Message72303
Le lundi 01 septembre 2008 à 21:22 +0000, Nick Coghlan a écrit :
> As Fredrik noted in the python-dev thread, the threading versions of
> these are already explicitly documented as being factory functions
> rather than classes (and as a reference to _thread.allocate_lock,
> threading.Lock has good reason to be a factory function rather than a
> class), so it may be appropriate to do the same thing for multiprocessing.
"Foolish consistency, etc.". :-)
I think we should fix this before the releases rather than live with a
stupid indirection that doesn't seem to serve any useful purposes.
Unless the fix is too complicated of course. |
|
Date |
User |
Action |
Args |
2008-09-01 22:11:26 | pitrou | set | recipients:
+ pitrou, ncoghlan, benjamin.peterson, jnoller |
2008-09-01 22:11:25 | pitrou | link | issue3589 messages |
2008-09-01 22:11:25 | pitrou | create | |
|