Gabriel Scherer

> Third (and in my opinion most confusing but also likely the hardest to warn about), it's unclear that the recursive dependence on the compare functions is inducing this error...

Right, it would make sense to point to the location of the submodule, rather than the location of its first non-safe value. Notice how the `loc` parameter of `init_shape_mod` (which...

This is an issue about how to improve the error message for recursive modules, when the issue occurs in a submodule. Several possible changes and improvements have been discussed, some...

This has been solved by merging #13608. Thanks @shivam-909!

I observed a new failure in https://github.com/ocaml/ocaml/pull/13341, which is a clean test failure rather than a segfault: ```diff -Observation MP+PA Never +Observation MP+PA Sometimes +Invalid behaviour on test MP+PA ```

The `mapM` example works with modular explicits, but not with first-class modules only. (It can also be expressed using a functor). I can clarify this.

I see some regularity in functions called `fooM`, where `foo` takes a higher-order function `(bar... -> baz)` and `fooM` a higher-order function `(bar... -> baz M.t)` -- for example we...

I modified the new introduction/tutorial part of this PR to address a non-trivial comment by @samsa1, and also to explain a potentially surprising example of code that fails to type...

Thanks @goldfirere and @kayceesrk, I took your comments into account and rebased. (My understanding of the present PR is that it is de-facto approved, in that it is waiting for...

A rendered version of the manual with this PR is available at . (The reason this is hard to build is that the manual PR must be based on the...