All commit authors signed the Contributor License Agreement.

Most changes to Python require a NEWS entry. Add one using the blurb_it web app or the blurb command-line tool.

If this change has little impact on Python users, wait for a maintainer to apply the skip news label instead.

This needs a NEWS entry, per the bot, as this is a user-facing bug. (Also, hi!)

Is this specific to a new REPL, or is it also a problem with the old REPL? If so, there's no need to backport it to 3.12.

Is this specific to a new REPL, or is it also a problem with the old REPL? If so, there's no need to backport it to 3.12.

Oh, I didn't check. Does the asyncio REPL use the new REPL?

This needs a NEWS entry, per the bot, as this is a user-facing bug. (Also, hi!)

Nevermind, I was being dumb, there's already a news entry. Ignore that!

Is this specific to a new REPL, or is it also a problem with the old REPL? If so, there's no need to backport it to 3.12.

Oh, I didn't check. Does the asyncio REPL use the new REPL?

Yes, it is.

Another fun fact I've also observed is that user_inputX = ... variables used in tests are not in fact atomic user inputs, in contrast to the analogous string variables in the standard REPL's tests, and every line break in them causes an immediate prompt EOF (try to indent the body of the set_var() function and see what happens – ... <rest of input> behavior doesn't really seem to get triggered, maybe it's something PTY-related?). I'm not sure how to handle that; it seems to be a part of a separate bug triggered by -i and -m asyncio used together in the circumstances of spawn_repl().

every line break in them causes an immediate prompt EOF (try to indent the body of the set_var() function and see what happens – ... <rest of input> behavior doesn't really seem to get triggered, maybe it's something PTY-related?)

$ ./python -m Lib.test.test_repl
F...........
======================================================================
FAIL: test_toplevel_contextvars_async (__main__.TestAsyncioREPLContextVars.test_toplevel_contextvars_async)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/bswck/Projects/cpython/Lib/test/test_repl.py", line 328, in test_toplevel_contextvars_async
    self.assertIn(expected, output, expected)
    ~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: 'toplevel contextvar test: ok' not found in 'asyncio REPL 3.14.0a0 (heads/main-dirty:8447c933da3, Sep 25 2024, 15:45:56) [GCC 11.4.0] on linux\nUse "await" directly instead of "asyncio.run()".\nType "help", "copyright", "credits" or "license" for more information.\n>>> import asyncio\n>>> >>> >>>   File "<stdin>", line 1\n    async def set_var():\n                        ^\nIndentationError: expected an indented block after function definition on line 1\n>>>   File "<stdin>", line 1\n    var.set(\'ok\')\nIndentationError: unexpected indent\n>>> Traceback (most recent call last):\n  File "/home/bswck/Projects/cpython/Lib/concurrent/futures/_base.py", line 448, in result\n    return self.__get_result()\n           ~~~~~~~~~~~~~~~~~^^\n  File "/home/bswck/Projects/cpython/Lib/concurrent/futures/_base.py", line 393, in __get_result\n    raise self._exception\n  File "<stdin>", line 1, in <module>\nNameError: name \'set_var\' is not defined\n>>> toplevel contextvar test: failed\n>>> \nexiting asyncio REPL...\nTraceback (most recent call last):\n  File "<frozen runpy>", line 198, in _run_module_as_main\n  File "<frozen runpy>", line 88, in _run_code\n  File "/home/bswck/Projects/cpython/Lib/asyncio/__main__.py", line 204, in <module>\n    sys.exit(return_code)\n    ~~~~~~~~^^^^^^^^^^^^^\nSystemExit: 0\n>>> \n' : toplevel contextvar test: ok

----------------------------------------------------------------------
Ran 12 tests in 0.353s

FAILED (failures=1)

I believe copy_context() makes more sense than manually constructing the context with Context() because the pythonstartup.py script can set some context variables, but I'm open to discussion about that.

I haven't followed the new repl change so can't review but copy_context doesn't allows back propagation across threads, not sure if that matters here.

As far as I can see, this isn't specific to the new REPL.

LGTM, thanks!

Thanks @bswck for the PR, and @asvetlov for merging it 🌮🎉.. I'm working now to backport this PR to: 3.12, 3.13.
🐍🍒⛏🤖

GH-124848 is a backport of this pull request to the 3.13 branch.

GH-124849 is a backport of this pull request to the 3.12 branch.

@asvetlov Thanks! This also closes out gh-87585 and gh-24773.