Closed jdblischak closed 1 month ago
https://github.com/conda-forge/tiledb-feedstock/pull/285#issuecomment-2117978773 was merged last night. This morning I restarted the failed linux-64 build on commit https://github.com/TileDB-Inc/tiledbsoma-feedstock/commit/f537d2dbc8d7e195dac350578fd8b0ca031cd67a, and it passed.
And strangely an entire new job was also started even though I clearly clicked in the GitHub checks UI to only re-run failed jobs. This same thing happened recently to me where I restarted a failed job on a merge commit in another feedstock. This is really frustrating since it is a waste of CI resources. And I don't want to cancel it because it replaces the CI links attached to the commit, so it would like it failed.
The linux-64 job failed on the merge commit (https://github.com/TileDB-Inc/tiledbsoma-feedstock/commit/f537d2dbc8d7e195dac350578fd8b0ca031cd67a) from my recent PR #154 to update for Python 3.12.
Thus it clearly needs tiledb-feedstock to be migrated to libgoogle-cloud 2.24. The PR is open, https://github.com/conda-forge/tiledb-feedstock/pull/285#issuecomment-2117978773, but it is blocked by an unexpected win-64 build failure that we first discovered last week in our nightly feedstock builds (https://github.com/TileDB-Inc/conda-forge-nightly-controller/issues/94)