13:23 <discocaml> <deepspacejohn> I noticed there are no warnings for unused object instance variables or methods. Is that not possible to implement due to their structural nature?

13:55 <discocaml> <sim642> There seems to be an unused-field warning

13:56 <discocaml> <sim642> But I guess it might only be triggered when explicitly setting an object type to your object or whatnot

13:57 <discocaml> <sim642> If it's exposed, such that it could be accessed from another module, then the compiler cannot say it's unused because it won't know that

16:26 <discocaml> <asadun5771> does anyone here have any experience with hardcaml?

16:27 <discocaml> <asadun5771> namely to program FPGA (in this case I have Arty A7-100T)

18:05 <discocaml> <deepspacejohn> @sim642 that warning is for unused record fields, not object instance variables, right?

18:09 <discocaml> <deepspacejohn> https://ocaml.org/play#code=W0BAQHdhcm5pbmcgIitBIl0KbW9kdWxlIE06IHNpZyBlbmQgPSBzdHJ1Y3QKICBjbGFzcyB4ID0KICAgIG9iamVjdAogICAgICB2YWwgdXNlZCA9ICJ1c2VkIgogICAgICB2YWwgdW51c2VkID0gInVudXNlZCIKICAgICAgbWV0aG9kIHVzZWQgPSBwcmludF9lbmRsaW5lIHVzZWQKICAgICAgbWV0aG9kIHVudXNlZCA9IHByaW50X2VuZGxpbmUgdW51c2VkCiAgICBlbmQKICBsZXQgKCkgPSAobmV3IHgpI3VzZWQKZW5k

18:09 <discocaml> <deepspacejohn> There aren't any warnings for the unused variables or methods

18:50 <discocaml> <sim642> Hmm, I must've misread then

19:08 <discocaml> <deepspacejohn> I'm sure this is not a hot take to anyone here, but every once in a while when I experiment with using OCaml objects I always feel like I'm just doing a huge downgrade. I need more type annotations, there's no LSP go-to-definition for methods, no warnings for unused methods, bundle size gets bigger (for JSOO), and all kinds of other rough edges.

19:08 <humasect> try Modules ?

19:08 <humasect> not sure about LSP for ocaml, but merlin is *awesome*

19:10 <discocaml> <sim642> It's not surprising, objects are quite rarely used. Although it's a feedback loop: they're also used less because they're less optimized

19:10 <discocaml> <sim642> But I'm surprised about needing more type annotations for objects

19:10 <discocaml> <sim642> I thought annotations weren't necessary there at all

19:11 <discocaml> <deepspacejohn> I assumed the lack of go-to-definition for objects was more due to their fundamentally dynamic nature. In many cases there's no "definition" for a method.

19:11 <discocaml> <sim642> That probably doesn't help the issue indeed

19:12 <humasect> because objects are dynamic

19:12 <humasect> modules are static

19:12 <discocaml> <deepspacejohn> Any time a class does anything polymorphic, including a polymorphic method, you usually need type annotations.

19:12 <discocaml> <sim642> Right, polymorphism is explicit there

19:14 <discocaml> <deepspacejohn> anyway, I'm just being grumpy. I have no special need to use objects except that once a year or so I get curious and decide to try them out.

19:15 <discocaml> <deepspacejohn> maybe I'll find a use case for them this time? (typically not)

19:16 <humasect> very_pretty#syntax ~stuff:

19:18 <discocaml> <lecondorduplateau> ocaml needs static objects

19:18 <discocaml> <lecondorduplateau> instead of poor unlabeled tuples

19:36 <discocaml> <sim642> Well, there will be labeled tuples now

19:53 <companion_cube> I'd also like static variables :p