Expanding a Collapsible and pressing Tab within one frame scrolls the newly focused widget off screen
Nobody has claimed this yet.
Assessment
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Newbie friendliness
- 42/100
Research direction
Start with Collapsible._watch_collapsed, which queues call_after_refresh(self.scroll_visible), and compare it with Screen.set_focus, which already checks if self.focused is widget before its own deferred scroll. The repro in the issue (run via asyncio with Pilot.press("enter", "tab")) is the test to reproduce against. Done means the newly focused widget stays on screen across repeated runs at low TEXTUAL_FPS, with a regression test for the one-frame Tab case; the maintainer has not yet picked between the two proposed fixes.
Written by the indexing model from the issue text.
Description
Hi! I hit this while writing generated tests against Textual's widgets: the same key presses ended at different scroll positions from run to run.
Expanding a Collapsible queues call_after_refresh(self.scroll_visible) (in Collapsible._watch_collapsed). If a Tab arrives before the next screen refresh, Screen.set_focus focuses the next widget and queues its own scroll with call_later. That scroll runs first, and then the Collapsible's scroll_visible runs on the refresh and scrolls back to the top of the first Collapsible. The focused widget ends up off screen.
set_focus already guards its own callback (if self.focused is widget ...), but the Collapsible's callback doesn't check whether anything has happened since.
import asyncio
import collections
import sys
import textual
from textual.app import App, ComposeResult
from textual.widgets import Collapsible, Static
class CollapsibleApp(App):
def compose(self) -> ComposeResult:
with Collapsible(title="First"):
yield Static("\n".join(f"line {n}" for n in range(40)))
with Collapsible(title="Second"):
yield Static("second contents")
async def once() -> tuple[str, float, bool]:
app = CollapsibleApp()
async with app.run_test(size=(40, 12)) as pilot:
await pilot.pause()
await pilot.press("enter", "tab")
await pilot.pause(0.5)
focused = app.focused
on_screen = focused.region.overlaps(app.screen.region)
return focused.parent.title, app.screen.scroll_y, on_screen
async def main() -> None:
runs = int(sys.argv[1]) if len(sys.argv) > 1 else 20
tally = collections.Counter()
for _ in range(runs):
tally[await once()] += 1
print(f"Textual {textual.__version__}, TEXTUAL_FPS={textual.constants.MAX_FPS}, {runs} runs")
for (title, scroll_y, on_screen), count in sorted(tally.items()):
print(f" focused={title!r} scroll_y={scroll_y} focused widget on screen={on_screen}: {count}")
asyncio.run(main())
At the default 60 fps the two keys nearly always land in the same frame window the right way round. Lowering the frame rate widens the window and shows both outcomes. Output on main (5e5b7ef), Python 3.13, macOS, 30 runs at each rate:
Textual 8.2.8, TEXTUAL_FPS=60, 30 runs
focused='Second' scroll_y=35 focused widget on screen=True: 30
Textual 8.2.8, TEXTUAL_FPS=10, 30 runs
focused='Second' scroll_y=0.0 focused widget on screen=False: 23
focused='Second' scroll_y=35 focused widget on screen=True: 7
Textual 8.2.8, TEXTUAL_FPS=2, 30 runs
focused='Second' scroll_y=0.0 focused widget on screen=False: 30
Textual 8.2.7 gives the same pattern (at 10 fps: 20 off screen, 10 on screen).
In a real terminal I think this needs both keys to arrive within one refresh, for example from a fast typist, a key macro, or input batched over SSH. In tests, Pilot.press("enter", "tab") waits only for wait_for_idle(0) between keys, not for pending call_after_refresh callbacks, so the result depends on the clock.
I expected the newly focused widget to stay in view, whatever the timing. Two ideas, either of which would fix the repro:
- in
Collapsible._watch_collapsed, skip the deferredscroll_visibleif focus has moved out of the Collapsible since it was queued, asset_focusdoes for its own callback; - or have
Pilot.presswait for the screen's pending after-refresh callbacks between keys, so tests are deterministic (this doesn't help real input).
I'm happy to send a PR for whichever you'd prefer, or to leave it if you consider the burst case acceptable.
- Dominant language
- Python
- Stars
- 37.7k
- Forks
- 1.5k
- PR merge metrics
- No merged PRs in 30d
Getting set up
- No Dockerfile or Docker Compose file
- Has a pull request template
- Read the contributing guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from Textualize/textual
-
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
Textualize/textual#6790 · 2 comments ·
-
Difficulty 2/5 1-3 hours Newbie friendliness 80/100
Textualize/textual#6789 · 1 comment ·
-
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
Textualize/textual#6787 · 1 comment ·
-
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
Textualize/textual#6782 · 1 comment ·
-
Difficulty 1/5 Under an hour Newbie friendliness 85/100
Textualize/textual#6778 · 1 comment ·
All issues in Textualize/textual
Similar issues
-
Difficulty 2/5 1-3 hours Newbie friendliness 76/100
GoogleCloudPlatform/professional-services-data-validator#1899 ·
Maintainers usually reply within 4 days
-
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
mohanraj00/entelos#200 ·
Maintainers usually reply within 1 day
-
Difficulty 1/5 Under an hour Newbie friendliness 68/100
-
sweep
Difficulty 2/5 1-3 hours Newbie friendliness 85/100
bketelsen/truenas-mcp#77 ·
Maintainers usually reply within 1 day
-
Console writes URL credentials to the local log, though AGENTS.md says secrets never reach diskOpen
Difficulty 1/5 Under an hour Newbie friendliness 86/100
sandbox-quantum/switch#2163 ·
Maintainers usually reply within 1 day