From 3172ce264b8b54083b73461c08e87bb888d34a2a Mon Sep 17 00:00:00 2001
From: sladro <sladro@163.com>
Date: Thu, 2 Apr 2026 11:52:10 +0800
Subject: [PATCH] fix: preserve artifact line endings and restore deletions

---
 engine/artifact_manager.py     | 17 +++++++++---
 tests/test_artifact_manager.py | 51 ++++++++++++++++++++++++++++++++++
 2 files changed, 64 insertions(+), 4 deletions(-)

diff --git a/engine/artifact_manager.py b/engine/artifact_manager.py
index bcfe4c7..7034458 100644
--- a/engine/artifact_manager.py
+++ b/engine/artifact_manager.py
@@ -1,10 +1,10 @@
 from __future__ import annotations
 
 from dataclasses import dataclass
+from difflib import unified_diff
 from fnmatch import fnmatch
 from hashlib import sha256
 from pathlib import Path
-from difflib import unified_diff
 
 from engine.models import BaselineSnapshot, TaskSpec
 
@@ -30,19 +30,28 @@ class ArtifactManager:
         file_contents: dict[Path, str] = {}
         file_hashes: dict[Path, str] = {}
         for path in self.resolve_paths():
-            content = path.read_text(encoding="utf-8")
+            with path.open("r", encoding="utf-8", newline="") as handle:
+                content = handle.read()
             file_contents[path] = content
             file_hashes[path] = sha256(content.encode("utf-8")).hexdigest()
         return BaselineSnapshot(file_contents=file_contents, file_hashes=file_hashes)
 
     def restore(self, snapshot: BaselineSnapshot) -> None:
+        current_paths = set(self.resolve_paths())
+        snapshot_paths = set(snapshot.file_contents)
+        for path in current_paths - snapshot_paths:
+            path.unlink()
         for path, content in snapshot.file_contents.items():
             path.parent.mkdir(parents=True, exist_ok=True)
-            path.write_text(content, encoding="utf-8")
+            with path.open("w", encoding="utf-8", newline="") as handle:
+                handle.write(content)
 
     def diff_summary(self, snapshot: BaselineSnapshot) -> str:
         lines: list[str] = []
-        current_contents = {path: path.read_text(encoding="utf-8") for path in self.resolve_paths()}
+        current_contents: dict[Path, str] = {}
+        for path in self.resolve_paths():
+            with path.open("r", encoding="utf-8", newline="") as handle:
+                current_contents[path] = handle.read()
         all_paths = sorted(set(snapshot.file_contents) | set(current_contents))
         for path in all_paths:
             before = snapshot.file_contents.get(path, "")
diff --git a/tests/test_artifact_manager.py b/tests/test_artifact_manager.py
index 0eca28a..a7ca848 100644
--- a/tests/test_artifact_manager.py
+++ b/tests/test_artifact_manager.py
@@ -1,4 +1,5 @@
 from pathlib import Path
+from hashlib import sha256
 import tempfile
 import unittest
 
@@ -42,6 +43,56 @@ class ArtifactManagerTest(unittest.TestCase):
             manager.restore(snapshot)
             self.assertEqual(target.read_text(encoding="utf-8"), "hello\n")
 
+    def test_restore_removes_newly_created_included_artifact(self) -> None:
+        with tempfile.TemporaryDirectory() as tmp:
+            root = Path(tmp)
+            artifact_dir = root / "artifacts"
+            artifact_dir.mkdir()
+            target = artifact_dir / "sample.md"
+            target.write_text("hello\n", encoding="utf-8")
+            manager = ArtifactManager(make_task(root))
+            snapshot = manager.snapshot()
+            extra = artifact_dir / "new.md"
+            extra.write_text("new\n", encoding="utf-8")
+            manager.restore(snapshot)
+            self.assertFalse(extra.exists())
+            self.assertEqual(target.read_text(encoding="utf-8"), "hello\n")
+
+    def test_snapshot_and_restore_preserve_crlf_content(self) -> None:
+        with tempfile.TemporaryDirectory() as tmp:
+            root = Path(tmp)
+            artifact_dir = root / "artifacts"
+            artifact_dir.mkdir()
+            target = artifact_dir / "sample.md"
+            original = "line1\r\nline2\r\n"
+            with target.open("w", encoding="utf-8", newline="") as handle:
+                handle.write(original)
+            manager = ArtifactManager(make_task(root))
+            snapshot = manager.snapshot()
+            self.assertEqual(snapshot.file_contents[target], original)
+            self.assertEqual(
+                snapshot.file_hashes[target],
+                sha256(original.encode("utf-8")).hexdigest(),
+            )
+            with target.open("w", encoding="utf-8", newline="") as handle:
+                handle.write("changed\r\n")
+            manager.restore(snapshot)
+            with target.open("r", encoding="utf-8", newline="") as handle:
+                restored = handle.read()
+            self.assertEqual(restored, original)
+
+    def test_resolve_paths_is_deterministic_and_respects_excludes(self) -> None:
+        with tempfile.TemporaryDirectory() as tmp:
+            root = Path(tmp)
+            artifact_dir = root / "artifacts"
+            artifact_dir.mkdir()
+            (artifact_dir / "b.md").write_text("b\n", encoding="utf-8")
+            (artifact_dir / "ignore.md").write_text("ignore\n", encoding="utf-8")
+            (artifact_dir / "a.md").write_text("a\n", encoding="utf-8")
+            manager = ArtifactManager(make_task(root))
+            paths = manager.resolve_paths()
+            self.assertEqual(paths, [artifact_dir / "a.md", artifact_dir / "b.md"])
+
     def test_diff_summary_contains_changed_line(self) -> None:
         with tempfile.TemporaryDirectory() as tmp:
             root = Path(tmp)