Message281926
Regarding the problem with the default value, can we use “unspecified”, as documented at <https://docs.python.org/dev/howto/clinic.html#writing-a-custom-converter>, or is that an old relic that no longer exists (hinted at the end of Issue 20293)?
Failing that, I prefer Raymond’s earlier suggestion: say that any support for −1 or None are implementation details, and not part of the API. The same statement could be added to the native Python doc strings as well.
BTW I am not opposed to the proposal for proper hi=None support either, although it does complicate the change. IMO allowing hi=None could be done as a second step. |
|
Date |
User |
Action |
Args |
2016-11-29 01:01:44 | martin.panter | set | recipients:
+ martin.panter, rhettinger, vstinner, larry, serhiy.storchaka, mdk |
2016-11-29 01:01:44 | martin.panter | set | messageid: <1480381304.71.0.255281457993.issue28754@psf.upfronthosting.co.za> |
2016-11-29 01:01:44 | martin.panter | link | issue28754 messages |
2016-11-29 01:01:44 | martin.panter | create | |
|