Albin Ahlbäck
Albin Ahlbäck
The CI I just started ran into this issue. See it [here](https://github.com/flintlib/flint/actions/runs/8026786828/job/21929863351?pr=1791). The workflow configuration can be seen [here](https://github.com/flintlib/flint/blob/0c5ed5184367c3beedc27c86eb6bec1742c874b4/.github/workflows/CI.yml#L369).
Oh, I didn't think of that, thanks! The README doesn't really state which versions it picks for `toolset` and `vsversion`. Perhaps add that to the README?
I totally agree.
To be clear, this only happens with `--enable-assert`?
I could reproduce, and I think it's a bug/feature in Valgrind. Reading [Valgrind's manual](https://valgrind.org/docs/manual/manual-core.html): > Rounding: Valgrind does observe the 4 IEEE-mandated rounding modes (to nearest, to +infinity, to -infinity,...
I'm closing this. Please reopen if I misinterpreted it.
> * Right now additions and subtractions are inlined up to n = 8 limbs. For larger n inlining additions everywhere might be overkill, but we should investigate whether using...
> Would the approach proposed in PR [#2339](https://github.com/flintlib/flint/pull/2339) be fine? Yes!!
I think Fredrik is the person for this, I don't have a lot of insight in Calcium.