args ...string) string {
+		t.Helper()
+		cmd := exec.Command("git", args...)
+		cmd.Dir = dir
+		out, err := cmd.CombinedOutput()
+		if err != nil {
+			t.Fatalf("git %v: %v\n%s", args, err, out)
+		}
+		return strings.TrimSpace(string(out))
+	}
+	git("add", ".")
+	git("commit", "-m", "add special paths")
+	commit := git("rev-parse", "HEAD")
+
+	tests := []struct {
+		name string
+		mode ReviewMode
+		ref  string
+	}{
+		{name: "workspace", mode: ModeWorkspace},
+		{name: "commit", mode: ModeCommit, ref: commit},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			p := NewFileFind(&FileReader{RepoDir: dir, Mode: tt.mode, Ref: tt.ref})
+
+			files, err := p.listGitFiles(context.Background())
+			if err != nil {
+				t.Fatal(err)
+			}
+			// Newline paths are asserted only at the enumeration layer because
+			// Execute intentionally joins multiple results with newlines.
+			for _, want := range append([]string{"main.go"}, specialPaths...) {
+				found := false
+				for _, got := range files {
+					if got == want {
+						found = true
+						break
+					}
+				}
+				if !found {
+					t.Errorf("listGitFiles() missing original path %q; got %q", want, files)
+				}
+			}
+
+			for _, query := range []struct {
+				name string
+				want string
+			}{
+				{name: "main.go", want: "main.go"},
+				{name: "unicode-", want: "unicode-测试.go"},
+				{name: "quote", want: "tab\tquote\"file.go"},
+			} {
+				got, err := p.Execute(context.Background(), map[string]any{"query_name": query.name})
+				if err != nil {
+					t.Fatal(err)
+				}
+				if got != query.want {
+					t.Errorf("Execute(query_name=%q) = %q, want %q", query.name, got, query.want)
+				}
+			}
+		})
+	}
+}
+
 func TestFileFind_GitRepo_WithRunner(t *testing.T) {
 	dir := setupFileFindRepo(t)
 	runner := gitcmd.New(4)