Skip to content

Commit

Permalink
Unpin numba<0.58 (rapidsai#15031)
Browse files Browse the repository at this point in the history
I think unpinning `numba` in the conda recipe was just missed in rapidsai#14616.

I discovered this issue [trying to build the `24.02` release](https://github.com/rapidsai/cudf/actions/runs/7878153691/job/21496377912#step:7:1674).

PRs & nightly builds are working because the `rapidsai-nightly` channel has an older version of `pynvjitlink` that supported `numba>=0.57` whereas the `rapidsai` channel only has the latest version which pins to `numba>=0.58`.

Authors:
   - Ray Douglass (https://github.com/raydouglass)

Approvers:
   - Vyas Ramasubramani (https://github.com/vyasr)
   - Bradley Dice (https://github.com/bdice)
  • Loading branch information
raydouglass authored Feb 13, 2024
1 parent 82f6a53 commit ac438c4
Showing 1 changed file with 1 addition and 2 deletions.
3 changes: 1 addition & 2 deletions conda/recipes/cudf/meta.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -78,8 +78,7 @@ requirements:
- typing_extensions >=4.0.0
- pandas >=1.3,<1.6.0dev0
- cupy >=12.0.0
# TODO: Pin to numba<0.58 until #14160 is resolved
- numba >=0.57,<0.58
- numba >=0.57
# TODO: Pin to numpy<1.25 until cudf requires pandas 2
- numpy >=1.21,<1.25
- {{ pin_compatible('pyarrow', max_pin='x') }}
Expand Down

0 comments on commit ac438c4

Please sign in to comment.