09:37 <adrien> I spotted "bash" and "unit testing" in the history; shell script's error management is too broken, there's no hope

10:52 <neiluj> hey! newbie question about objects: so in this code snippet: https://tpaste.us/Xy9e the syntax is not correct in the EllipticCurve module where I try to extend the meth module with a new method random

10:52 <neiluj> how to extend classes signatures here?

10:53 <neiluj> first it should be inherit t group

12:19 <discocaml> <xavierm02> Given several implementations of the same interface given by OCaml modules, it is (often) possible to define a dynamic implementation of that interface by wrapping the corresponding implementation with each value. When some binary operation of the dynamic implementation act differently depending on whether both operand use the same implementation or not, how should the equality between implementations be done?

12:20 <discocaml> <xavierm02> Here is a toy example where the interface is monoids and the dynamic implementation is the free product of all monoids : https://pastebin.com/1dEAa2ZP

12:23 <discocaml> <xavierm02> Since the polymorphic equality did not seem to work on first-class modules, I added a uid to each of them to be able to test for equality. And I would prefer avoiding the use of `Obj.magic`, but I'm not sure how to do that

12:24 <discocaml> <xavierm02> Given several implementations of the same interface given by OCaml modules, it is (often) possible to define a dynamic implementation of that interface by wrapping the corresponding implementation with each value. When some binary operation of the dynamic implementation act differently depending on whether both operand use the same implementation or not, how should the equality between implementations be tested?

12:36 <discocaml> <octachron> @xavierm02 , you can use an extensive variant to rediscover identity.

12:39 <discocaml> <octachron> neiluj: see the manual object chapter about object types (inheritance is for code reuse in implementation not for types).

12:44 <neiluj> thanks octachron!

19:40 <companion_cube> damn, I got schooled by nojb 😂

20:01 <discocaml> <anmonteiro> dang

20:02 <discocaml> <anmonteiro> what an interesting bit of history 😛

20:03 <companion_cube> yeah I didn't realize that C99 added several useful things

20:45 <discocaml> <Alistair> What does cmm and cmo actually stand for?

20:51 <discocaml> <anmonteiro> I _think_ `cmm` is C-minus-minus?

20:51 <discocaml> <anmonteiro> and cmo for "caml object"?