make stdout and stderr decode handle errors by replacing with unicode…

… replacement char; add tests

packages/python-sdk/e2b/sandbox_sync/commands/command_handle.py

@@ -65,11 +65,11 @@ def _handle_events(
         for event in self._events:
             if event.event.HasField("data"):
                 if event.event.data.stdout:
-                    out = event.event.data.stdout.decode()
+                    out = event.event.data.stdout.decode('utf-8', 'replace')
                     self._stdout += out
                     yield out, None, None
                 if event.event.data.stderr:
-                    out = event.event.data.stderr.decode()
+                    out = event.event.data.stderr.decode('utf-8', 'replace')
                     self._stderr += out
                     yield None, out, None
                 if event.event.data.pty:

packages/python-sdk/tests/sync/sandbox_sync/commands/test_run.py

@@ -20,6 +20,14 @@ def test_run_with_special_characters(sandbox: Sandbox):
     assert cmd.exit_code == 0
     assert cmd.stdout == f"{text}\n"
 
+def test_run_with_broken_utf8(sandbox: Sandbox):
+    # Create a string with 8191 'a' characters followed by the problematic byte 0xe2
+    long_str = 'a' * 8191 + '\\xe2'
+    result = sandbox.commands.run(f'printf "{long_str}"')
+    assert result.exit_code == 0
+
+    # The broken UTF-8 bytes should be replaced with the Unicode replacement character
+    assert result.stdout == ('a' * 8191 + '\ufffd')
 
 def test_run_with_multiline_string(sandbox):
     text = "Hello,\nWorld!"