fix: use missing_ok for temp file cleanup to avoid masking errors

[Quratulain-bilal]

Replace if temp_path.exists(): temp_path.unlink() with emp_path.unlink(missing_ok=True) in 3 files: integrations/manifest.py, shared_infra.py, _utils.py. The old pattern could raise OSError if unlink() fails, masking the original exception in inally/except blocks.

[Copilot]

Pull request overview

This PR attempts to prevent temporary-file cleanup from masking write failures.

Changes:

• Replaces existence checks with Path.unlink(missing_ok=True) in three atomic-write paths.
• However, non-FileNotFoundError cleanup failures can still mask original exceptions.

Show a summary per file

File	Description
src/specify_cli/shared_infra.py	Updates shared-file cleanup.
src/specify_cli/integrations/manifest.py	Updates manifest cleanup.
src/specify_cli/_utils.py	Updates settings JSON cleanup.

Review details

Tip

Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

• Files reviewed: 3/3 changed files
• Comments generated: 3
• Review effort level: Medium

[Copilot]

Review details

Suppressed comments (3)

src/specify_cli/shared_infra.py:268

• Add a regression test for this new cleanup-failure path. Have the write/replace operation raise a sentinel exception and Path.unlink raise OSError, then assert the sentinel remains the propagated error; otherwise the behavior this PR is intended to guarantee can regress unnoticed.

        try:
            temp_path.unlink(missing_ok=True)
        except OSError:
            pass

src/specify_cli/integrations/manifest.py:445

• Add a manifest persistence regression test that forces the save operation to fail and the temporary-file unlink to raise OSError, and verifies the original save error is preserved. The existing persistence tests cover successful round trips but not the masking scenario introduced here.

            try:
                temp_path.unlink(missing_ok=True)
            except OSError:
                pass

src/specify_cli/_utils.py:199

• Please cover this exception branch by making os.replace fail with a distinguishable original error and cleanup unlink fail with OSError, then asserting handle_vscode_settings still propagates the replace error. Current settings tests only exercise a successful atomic write, so this fix has no regression guard.

            if temp_path:
                try:
                    temp_path.unlink(missing_ok=True)
                except OSError:
                    pass

• Files reviewed: 3/3 changed files
• Comments generated: 0 new
• Review effort level: Balanced

[Copilot]

Review details

Suppressed comments (3)

src/specify_cli/integrations/manifest.py:457

• Add a regression test for save() where the manifest write fails and cleanup also raises OSError, asserting that callers still receive the original write failure. Current manifest persistence tests only cover successful saves, so this new exception behavior is unverified.

            try:
                temp_path.unlink(missing_ok=True)
            except OSError:
                pass

src/specify_cli/shared_infra.py:268

• Add a regression test that makes the primary write fail and the temporary-file unlink raise OSError, then asserts the primary exception is preserved. The existing shared-infrastructure tests cover normal and refresh-failure paths, but none exercises the error-masking behavior this branch is intended to fix.

        try:
            temp_path.unlink(missing_ok=True)
        except OSError:
            pass

src/specify_cli/_utils.py:199

• Please cover this error path by forcing os.replace to fail and temporary-file cleanup to raise OSError, then asserting that handle_vscode_settings propagates the original replace error. The current atomic-write test verifies mode preservation only, so the stated regression could return unnoticed.

            if temp_path:
                try:
                    temp_path.unlink(missing_ok=True)
                except OSError:
                    pass

• Files reviewed: 3/3 changed files
• Comments generated: 0 new
• Review effort level: Balanced