The execution results of dmypy are not as expected

Operation steps

1. I start a daemon server

    command = [
        'dmypy', '--status-file', status_file_name, 'start', '--log-file', mypy_server_log_name,
        '--', '--show-column-numbers', '--show-error-end'
    ]
    process = subprocess.Popen(command, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
    out, err = process.communicate()

2. use command line to check

dmypy --status-file xxxx.json check -v xxx.py

It can output the correct result

3. use python to write file,and use command line to check again

with open(path, 'w') as file:
    file.write(content)

It can not output any issue,like:

Success: no issues found in 1 source file
err                     : 
out                     : Success: no issues found in 1 source file\n
platform                :     darwin
python_version          :       3_11
roundtrip_time          :      1.148
stats                   : {'find_module_time': 0.006188869476318359, 'find_module_calls': 673, 'update_isolated_time': 0.0071163177490234375, 'propagate_time': 1.1365551948547363, 'find_changes_time': 0.00043392181396484375, 'fg_update_time': 6.198883056640625e-06, 'refresh_suppressed_time': 0.0006737709045410156, 'find_added_supressed_time': 0.0018012523651123047, 'cleanup_time': 1.1443119049072266}
status   

4. BUT if I use mypy instead of dmypy，It can output correct result：

mypy xxx.py

Found 8 errors in 6 files (checked 1 source file)
err                     : 
out                     : c8f34f1c-8173-4437-b07f-8d7f8c7f7210/FLO ... 1725 more characters
platform                :     darwin
python_version          :       3_11
roundtrip_time          :      0.160
stats                   : {'validate_meta_time': 8.96453857421875e-05, 'files_parsed': 34, 'modules_parsed': 22, 'stubs_parsed': 12, 'parse_time': 0.01146078109741211, 'find_module_time': 0.01342916488647461, 'find_module_calls': 3688, 'semanal_time': 0.018131256103515625, 'typecheck_time': 0.014755964279174805, 'finish_passes_time': 0.007950305938720703, 'load_fg_deps_time': 1.5020370483398438e-05, 'update_isolated_time': 0.1284799575805664, 'propagate_time': 0.02331995964050293, 'find_changes_time': 0.0006420612335205078, 'fg_update_time': 0.1564321517944336, 'refresh_suppressed_time': 0.0005068778991699219, 'find_added_supressed_time': 0.0013458728790283203, 'cleanup_time': 0.0001461505889892578}
status                  :          1

5. If I manually change the python file slightly，the dmypy can ouput correct result again

Your Environment

• Mypy version used: 1.10.1

Is there anything special about the caching logic of this dmypy? Will all information be lost if the file is regenerated?

Reproduce the reported sequence with dmypy start, dmypy check, the generated Python file, and the status file, then compare it with mypy on version 1.10.1. Start by investigating dmypy's incremental caching behavior and whether rewriting the file invalidates its cached state. Done means dmypy reports the same type errors as mypy after the file is regenerated.