Matti Picus
Matti Picus
> This is definitely a bug, both conda and mamba solvers should respect (and prioritize) the pythons configuration option Does this work now?
I guess we could do this. I somewhat prefer to encourage people to use global `.gitignore` files for stuff on their systems: ide projects, editor backups and the such, and...
The impacted functions are `np.std`, `np.round`, and the getters for `np.scalar` for `object` dtype. Running `$ spin bench -t time_std --compare` I see no real change in the benchmarks.
The original request did note > wherever the lookup may be in a performance relevant code branch I don't think that is the case here.
Are there other parts of the asv web interface we could move out to a separate `asv[analyze]` package?
Did you mean to mark this for 0.6.3 or 1.6.2?
Should we put this in for the next release?
Sounds good.
This now has a merge conflict (no `u""` needed). I think it is fine to put this in, the change does not reduce the copyright of any of the original...
Thanks @HaoZeke