CM-52765 - Fix git mv handling for commit range scans

Author: MarshalX

cycode/cli/apps/scan/commit_range_scanner.py

@@ -282,7 +282,11 @@ def _scan_secret_pre_commit(ctx: typer.Context, repo_path: str) -> None:
     for diff in diff_index:
         progress_bar.update(ScanProgressBarSection.PREPARE_LOCAL_FILES)
         documents_to_scan.append(
-            Document(get_path_by_os(get_diff_file_path(diff)), get_diff_file_content(diff), is_git_diff_format=True)
+            Document(
+                get_path_by_os(get_diff_file_path(diff, repo=repo)),
+                get_diff_file_content(diff),
+                is_git_diff_format=True,
+            )
         )
     documents_to_scan = excluder.exclude_irrelevant_documents_to_scan(consts.SECRET_SCAN_TYPE, documents_to_scan)
 

cycode/cli/files_collector/commit_range_documents.py

@@ -87,7 +87,7 @@ def collect_commit_range_diff_documents(
         for diff in diff_index:
             commit_documents_to_scan.append(
                 Document(
-                    path=get_path_by_os(get_diff_file_path(diff)),
+                    path=get_path_by_os(get_diff_file_path(diff, repo=repo)),
                     content=get_diff_file_content(diff),
                     is_git_diff_format=True,
                     unique_id=commit_id,
@@ -166,7 +166,7 @@ def get_commit_range_modified_documents(
     for diff in modified_files_diff:
         progress_bar.update(progress_bar_section)
 
-        file_path = get_path_by_os(get_diff_file_path(diff))
+        file_path = get_path_by_os(get_diff_file_path(diff, repo=repo))
 
         diff_documents.append(
             Document(
@@ -211,16 +211,24 @@ def parse_pre_receive_input() -> str:
     return pre_receive_input.splitlines()[0]
 
 
-def get_diff_file_path(diff: 'Diff', relative: bool = False) -> Optional[str]:
+def get_diff_file_path(diff: 'Diff', relative: bool = False, repo: Optional['Repo'] = None) -> Optional[str]:
     if relative:
         # relative to the repository root
         return diff.b_path if diff.b_path else diff.a_path
 
+    # Try blob-based paths first (most reliable when available)
     if diff.b_blob:
         return diff.b_blob.abspath
     if diff.a_blob:
         return diff.a_blob.abspath
 
+    # Fallback: construct an absolute path from a relative path
+    # This handles renames and other cases where blobs might be None
+    if repo and repo.working_tree_dir:
+        target_path = diff.b_path if diff.b_path else diff.a_path
+        if target_path:
+            return os.path.abspath(os.path.join(repo.working_tree_dir, target_path))
+
     return None
 
 
@@ -244,7 +252,7 @@ def get_pre_commit_modified_documents(
     for diff in diff_index:
         progress_bar.update(progress_bar_section)
 
-        file_path = get_path_by_os(get_diff_file_path(diff))
+        file_path = get_path_by_os(get_diff_file_path(diff, repo=repo))
 
         diff_documents.append(
             Document(

tests/cli/files_collector/test_commit_range_documents.py

@@ -6,7 +6,11 @@
 from git import Repo
 
 from cycode.cli import consts
-from cycode.cli.files_collector.commit_range_documents import get_safe_head_reference_for_diff
+from cycode.cli.files_collector.commit_range_documents import (
+    get_diff_file_path,
+    get_safe_head_reference_for_diff,
+)
+from cycode.cli.utils.path_utils import get_path_by_os
 
 
 @contextmanager
@@ -128,3 +132,26 @@ def test_sequential_operations_on_same_repository(self) -> None:
             assert head_ref_after == consts.GIT_HEAD_COMMIT_REV
             assert len(diff_after) == 1
             assert diff_after[0].b_path == 'new.py'
+
+
+def test_git_mv_pre_commit_scan() -> None:
+    with temporary_git_repository() as (temp_dir, repo):
+        newfile_path = os.path.join(temp_dir, 'NEWFILE.txt')
+        with open(newfile_path, 'w') as f:
+            f.write('test content')
+
+        repo.index.add(['NEWFILE.txt'])
+        repo.index.commit('init')
+
+        # Rename file but don't commit (this is the pre-commit scenario)
+        renamed_path = os.path.join(temp_dir, 'RENAMED.txt')
+        os.rename(newfile_path, renamed_path)
+        repo.index.remove(['NEWFILE.txt'])
+        repo.index.add(['RENAMED.txt'])
+
+        head_ref = get_safe_head_reference_for_diff(repo)
+        diff_index = repo.index.diff(head_ref, create_patch=True, R=True)
+
+        for diff in diff_index:
+            file_path = get_path_by_os(get_diff_file_path(diff, repo=repo))
+            assert file_path == renamed_path