Delete always skipped test, fix trio test, fix framelocal has not .clear()

[Carreau]

There are few chances we will ever use them again.

Also a small refactor and unskip of trio test using trio nursery.

[Carreau]

Well that's strange, trio test passes locally...

well on CI, anyio decides to switch in the middle of the call stack from trio to asyncio:

 | tg.start_soon(self.kernel.start)
| File ".../site-packages/anyio/_backends/_trio.py", line 191, in __aexit__
| return await self._nursery_manager.__aexit__(exc_type, exc_val, exc_tb) # type: ignore[return-value]
| File ".../site-packages/trio/_core/_run.py", line 1043, in __aexit__
| raise combined_error_from_nursery
| exceptiongroup.ExceptionGroup: Exceptions from Trio nursery (1 sub-exception)
+-+---------------- 1 ----------------
| Traceback (most recent call last):
| File ".../ipykernel/ipykernel/kernelapp.py", line 753, in _wait_to_enter_eventloop
| await self.kernel._eventloop_set.wait()
| File ".../site-packages/anyio/_core/_synchronization.py", line 139, in wait
| await self._event.wait()
| File ".../site-packages/anyio/_backends/_asyncio.py", line 1764, in wait
| await AsyncIOBackend.checkpoint()
| File ".../site-packages/anyio/_backends/_asyncio.py", 

[Carreau]

Well something else is wrong on CI, some pytest items are failing, but it's shown as green on 3.10: https://github.com/ipython/ipykernel/actions/runs/13292836924/job/37117594041?pr=1322

[davidbrochart]

Well something else is wrong on CI, some pytest items are failing, but it's shown as green on 3.10: https://github.com/ipython/ipykernel/actions/runs/13292836924/job/37117594041?pr=1322

It's because we rerun last failed tests here.

[Carreau]

fix framelocal has not .clear()

It would be really great to get that merged, as this fixes failures of other PRs

[Carreau]

(docs breckage is due to new sphinx release and unrelated to us)

[krassowski]

This looks good to me. The sphinx issue is tracked in tox-dev/sphinx-autodoc-typehints#523

[Carreau]

Thanks for the reviews. Merging.