@pytest.mark.tui

Per-test configuration without depending on tui_config directly.

import pytest


@pytest.mark.tui(cols=140, rows=44, timeout=10, strict_mouse=True)
async def test_big_screen(tui):
    await tui.start("myapp")
    ...

Options

Key	Default	Purpose
cols	80	Terminal width
rows	24	Terminal height
timeout	5.0	Default wait_for_* timeout (seconds)
strict_mouse	False	Raise instead of warn when mouse input is sent before the app enabled tracking

The marker takes precedence over CLI flags (--tui-cols, --tui-rows, --tui-timeout).

When to use it vs. inline tui.start(cols=…)

The marker is right when every action in the test needs the same size:

@pytest.mark.tui(cols=160, rows=50)
async def test_wide_layout(tui):
    await tui.start("myapp")
    await tui.wait_for_text("...")
    # any subsequent resize is still possible, but starts at 160×50

The tui.start(cols=…) keyword is right when you want to override once within an otherwise default test:

async def test_resize_flow(tui):
    await tui.start("myapp", cols=80, rows=24)   # explicit start size
    await tui.resize(160, 50)                    # then grow

Common configurations

# A larger viewport for layout snapshots
@pytest.mark.tui(cols=120, rows=30)

# Strict mouse mode for a click-heavy suite
@pytest.mark.tui(strict_mouse=True)

# Slow CI runner — bump the wait timeout
@pytest.mark.tui(timeout=15)

# Combined
@pytest.mark.tui(cols=140, rows=44, timeout=10, strict_mouse=True)

Registering markers

The plugin registers tui as a known marker, so you won't get "unknown marker" warnings. If you've enabled strict markers (filterwarnings = error::pytest.PytestUnknownMarkWarning), this is seamless.

Next: CLI flags →