Message411927
| Author | eric.smith |
|---|---|
| Recipients | AlexWaygood, GBeauregard, JelleZijlstra, eric.smith, gvanrossum, kj, sobolevn |
| Date | 2022-01-27.21:17:04 |
| SpamBayes Score | -1.0 |
| Marked as misclassified | Yes |
| Message-id | <1643318224.69.0.0942996363246.issue46553@roundup.psfhosted.org> |
| In-reply-to |
| Content | |
|---|---|
dataclasses is no doubt too lenient. But it's just trying to accept valid strings that look like ClassVar. Way back when this was initially implemented, we decided that calling get_type_hints would be too expensive for every dataclass, and would also neccesitate importing typing, which we didn't want to require. I think someone needs to do an analysis of how expensive it would be for dataclasses to import typing and to call get_type_hints. Perhaps we'd make a different decision today. |
|
| History | |||
|---|---|---|---|
| Date | User | Action | Args |
| 2022-01-27 21:17:04 | eric.smith | set | recipients: + eric.smith, gvanrossum, JelleZijlstra, sobolevn, kj, AlexWaygood, GBeauregard |
| 2022-01-27 21:17:04 | eric.smith | set | messageid: <1643318224.69.0.0942996363246.issue46553@roundup.psfhosted.org> |
| 2022-01-27 21:17:04 | eric.smith | link | issue46553 messages |
| 2022-01-27 21:17:04 | eric.smith | create | |