Fatal python error in com_make_closure()

[glchapman]

BPO	233096
Nosy	@mwhudson, @freddrake

^{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 = None
closed_at = <Date 2001-02-19.23:38:29.000>
created_at = <Date 2001-02-19.17:03:35.000>
labels = ['interpreter-core']
title = 'Fatal python error in com_make_closure()'
updated_at = <Date 2001-02-19.23:38:29.000>
user = 'https://bugs.python.org/glchapman'

bugs.python.org fields:

activity = <Date 2001-02-19.23:38:29.000>
actor = 'jhylton'
assignee = 'jhylton'
closed = True
closed_date = None
closer = None
components = ['Interpreter Core']
creation = <Date 2001-02-19.17:03:35.000>
creator = 'glchapman'
dependencies = []
files = []
hgrepos = []
issue_num = 233096
keywords = []
message_count = 6.0
messages = ['3492', '3493', '3494', '3495', '3496', '3497']
nosy_count = 4.0
nosy_names = ['mwh', 'jhylton', 'fdrake', 'glchapman']
pr_nums = []
priority = 'high'
resolution = 'fixed'
stage = None
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue233096'
versions = []

[glchapman]

This construct causes python (2.1 alpha 2) to crash:

>>> def test(n):
...     return lambda fact: fact(fact, long(n))
...
lookup 'n' in test 1 -1
Fatal Python error: com_make_closure()

To try to help narrow it down, the following works with no problem:

>>> def test(n):
...     return lambda fact: fact(fact, n)

So it appears that adding the call to the long function is breaking something.

[glchapman]

I probably should have added that I'm using the Win32 compiled version downloaded from SourceForge (operating system is Windows 2000).

[mwhudson]

Nah, it crashes on linux too.

It doesn't crash if you change almost anything about the test-case, so it's probably some bizarre kind of dictionary traversal bug again.

[mwhudson]

OTOH, I've just rebuilt and it seems to have gone away.

A bug of this kind was fixed in revision 2.163 of compile.c (on 2001/02/12)- can you try after that?

[freddrake]

Assigned to Mr. Compiler.

[jhylton]

This appears to work with the current CVS repository. One of the
bug fixes this morning must have caught it.