3 Commits

Author	SHA1	Message	Date
Peter Bell	c936396af2	Always convert truthy booleans to 1 ... Ref #54789 A `bool` has only two valid values, 1 or 0. Any in-memory value outside of those leads to undefined behavior. So, instead of `reinterpret_cast`-ing to `bool*` I introduce `c10::load<scalar_t>` which will read as `unsigned char` and convert to a valid `bool`. This gets >90% of operators working, but the remaining operators where skips and xfails have been added will require individual attention. Pull Request resolved: https://github.com/pytorch/pytorch/pull/77122 Approved by: https://github.com/mruberry	2022-06-07 16:00:30 +00:00
PyTorch MergeBot	78824a7d54	Revert "Always convert truthy booleans to 1" ... This reverts commit 3c3c6cd982. Reverted https://github.com/pytorch/pytorch/pull/77122 on behalf of https://github.com/mruberry due to broke some jobs, like https://github.com/pytorch/pytorch/runs/6706333043?check_suite_focus=true	2022-06-02 13:45:54 +00:00
Peter Bell	3c3c6cd982	Always convert truthy booleans to 1 ... Ref #54789 A `bool` has only two valid values, 1 or 0. Any in-memory value outside of those leads to undefined behavior. So, instead of `reinterpret_cast`-ing to `bool*` I introduce `c10::load<scalar_t>` which will read as `unsigned char` and convert to a valid `bool`. This gets >90% of operators working, but the remaining operators where skips and xfails have been added will require individual attention. Pull Request resolved: https://github.com/pytorch/pytorch/pull/77122 Approved by: https://github.com/mruberry	2022-06-02 04:18:34 +00:00