Closed count0 closed 1 month ago
We need help from the @conda-forge/bot team.
On the bot status page for the numpy-2 migration, graph-tool
is listed under the Not solvable
category. It is showing errors like this:
linux_64_microarch_level1numpy2.0python3.12.____cpython: Cannot solve the request because of: _x86_64-microarch-level >=1 cannot be installed because there are no viable options:
└─ _x86_64-microarch-level 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 2 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 3 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 | 4 would require
└─ __archspec ==1 core2, for which no candidates were found.
So it seems that somehow our enhancements for microarch-optimized builds (#140) are complicating things for the bot. The bot was successfully triggered on May 29, before we merged our changes in #140, but was not able to run again after that.
It might be possible to just inspect the diff for #142 and manually apply those changes in a new PR, but it won't be practical to do that for all migrations in the future.
Can you make an issue here in this repo https://github.com/regro/conda-forge-feedstock-check-solvable?
I suspect we need to add __archspec
virtual packages to the solver testing logic.
Closed via PR #152
Comment:
There was a bot a while back trying to automatically port the recipe to numpy 2 (#142), but upstream was not ready. It is ready now (since version 2.71) but it seems the bot has either given up or has been discontinued.
It's not clear if still something we need to do... If so, we should try.
@stuarteberg