[Draft] use algebraic number for reals #153
Conversation
|
Looking at the CI, and I'm wondering: is |
|
I have not made the effort of making it work there, but I have also not actively forbid it so I don't yet understand. |
|
Are the package updated in the windows CI? Perhaps this should be added: |
|
I think it is just because the original windows repo stop updating last year: https://ocaml.org/docs/ocaml-on-windows#opam-repository-mingw . I haven't been able to test adding the current one in my virtual machine. (ocaml-ci can also be a good CI since it is the one used by opam) (EDIT: sorry for the repetition, github slowly updated the page 🤦 ) |
|
yeah, it's not the first time the windows CI has had problems because the repo used by |
Gbury
force-pushed
the
algebraic_number
branch
2 times, most recently
from
be89444
to
5331850
Compare
There was a problem hiding this comment.
I pushed a few changes, mainly some refactoring and documentation, as well as proper errors instead of the invalid_arg. The code is good for merging, but we need to figure out how to make CI runs succeed; the installation of flint/calcium takes way too long: 6h+ on the windows CI which end up being forcefully shut down before completion. Ideally, the flint/calcium/arb packages could try and make use of depexts to rely on systems binary packages for the big C (or C++ ?) libraries.
|
ocaml/opam-repository#24509 might solve the CI problems. Will try again once it is merged. |
Tentative introduction of root-of-with-order Model support algebraic number root-of-with-order and root-of-with-enclosure Error path should be improved [Model] Set version of calcium needed Accept root-of-with-ordering [Algebraic] accept also root-of-with-interval [Algebraic] accept "-3" coefficient for experiment
Gbury
force-pushed
the
algebraic_number
branch
from
51a7d7d
to
e056909
Compare
ocaml-flint version 0.3 is needed