Bump to (OCaml4.14 + Core.v0.16) / (OCaml5.2 + Core.v0.17) - #912
Conversation
06367b7 to
99aff38
Compare
99aff38 to
0976bb4
Compare
850a6c3 to
3cac602
Compare
| end | ||
|
|
||
| module Make | ||
| (Constant : Field_intf.S) (F : sig |
|
FWIW: OCaml 5.2.x is already 2 years old, see https://ocaml.org/releases. OCaml 5.5.0 brings a set of interesting features, and there have been other improvements since 5.2.x. |
|
We're bumping up step by step. But I'd love to bump to 5.5.0 if there's no obvious blockers as well -- it just need to be investigated. Thanks for the recommendation @dannywillems :) |
|
@glyh question if existing tests are enough to cover this bump? |
|
@dkijania Good question — I went through the full diff to answer it. I believe the existing tests are sufficient for this bump, for two reasons: 1. The changes are mechanical, not semantic. The ~430 changed lines fall entirely into these buckets:
There are no algorithmic or data-structure changes, so what needs coverage is "does everything still compile and behave the same", not new logic. 2. What the suite exercises is the right regression surface for this kind of change. Two caveats, one of which is now addressed:
|
Pin core explicitly in each matrix leg — Core v0.17 also solves on OCaml 4.14, so without the pin both legs would test the same series. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This is part of the bigger plan bumping Mina to OCaml 5.2.1 with Corev0.17.1