Message266612
I agree. This assert only indirectly verifies that something bad doesn't happen.
The bad thing that might happen is an in-place negating of an element of small_ints, so the most direct assert should be 'assert(Py_REFCNT(z) == 1);'.
This is exactly what Victor did in long_lshift back in revision 84698...
What do you think? |
|
Date |
User |
Action |
Args |
2016-05-29 18:48:29 | Oren Milman | set | recipients:
+ Oren Milman, mark.dickinson, vstinner, serhiy.storchaka, yselivanov |
2016-05-29 18:48:29 | Oren Milman | set | messageid: <1464547709.37.0.45192072923.issue27073@psf.upfronthosting.co.za> |
2016-05-29 18:48:29 | Oren Milman | link | issue27073 messages |
2016-05-29 18:48:29 | Oren Milman | create | |
|