There doesn't seem to be a way to easily change capture state within a test

[demianbrecht]

I might be missing something blatantly obvious, but it would be nice to be able to change the capture state at runtime. But first, a little context!

I have some code that fiddles with term settings. However, because capture uses a file-like object (that doesn't implement fileno), it will fail when twiddling bit flags. Trivial example:

import termios

def test_example():
    import sys

    termios.tcsetattr(
        sys.stdin, termios.TCSADRAIN, termios.tcgetattr(sys.stdin))

The above results in:

self = <_pytest.capture.DontReadFromInput instance at 0x102436440>

    def fileno(self):
>       raise ValueError("redirected Stdin is pseudofile, has no fileno()")
E       ValueError: redirected Stdin is pseudofile, has no fileno()

Of course, if you run pytest -s, all is wonderful. However, that's not an option in my case due to the drastic increase in log noise.

[codewarrior0]

There are fixture functions capfd and capsys that allow you to enable capturing for a particular test, but there doesn't seem to be one for disabling it. I'd like to have one for disabling output capture - one of my tests relies on a subprocess having its stdio connected to a terminal (so it can verify that the subprocess sets its I/O encodings correctly).

[nicoddemus]

Something like what's been discussed in #167?

[demianbrecht]

@nicoddemus: Yep, looks like the solution to that would also solve this as well. I'll add a note to #167 and close this one.

[nicoddemus]

Thanks @demianbrecht