Gabriel Scherer

Some directories have non-parallel and parallel tests for the same feature (for example: weak-ephe-final/, gc-roots/). I find this a natural organization. We could have `parallel/` subdirectory in the test directory,...

My intuition is that this change is more principled/robust than #14304 (which is arguably a hack to avoid a known-bad interaction with the GC, rather than a guarantee that the...

Thanks for the review (and also @jhjourdan!). I will try to please the CI, add an extra comment I felt was missing on re-review (clarify that we intentionally reset the...

Thanks! This one is for @garrigue to review (it's a very simple code change, the question is whether it is correct).

Ah, the problem here is that we are making an assumption about the algebra of polymorphic variants, that empty types `[< ]` are not allowed. Otherwise it is incorrect that...

Yep, I've been procrastinating on getting our experimental fork ( mentioned/advertised in https://github.com/ocaml/RFCs/pull/14#issuecomment-920643103 ) in a state in which I can submit the feature as a PR. I wanted to...

I'm not sure; my understanding is that: 1. Most industrial members of the Consortium, who were mostly interested in providing some financial support for the OCaml ecosystem, migrated to become...

I rebased the PR to take the comments into account (thanks!). The paragraph reassuring downstream forks regarding the linking exception now reads as follows: > Finally: If you have your...

I took @hannesm's comments into account (thanks!) and rebased the PR. I am thinking of merging when the CI agrees.

Sorry about that. Would you have a repro case to test?