Threading misbehavior with lambdas

[fijal]

BPO	1339045
Nosy	@josiahcarlson

^{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 2005-10-27.19:32:48.000>
created_at = <Date 2005-10-27.00:21:38.000>
labels = ['interpreter-core']
title = 'Threading misbehavior with lambdas'
updated_at = <Date 2005-10-27.19:32:48.000>
user = 'https://bugs.python.org/fijal'

bugs.python.org fields:

activity = <Date 2005-10-27.19:32:48.000>
actor = 'josiahcarlson'
assignee = 'none'
closed = True
closed_date = None
closer = None
components = ['Interpreter Core']
creation = <Date 2005-10-27.00:21:38.000>
creator = 'fijal'
dependencies = []
files = []
hgrepos = []
issue_num = 1339045
keywords = []
message_count = 2.0
messages = ['26746', '26747']
nosy_count = 2.0
nosy_names = ['josiahcarlson', 'fijal']
pr_nums = []
priority = 'normal'
resolution = None
stage = None
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue1339045'
versions = ['Python 2.4']

[fijal]

suppose i write:
def f(x):
print x()

for i in range(3):
  f ( lambda : i )

I got 0,1,2

But when I write

for i in range(3):
  thread . start_new_thread ( f , ( lambda : i ) )

I got 2,2,2

Probably I don't get well design principles, but isn't
it against thread consistency? (as long as threads does
not interact with each other, interlace doesn't matter).

[josiahcarlson]

Logged In: YES
user_id=341410

This is a bug in your understanding of lambdas, not in how
threads work. More specifically, lambdas do late binding.
By the time the threads have actually started executing and
call the lambda, the name 'i' is bound to the value 2.

If you need early binding, then you should bind early:

for i in xrange(3):
    thread.start_new_thread(f, (lambda i=i:i))