New comment by tornaria on void-packages repository

https://github.com/void-linux/void-packages/pull/34030#issuecomment-979968082

Comment:
> Sounds reasonable. On musl, the failures are not real (I think) (and similar for i686)
> 
> ```
> sage -t --random-seed=53597881869110074179920514523677232559 src/sage/parallel/map_reduce.py  # Timed out
> sage -t --random-seed=53597881869110074179920514523677232559 src/sage/schemes/affine/affine_point.py  # Killed due to kill signal
> ```

Why do you mean with "not real" ? Occasionally a timeout could be a false positive: when a test takes too long it could sometimes goes over the timeout (in that case it should has to be tagged as "# long test" so it's still a bug). All of these may be very real, just random. The failures on i686 are definitely real.
  
> Do you think python3.10 will make it into sage9.5?

I don't know, it's progressing slowly.

> Can you bring the `depends` down to what is really necessary,

What should I remove? We have to make sure all tests pass because a failed tests is a missing feature.

> and make sure it removes properly with xbps-remove? (last time I tried, it left behind `/usr/lib/sage..`)

Can you tell me which files are left behind?