This code seems to get a RESUME opcode where no function call is happening:

    a = 1
    fn = lambda 2
    b = 3

Here is the disassembly. Offset 6 has a RESUME opcode for line 2:

Python 3.11.0a3+ (heads/main:0fc58c1e05, Jan  8 2022, 19:45:58) [Clang 12.0.0 (clang-1200.0.32.29)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import dis
>>> dis.dis("""\
... a = 1
... fn = lambda: 2
... b = 3
... """)
              0 RESUME                   0

1 2 LOAD_CONST 0 (1)
4 STORE_NAME 0 (a)

2 6 RESUME 0
8 LOAD_CONST 1 (<code object <lambda> at 0x10772c2d0, file "<dis>", line 2>)
10 MAKE_FUNCTION 0
12 STORE_NAME 1 (fn)

3 14 LOAD_CONST 2 (3)
16 STORE_NAME 2 (b)
18 LOAD_CONST 3 (None)
20 RETURN_VALUE

Disassembly of <code object <lambda> at 0x10772c2d0, file "<dis>", line 2>:
2 0 RESUME 0
2 LOAD_CONST 1 (2)
4 RETURN_VALUE

This causes an extra "call" event when tracing:

---< bug233.py >--------------------------------------------------------

import linecache, os.path, sys

def trace(frame, event, arg):
    if (event != "line") and ("bug233" in frame.f_code.co_filename):
        lineno = frame.f_lineno
        first = frame.f_code.co_firstlineno
        source = linecache.getline(frame.f_code.co_filename, lineno) if lineno else "******"
        file = os.path.basename(frame.f_code.co_filename)
        print("{} {}@{!s:4} (first={}): {}".format(event[:4], file, lineno, first, source.rstrip()))
    return trace

print(sys.version)
sys.settrace(trace)

def f():
    a = 1
    fn = lambda: 2
    b = 3
f()

% python3.10 bug233.py
3.10.1 (main, Dec 14 2021, 08:30:13) [Clang 12.0.0 (clang-1200.0.32.29)]
call bug233.py@15 (first=15): def f():
retu bug233.py@18 (first=15): b = 3

% python3.11 bug233.py
3.11.0a3+ (heads/main:0fc58c1e05, Jan 8 2022, 19:45:58) [Clang 12.0.0 (clang-1200.0.32.29)]
call bug233.py@15 (first=15): def f():
call bug233.py@17 (first=15): fn = lambda: 2
retu bug233.py@18 (first=15): b = 3

Thanks, Ned.

Pablo, I've marked this as a release blocker so this bug doesn't get into 3.11a4.

Unfortunately the release of 3.11.0a4 is already underway and we are waiting for Steve's windows binaries, so this will need to wait for alpha 5

New changeset ec0c392 by Mark Shannon in branch 'main':
bpo-46314: Remove extra RESUME when compiling a lamdba. (GH-30513)
ec0c392

Ok, I'll add a news item in that case.

This fixes the problems I was seeing, thanks.

This fixes the problems I was seeing, thanks.

Good; removing release blocker status. Keeping this open until Mark has added a NEWS item.

The news item was added in PR 30515