02:48 <d_bot> <darrenldl> i use mparser a lot, found it to be a pretty good toolbox

06:53 <d_bot> <thangngoc89> Are there any implementations of suffix-array available?

09:53 <d_bot> <Competitive Complications> I'm using `Printexc.record_backtrace true`, dune printenv shows `-g`, I'm even using bytecode build, I don't use Lwt, and I still get very bad stack traces

09:53 <d_bot> <Competitive Complications> Am I missing something?

09:55 <d_bot> <octachron> What do you mean by bad stack traces?

10:13 <d_bot> <Competitive Complications> @octachron Lots of function calls missing

10:14 <d_bot> <octachron> There might be some exception reraises that mess up the trace.

10:22 <d_bot> <Competitive Complications> hmmm

10:22 <d_bot> <Competitive Complications> there are none unfortunately

10:22 <d_bot> <Competitive Complications> only simple failwiths

11:25 <d_bot> <ggole> Are the missing functions tail called?

12:39 <indicator> I'm trying to build https://github.com/lemaetech/jsoo_todomvc.git, opam install js_of_ocaml-ppx works fine, but dune build (same opam switch, ocaml-base-compiler.4.12.0, and current working directory) thinks 'Error: Library "js_of_ocaml-ppx" not found.'. Any clues?

12:40 <indicator> https://pastebin.com/Yqc0mpP1

12:45 <d_bot> <octachron> This looks like you are missing an `eval $(opam env)`.

12:47 <d_bot> <Competitive Complications> some of them yeah

12:48 <d_bot> <ggole> It's natural for a function to disappear from the stack if it ends in a tail call

12:51 <d_bot> <Competitive Complications> how can i make it so that's not the case for specified functions?

12:51 <d_bot> <Competitive Complications> (I just realized i have way more tail calls than I naively expected)

12:52 <d_bot> <ggole> You could stick something after the tail call (to make it not a tail call)

12:53 <d_bot> <Competitive Complications> yeah, that's an interpreting function, so that would imply changing all the places it calls itself recursively

12:53 <d_bot> <Competitive Complications> i was more wondering about a `non-tailcall` annotation or something

12:54 <d_bot> <ggole> I don't remember seeing anything like that

12:55 <d_bot> <ggole> In general the stack can't be relied on the capture the past of a program; only the future

12:59 <indicator> No, I have run eval $(opam env)

13:00 <indicator> And I ran everything in that paste in that same shell

13:27 <d_bot> <Alistair> Anyone know how to do this? (sorry for reposting, but didn't really get an answer the first time)

13:29 <d_bot> <Alistair> (For context, looking at adding tests to: https://github.com/johnyob/ppx-open)

13:29 <indicator> Interestingly, `ls ~/.opam/4.12.0/lib/ | grep js_of` only shows `js_of_ocaml` and `js_of_ocaml-compiler` in there, no `js_of_ocaml-ppx`

13:37 <indicator> Even if I create a new switch and set that switch, opam install still thinks js_of_ocaml-ppx is installed. Makes me think it's found from a system install, but I don't see one.

14:03 <indicator> Ahh, I had an `_opam` directory in my working dir (at the root of the jsoo_todomvc project). I think that came from me running `make dev-switch` in jsoo_todomvc (experimentally, not really knowing exactly what it would do, then forgetting about it), which runs `opam switch create . ocaml-base-compiler.4.10.0 --deps-only --with-test --working-dir`. So, knowing that, I see from the docs that opam thinks that a directory switch (a concept

14:03 <indicator> I didn't know about until just now) overrides the env var configuration from `opam env`, but... dune doesn't? Still confused about that last part I think.

17:38 <d_bot> <EduardoRFS> that's not an option, OCaml just emits jmp instead of call every time it's possible

17:38 <d_bot> <EduardoRFS> which is at almost any case

19:19 <remexre> did something happen to Typedtree? I remember it being part of compiler-libs, but it's not on https://ocaml.org/api/compilerlibref/index.html

20:35 <octachron> It is still part of the library, but it is in the documented part

20:35 <octachron> Partly because it is even more unstable than the parsetree.

20:37 <remexre> hm, okay...

20:39 <octachron> it is *not* in the documented part

20:41 <remexre> er yeah I misread that the inverse way you mistyped it lol

21:41 <d_bot> <monk> i think i'm going to try reading that opal library till i understand what it's doing for all 150loc lol

22:59 <d_bot> <monk> silly question, but can someone explain why `()` needs to be added for application to occur in this context?

22:59 <d_bot> <monk> ```ocaml

22:59 <d_bot> <monk> utop[46]>let ite x e1 e2 = match x with true -> e1 | false -> e2 ;;

22:59 <d_bot> <monk> - : val ite : bool -> 'a -> 'a -> 'a = <fun>

22:59 <d_bot> <monk> utop[47]> ite true (fun _ -> 4) (fun _ -> List.hd []) ;;

22:59 <d_bot> <monk> - : '_weak3 -> int = <fun>

22:59 <d_bot> <monk> utop[48]> (ite true (fun _ -> 4) (fun _ -> List.hd []))() ;;

22:59 <d_bot> <monk> - : int = 4

22:59 <d_bot> <monk> ```

22:59 <d_bot> <monk> does `()` get used to represent both unit and function application?

23:00 <d_bot> <monk> i took an example that had `()` originally in the expression arms of the match statement, ie `true -> e1() | false -> e2()`just to mess around

23:04 <d_bot> <Splingush> Both your arguments are functions. So evaluating `ite true fun1 fun2` returns one of the functions. The `()` is then the argument used the evaluate the returned function.

23:05 <d_bot> <monk> ah, thank you

23:52 <d_bot> <monk> what about caching/laziness makes this function no longer print out `"hello world"` after the thunk is forced?

23:52 <d_bot> <monk> ```ocaml

23:52 <d_bot> <monk> utop[67]> let lazy_hello = lazy begin print_endline "hello world"; 1 end ;;

23:52 <d_bot> <monk> val lazy_hello : int lazy_t = <lazy>

23:52 <d_bot> <monk> utop[68]> lazy_hello ;;

23:52 <d_bot> <monk> - : int lazy_t = <lazy>

23:52 <d_bot> <monk> utop[69]> Lazy.force lazy_hello ;;

23:52 <d_bot> <monk> hello world

23:52 <d_bot> <monk> - : int = 1

23:52 <d_bot> <monk> utop[70]> Lazy.force lazy_hello ;;

23:53 <d_bot> <monk> - : int = 1

23:53 <d_bot> <monk> ```