gen_send_ex: Assertion `f->f_back != ((void *)0)' failed.

[nascheme]

BPO	1257960
Nosy	@nascheme, @rhettinger, @pjeby

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = 'https://github.com/pjeby'
closed_at = <Date 2005-08-13.03:29:21.000>
created_at = <Date 2005-08-12.19:20:46.000>
labels = ['interpreter-core', 'release-blocker']
title = "gen_send_ex: Assertion `f->f_back != ((void *)0)' failed."
updated_at = <Date 2005-08-13.03:29:21.000>
user = 'https://github.com/nascheme'

bugs.python.org fields:

activity = <Date 2005-08-13.03:29:21.000>
actor = 'pje'
assignee = 'pje'
closed = True
closed_date = None
closer = None
components = ['Interpreter Core']
creation = <Date 2005-08-12.19:20:46.000>
creator = 'nascheme'
dependencies = []
files = []
hgrepos = []
issue_num = 1257960
keywords = []
message_count = 3.0
messages = ['26052', '26053', '26054']
nosy_count = 3.0
nosy_names = ['nascheme', 'rhettinger', 'pje']
pr_nums = []
priority = 'release blocker'
resolution = 'fixed'
stage = None
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue1257960'
versions = ['Python 2.5']

[nascheme]

Triggering is trival. Create a script that contains:

    def f():
        yield 1
    g = f()

Run with a debug version of Python:

python: ../Objects/genobject.c:85: gen_send_ex:
Assertion `f->f_back != ((void *)0)' failed.
Aborted (core dumped)

[rhettinger]

Logged In: YES
user_id=80475

I believe this is related to IDLE crash that I've been seeing.
The problem did not occur just before the checkin:
cvs up -D "2005-08-01 18:00"
But emerged immediately after:
cvs up -D "2005-08-01 21:00"

[pjeby]

Logged In: YES
user_id=56214

Sadly, this is not the cause of the IDLE problem, because
it's the assert that's wrong here. The problem that's
occurring is that f->f_back is NULL because the final
garbage collection at shutdown is occurring with a NULL
tstate->frame. Changing the assert to check that
f->f_back==tstate->frame makes the (meaningless) error go away.

Basically, the problem here is because this code used to be
the iternext routine, and it was never called by the GC.
Now, generators are closed when they are garbage collected,
so they can be executed during interpreter shutdown.

I've checked in a corrected assertion.