03:26 <discocaml> <rgrinberg> there's a pure ocaml solver the "0install solver" I believe

04:35 <companion_cube> yes, but it's really not great (or wasn't, years ago)

04:55 <discocaml> <rgrinberg> what's wrong with it?

05:01 <companion_cube> It's a sat solver implemented with objects :). There are better ways. That said if it does the job...

05:01 <companion_cube> I don't think it'd be enough for opam which needs more than sat though

05:42 <discocaml> <rgrinberg> yeah, there are some features it doesn't support but at least it's fast apparently

10:41 <discocaml> <AradArbel10> I'm slightly concerned about copying data structures. my situation is some structured data, built mostly of nested `list`s and recursive ADTs, everything immutable. if I take one node of this tree-like thing and assign it into another node, would ocaml do a deep copy?

10:41 <discocaml> <AradArbel10> how can I check if this kind of sharing takes place? and if it doesn't, how can I make it happen (maybe some `ref` shenanigans, despite not needing mutability...)

10:42 <discocaml> <AradArbel10> I'm slightly concerned about copying data structures. my situation is some structured data, built mostly of nested `list`s and recursive ADTs, everything immutable. if I take one node of this tree-like thing and construct a new node containing it, would ocaml do a deep copy?

10:42 <discocaml> <AradArbel10> how can I check if this kind of sharing takes place? and if it doesn't, how can I make it happen (maybe some `ref` shenanigans, despite not needing mutability...)

10:51 <discocaml> <octachron> Ocaml never does deep copy by itself.

10:53 <discocaml> <octachron> A value is either an integer or an immutable pointer towards a GC-owned piece of memory.

11:00 <discocaml> <AradArbel10> ah nice, so I think I'm good

17:47 <discocaml> <ec> Is there a popular/standard-ish monadic transformers library that _isn't_ dependent on `core_kernel`? I see `mnd` … and `monadlib`, but it doesn't look very active, and also depends on `batteries` …

18:43 <discocaml> <nsmmrs> Containers has functions like Option.map/bind and their infix operator equivalents.

19:12 <discocaml> <darrenldl> finally finished restructuring a nottui mess into a component tree structure. definitely should have pinged the frontend dev friend earlier.

19:55 <discocaml> <mbacarella> suppose I have a suite of libraries that I want all contained under a top-level module called Swiss with the option to install any of the library suite later. like someone can opam install swiss-foo and swiss-bar and the result there will be they can access the code via Swiss.Foo and Swiss.Bar. is the way to do this through the dune selection system where if the library Swiss_foo exists it'll bind it to Swiss.Foo and if it doesn't exist it'

20:16 <discocaml> <mbacarella> I guess also list swiss-foo and swiss-bar as optional dependencies of swiss in opam so that it rebuilds swiss after you install swiss-foo?

20:23 <discocaml> <mbacarella> suppose I have a suite of libraries that I want all contained under a top-level module called Swiss with the option to install any of the library suite later. like someone can opam install swiss-foo and swiss-bar and the result there will be they can access the code via Swiss.Foo and Swiss.Bar. is the way to do this through the dune selection system where if the library Swiss_foo exists it'll bind it to Swiss.Foo and if it doesn't exist it'

20:56 <companion_cube> @darrenldl is the code still using nottui?

22:12 <discocaml> <masterbuilder> is there a way to make `assert` print the failing assertion?

22:13 <discocaml> <NULL> You mean having an exception which carries a string?

22:14 <discocaml> <masterbuilder> No, like, Assertion failed: lhs =/= rhs, printing the actual values that were compared, I mean, I guess they might not be printable in all cases

22:15 <discocaml> <masterbuilder> it's nice to know why the assertion failed

22:15 <discocaml> <masterbuilder> there isn't really a good way to write a function that can do that for all types yourself as far as I can tell

22:16 <discocaml> <NULL> Yeah, that would need a universal printer, which OCaml doesn't have

22:16 <discocaml> <masterbuilder> so is there any workaround?

22:16 <discocaml> <NULL> Also, assert takes a bool, so it wouldn't know what to print

22:16 <discocaml> <masterbuilder> yes fair point, but you get the idea

22:17 <discocaml> <NULL> Construct your own exception and build a string with the info you need

22:17 <discocaml> <masterbuilder> hmm, alright

23:02 <companion_cube> But assert tells you the file and line, doesn't it?

23:09 <discocaml> <masterbuilder> Yes, but if it's not a statically determinable assertion and it depends on runtime conditions and it's checked a bunch of times before failing then knowing where the assert is located in the code is not particularly insightful

23:10 <discocaml> <masterbuilder> exceptions are probably better for that, I just placed some quick and dirty asserts here and there

23:10 <companion_cube> Ah, right

23:11 <discocaml> <RegularSpatula> Are you already using Janestreet libs? If so you can use ppx_assert

23:12 <discocaml> <RegularSpatula> It will print inputs on failures, tho you can’t turn them off with noassert option

23:33 <companion_cube> I don't use any jst lib, if I'm debugging that kind of stuff I add a log statement in the assert

23:33 <companion_cube> `assert (let ok =... in if not ok then log; ok)`