Why do we need "try-finally" when using @contextmanager decorator?

Question

I wonder why we need to use a try-finally when using a the @contextmanager decorator.

The provided example suggests:

from contextlib import contextmanager

@contextmanager
def managed_resource(*args, **kwds):
    resource = acquire_resource(*args, **kwds)
    try:
        yield resource
    finally:
        release_resource(resource)

It seems to me, however, that this will do the exact same thing:

@contextmanager
def managed_resource(*args, **kwds):
    resource = acquire_resource(*args, **kwds)
    yield resource
    release_resource(resource)

I'm sure I must be missing something. What am I missing?

score 2 · Accepted Answer · answered Nov 23 '22 at 08:56

2

Because a finally statement is guaranteed to run no matter what (except a power outage), before the code can terminate. So writing it like this guarantees that the resource is always released

answered Nov 23 '22 at 08:56

oskros

3,101
2
9
28

So even when there's no second `next()` call? – Kris Nov 23 '22 at 08:58
1

@Kris yes, if you don't save a reference to the generator - see here: https://stackoverflow.com/a/56063400/9490769 – oskros Nov 23 '22 at 09:01

score 2 · Answer 2 · answered Nov 23 '22 at 09:02

finally makes sure that the code under it is always executed even if there's an exception raised:

from contextlib import contextmanager

@contextmanager
def exception_handler():
    try:
        yield
    finally:
        print("cleaning up")

with exception_handler():
    result = 10 / 0

If there were no try-finally, the above example wouldn't cleanup itself afterwards.

Why do we need "try-finally" when using @contextmanager decorator?

2 Answers2