Gabriel Scherer

I am slightly nervous about the correctness of this patch. Here is my current understanding: - The strings you are freeing were allocated with `caml_stat_alloc`, so they are not actually...

I looked at when `caml_remove_debug_info` is called. The call I found is (through `Meta.release_bytecode`) in toplevel/byte/Topeval.load_lambda, where it is conditioned on the `can_free` boolean returned by `Bytegen.compile_phrase`, which seems to...

The proposal suggests to allow `include functor FT` in signatures, where `FT` is the type/signature of a functor, but it also suggests that naming functor types is uncommon (this is...

I find that @yallop's suggestion of naming the "current module so far" has merit, at least when equipped with the decent syntax `module` proposed by @goldfirere. In modules, one would...

@fpottier: `include Foo(struct include this let x = 0 end)`: wait a minute, clearly `this` is the empty module in this context, right? You probably meant `this/2` ;-)

I also thought of `struct (Self) ... end` as for objects, and I don't see a big issue with the fact that the name refers to the fraction of the...

> Interestingly, the latest local assertion failure error was as follows: 🤔 [..] `Illegal instruction` This is probably a consequence of https://github.com/ocaml/ocaml/pull/13344 , where failed asserts now cause a `trap`...

In `s->running`, `s` is of type `struct interruptor *`, which is part of the `struct dom_internal` state, and shared between a domain and its backup thread. The interruptor of a...

@zazedd the `check-typo` script complains that one line is now over 80 columns: ``` Checking ee7e977d4979bdd06cda6ef280817b616e639453: stdlib/printf.mli ./stdlib/printf.mli:220.81: [long-line] line is over 80 columns ``` While we are at it:...

Your suggestion would break code that uses `open Format` and then lists.