| 1 |
On 5/7/19 1:01 PM, Zac Medico wrote: |
| 2 |
> On 5/7/19 7:55 AM, Alec Warner wrote: |
| 3 |
> |
| 4 |
>> Also curious why we are not implementing enter and exit so we can avoid |
| 5 |
>> unbalanced pairs by using context managers. |
| 6 |
>> |
| 7 |
>> e.g. in match(), we could likely write: |
| 8 |
>> |
| 9 |
>> with self.dbapi.lock(): |
| 10 |
>> # try to match |
| 11 |
>> update_pkgindex = self._populate_local() |
| 12 |
>> if update_pkgindex: |
| 13 |
>> self._pkgindex_write(update_pkgindex) |
| 14 |
>> |
| 15 |
>> If the block raises an exception, __exit__ is still called, so we can |
| 16 |
>> eliminate the finally: blocks and replace them with a careful __exit__ |
| 17 |
>> implementation? |
| 18 |
>> |
| 19 |
>> -A |
| 20 |
> |
| 21 |
> Yeah, the reference counted lock class could have a method that returns |
| 22 |
> a context manager which increments/decrements the lock count. The |
| 23 |
> context manager can be implemented with the contextlib.contextmanager |
| 24 |
> decorator like this: |
| 25 |
> |
| 26 |
> @contextlib.contextmanager |
| 27 |
> def contextmanager(self): |
| 28 |
> self.lock() |
| 29 |
> try: |
| 30 |
> yield self |
| 31 |
> finally: |
| 32 |
> self.unlock() |
| 33 |
> |
| 34 |
|
| 35 |
Since we really don't want asynchronous code to block waiting for a |
| 36 |
lock, we really should use a (python3.5+ only) asynchronous context manager: |
| 37 |
|
| 38 |
https://www.python.org/dev/peps/pep-0492/#asynchronous-context-managers-and-async-with |
| 39 |
|
| 40 |
Given that python2.7 is scheduled for retirement in 2020 |
| 41 |
(https://pythonclock.org/), maybe it's time to drop support for |
| 42 |
python3.4 and earlier. |
| 43 |
-- |
| 44 |
Thanks, |
| 45 |
Zac |
Archives