10:42 <discocaml> <lost.ray> Any problems with OCaml on MacOS with ARM chips? Thinking about buying M4 Pro...

11:36 <discocaml> <octachron> There are quite few core maintainers on MacOS ARM64, so there are no major issues. There are from time to time ephemeral minors issue after Xcode update because Apple is bit too keen on breaking backward compatibility on their compiler toolchain.

11:51 <hanker> Hey hey

11:51 <hanker> for anyone pro with Lwt, is there a way to turn an `'a -> 'b Lwt.t` into an `('a -> 'b) Lwt.t` ?

12:04 <octachron> That doesn't sound possible in general: if the monadic effect of "'a -> 'b Monad.t" is dependent on the input "'a", you cannot compute before the input is available.

12:06 <hanker> Oh, yeah, that makes sense

12:06 <hanker> What about the more specific case `unit -> 'a Lwt.t` to `(unit -> 'a) Lwt.t` ?

12:09 <hanker> I have an interface that needs an `f` to call repeatedly to get elements, and I'd like to feed these elements from an `Lwt_stream`

12:15 <octachron> `f () >>| fun x () -> x` (or `f () >>= Fun.const`) aka starts the effect and wraps the result inside a `fun () -> ... ` lambda.

13:13 <hanker> Ohh, thanks a lot!

13:57 <hanker> oh, err-

13:58 <hanker> that's slightly different from what I was looking for

13:58 <hanker> basically, it wraps the return of `f ()` into the `Fun.const`, so calling the generated function repeatedly yields the same result every time

14:40 <hanker> I'm not sure that'd be possible though

18:58 <discocaml> <gabyfle> Hello. I'm using the OCaml FFI to build up a Bigarray. I'm also using ASAN to check for possible leaks.

18:58 <discocaml> <gabyfle> I'm allocating the bigarray using this:

18:58 <discocaml> <gabyfle> ```cpp

18:58 <discocaml> <gabyfle> barray = caml_ba_alloc(type_flag | CAML_BA_C_LAYOUT, ndims, NULL, dims);

18:58 <discocaml> <gabyfle> ```

18:58 <discocaml> <gabyfle> from the documentation, it says that if the `data` pointer is null, the memory will be managed by OCaml. However, using ASAN, it detects that the memory that should've been managed by OCaml has leaked:

18:58 <discocaml> <gabyfle>

18:58 <discocaml> <gabyfle> ```

18:58 <discocaml> <gabyfle> ==279035==ERROR: LeakSanitizer: detected memory leaks

18:58 <discocaml> <gabyfle>

18:58 <discocaml> <gabyfle> Direct leak of 10274952 byte(s) in 2 object(s) allocated from:

18:58 <discocaml> <gabyfle> #0 0x7f4d5a9649c7 in malloc ../../../../src/libsanitizer/asan/asan_malloc_linux.cpp:69

18:58 <discocaml> <gabyfle> #1 0x55cb82d6af30 in caml_ba_alloc runtime/bigarray.c:246

18:58 <discocaml> <gabyfle> #2 0x55cb825f95eb in long caml_read_audio_file<float>(long, long) /home/gabyfle/Code/soundml/_build/default/src/io/soundml_io.cxx:348

18:58 <discocaml> <gabyfle> #3 0x55cb82dd2bbe in caml_c_call runtime/amd64.S:736

18:58 <discocaml> <gabyfle>

18:58 <discocaml> <gabyfle> Direct leak of 40 byte(s) in 1 object(s) allocated from:

18:58 <discocaml> <gabyfle> #0 0x7f4d5a9649c7 in malloc ../../../../src/libsanitizer/asan/asan_malloc_linux.cpp:69

18:58 <discocaml> <gabyfle> #1 0x55cb82dc4dfc in sync_mutex_create runtime/sync_posix.h:44

18:58 <discocaml> <gabyfle> #2 0x55cb82dc4dfc in caml_ml_mutex_new runtime/sync.c:77

18:58 <discocaml> <gabyfle> #3 0x55cb82dd2bbe in caml_c_call runtime/amd64.S:736

18:58 <discocaml> <gabyfle>

18:58 <discocaml> <gabyfle> Direct leak of 32 byte(s) in 1 object(s) allocated from:

18:58 <discocaml> <gabyfle> #0 0x7f4d5a9649c7 in malloc ../../../../src/libsanitizer/asan/asan_malloc_linux.cpp:69

18:58 <discocaml> <gabyfle> #1 0x55cb82d6af30 in caml_ba_alloc runtime/bigarray.c:246

18:58 <discocaml> <gabyfle> #2 0x55cb82d6b42b in caml_ba_create runtime/bigarray.c:671

18:58 <discocaml> <gabyfle> #3 0x55cb82dd2bbe in caml_c_call runtime/amd64.S:736

18:58 <discocaml> <gabyfle>

18:58 <discocaml> <gabyfle> SUMMARY: AddressSanitizer: 10275024 byte(s) leaked in 4 allocation(s).

18:58 <discocaml> <gabyfle> ```

18:58 <discocaml> <gabyfle>

18:58 <discocaml> <gabyfle> Hello. I'm using the OCaml FFI to build up a Bigarray. I'm also using ASAN to check for possible leaks.

18:58 <discocaml> <gabyfle> I'm allocating the bigarray using this:

18:58 <discocaml> <gabyfle> ```cpp

18:58 <discocaml> <gabyfle> barray = caml_ba_alloc(type_flag | CAML_BA_C_LAYOUT, ndims, NULL, dims);

18:58 <discocaml> <gabyfle> ```

18:58 <discocaml> <gabyfle> from the documentation, it says that if the `data` pointer is null, the memory will be managed by OCaml. However, using ASAN, it detects that the memory that should've been managed by OCaml has leaked:

18:58 <discocaml> <gabyfle>

18:58 <discocaml> <gabyfle> ```

18:58 <discocaml> <gabyfle> ==279035==ERROR: LeakSanitizer: detected memory leaks

18:58 <discocaml> <gabyfle>

18:58 <discocaml> <gabyfle> Direct leak of 10274952 byte(s) in 2 object(s) allocated from:

18:58 <discocaml> <gabyfle> #0 0x7f4d5a9649c7 in malloc ../../../../src/libsanitizer/asan/asan_malloc_linux.cpp:69

18:58 <discocaml> <gabyfle> #1 0x55cb82d6af30 in caml_ba_alloc runtime/bigarray.c:246

18:58 <discocaml> <gabyfle> #2 0x55cb825f95eb in long caml_read_audio_file<float>(long, long) /home/gabyfle/Code/soundml/_build/default/src/io/soundml_io.cxx:348

18:58 <discocaml> <gabyfle> #3 0x55cb82dd2bbe in caml_c_call runtime/amd64.S:736

18:59 <discocaml> <gabyfle>

19:09 <discocaml> <octachron> The OCaml garbage collector and runtime don't clean up at exit (because this would delay closing the program). So if the array was still alive at the end of the process this is the expected behavior. There is cleanup at exit mode for OCaml 4 (which should be back with OCaml 5.4).

19:15 <discocaml> <gabyfle> nice, thanks you :p I was afraid I was misusing the allocation of bigarrays somewhere

19:31 <discocaml> <octachron> An option to remove this false positive might be to make sure that the array is not alive and call `Gc.full_major` before exiting.

22:43 <discocaml> <tornato> alright i just wrote a rudimentary datalog interpreter in Ocaml, and I gotta say, I really like this language

22:43 <discocaml> <tornato> alright i just wrote a rudimentary datalog interpreter in Ocaml, and I gotta say, I really like this language (Ocaml)

22:44 <discocaml> <tornato> alright i just vibe coded a rudimentary datalog interpreter in a couple hours in Ocaml, and I gotta say, I really like this language (OCaml)

22:50 <discocaml> <yawaramin> it hits a sweet spot

22:51 <discocaml> <tornato> I was skeptical at first but the tooling is actually phenomenal

22:51 <discocaml> <yawaramin> they've put in a huge amount of work into it over the past few years

22:53 <discocaml> <tornato> indeed, also, the little things like being able to do

22:53 <discocaml> <tornato>

22:53 <discocaml> <tornato> ```

22:53 <discocaml> <tornato> match something with

22:53 <discocaml> <tornato> | Const c | Ident c -> ... -- this thing here, I forget what this is called

22:53 <discocaml> <tornato> | Str c -> ...

22:53 <discocaml> <tornato> ```

22:54 <discocaml> <yawaramin> i think or-patterns