Message366999
The patch assumed using the magic method attribute as the way to evaluate an object in a context which I got to know is wrong since evaluations in context like boolean are not only dependent on one magic method but has a precedence over several others. In the event of 3.9 alpha 6 being the next candidate I guess the change can be reverted to be revisited later about the desired behaviour with backwards compatibility. The docs could perhaps be clarified that calling magic method on a wrap gives default set of values.
Thoughts on reversion or other possible approaches? |
|
Date |
User |
Action |
Args |
2020-04-22 12:54:30 | xtreak | set | recipients:
+ xtreak, rbcollins, cjw296, r.david.murray, michael.foord, serhiy.storchaka, pconnell, Darragh Bailey, lisroach, anthonypjshaw, mariocj89, aviso, Amir |
2020-04-22 12:54:30 | xtreak | set | messageid: <1587560070.52.0.353841996171.issue39966@roundup.psfhosted.org> |
2020-04-22 12:54:30 | xtreak | link | issue39966 messages |
2020-04-22 12:54:29 | xtreak | create | |
|