python runtime_error ai_generated true

AssertionError: assert '' == 'hello\n' - + hello

ID: python/pytest-capsys-flush-buffering

Also available as: JSON · Markdown · 中文
80%Fix Rate
85%Confidence
0Evidence
2024-10-05First Seen

Version Compatibility

VersionStatusIntroducedDeprecatedNotes
7.x active — — —
8.x active — — —

Root Cause

`capsys.readouterr()` returns empty because the code under test uses `print(..., flush=False)` and the buffer has not been flushed, or it writes to `sys.__stdout__` instead of `sys.stdout`.

generic

中文

`capsys.readouterr()` 返回空，因为被测代码使用了 `print(..., flush=False)` 且缓冲区尚未刷新，或者它写入的是 `sys.__stdout__` 而不是 `sys.stdout`。

Workarounds

  1. 95% success 
    Force flush in the code under test or use print(..., flush=True):

print('hello', flush=True)
  2. 90% success 
    Use capfd and ensure the code writes to sys.stdout:

def test_out(capfd):
    my_func()
    out, err = capfd.readouterr()
    assert out == 'hello\n'

Dead Ends

Common approaches that don't work:

  1. 80% fail

    Flushing is not time-based; the buffer may still be unflushed and the test is now flaky.

  2. 60% fail

    capfd captures file descriptors; if the code uses sys.__stdout__ directly it still bypasses both.