Results 1654 comments of Gabriel Scherer

I rebased the present PR on top of #13338, which fixes a code-quality regression introduced in #13076.

I implement and document a warning as discussed in #13341. That PR is on top of #13153, which removes the `check_partial` machinery (an older, incomplete heuristic to degrade partiality information,...

I just rebased this PR on top of the (new) second half of #13338, and also there was a small conflict with #13342 (a nice-to-have conflict, the new version is...

As far as I can tell, we don't have an actual reproduction example for the original issue: > if the [@@unboxed] attribute is applied to a type for which unboxing...

One reason I like this heuristic is that it helps me reason about "structural matches" that do not contain splits -- all possible cases are explored in full. For example:...

Rebasing this on top of #13338 (transitively) made the implementation quite a bit less invasive, because I can add the temporality as an extra field to the `partiality` record instead...

Thanks! The present PR sits on top of #13341, so I should wait for that PR to be merged before merging the present one.

This PR can be merged now that #13341 has been approved and merged. I am planning to do this after/if the CI passes.

This PR is sitting in a suspended state due to a lack of reviewers, despite demonstrating an improvement over the current trunk on realistic use-cases. Any volunteers to review the...