01:13 <d_bot> <cemerick> @companion_cube just out of curiosity, do you only use ocp-indent?

01:32 <d_bot> <cemerick> I ask because containers, etc., seems quite nicely formatted, and in relatively sane ways. Every few months, I try to use ocamlformat again, and am quickly hit with cases where it does IMO really unpleasant stuff.

02:04 <d_bot> <darrenldl> oh huh, i ended up going from ocp-indent to ocamlformat

02:05 <d_bot> <darrenldl> curious where it fails for you @cemerick

02:10 <companion_cube> @cemerick, yes

02:10 <companion_cube> I still format by hand mostly (just leaving indentation to ocp-indent)

02:15 <d_bot> <cemerick> ok, that's explains it

02:15 <d_bot> <cemerick> I also mostly hand-format still, but would like to not

02:18 <companion_cube> I think the same, but ocamlformat always makes me unhappy

02:18 <companion_cube> I think the grammar doesn't lend itself well to that :(

02:19 <d_bot> <cemerick> @darrenldl eh, it's a different thing each time; the current frustration is that it does bad things to "pipeline" expressions (series of operations chained with `|>`). Depending on the subexpressions' length, it either leaves them all on one line (which is fine for 2, *maybe* three very short exprs), or it wraps them into irregular separate lines: two exprs on one line, then a few single-line exprs, then another pair on one line

02:25 <companion_cube> for me the frustration is ( ) around ifs

02:25 <companion_cube> and the begin/end around match that I add manually

02:25 <d_bot> <darrenldl> i thought you can force |> to be on separate lines

02:25 <d_bot> <darrenldl> (not 100% tho

02:25 <d_bot> <darrenldl> 100% sure*

02:28 <d_bot> <darrenldl> companion_cube: ah true, the removal(?) of () in if is not the most convenient

02:29 <d_bot> <darrenldl> side digression: portal stories mel is really good if you haven't played it already : D

02:30 <d_bot> <cemerick> @darrenldl if you know the knob that affects that, I'd be happy to try it.

02:30 <d_bot> <cemerick> I've looked, but hardly comprehensively

02:50 <d_bot> <cemerick> ech, it seems the relevant option is deprecated, apparently part of a broader effort to minimize the configurability of the tool in general 🙃

02:51 <companion_cube> -_-

03:50 <d_bot> <darrenldl> > apparently part of a broader effort to minimize the configurability of the tool in general

03:50 <d_bot> <darrenldl> ...right i vaguely recall trying to follow the footsteps of rust fmt and go fmt tools...

03:50 <d_bot> <darrenldl> recall mentioning of trying*

03:51 <d_bot> <darrenldl> i uh...i don't know about that tbh

03:51 <d_bot> <darrenldl> at this rate i might have to write my own formatter for security code : D (half jk

03:52 <companion_cube> tbh you'd be better off writing a new syntax that has one obvious formatting

03:53 <companion_cube> (Reason had it right on that point)

03:53 <companion_cube> new syntax also means you can have a convention from the start

03:53 <companion_cube> :p

03:53 <d_bot> <darrenldl> is reason still uh...dependable?

03:54 <companion_cube> not sure tbh

03:55 <companion_cube> but they had that idea right at least

03:55 <companion_cube> go and rust are also C-like (ish) and imho it formats more easily

03:55 <d_bot> <darrenldl> yep

04:00 <d_bot> <darrenldl> so dune ingests .re files with no problem whatsoever?

04:26 <companion_cube> I think so, these days

04:26 <d_bot> <cemerick> @darrenldl yeah, you can mix and match .ml and .re in the same project without any configuration or anything

04:27 <d_bot> <cemerick> very handy for using tyxml's reason ppx (IMO the best embedded templating option available)

04:27 <d_bot> <darrenldl> thats neat : D

04:27 <companion_cube> you mean the JSX like thing?

04:28 <d_bot> <cemerick> yes, sorry, tyxml-jsx is the actual name

04:29 <d_bot> <darrenldl> partial application looks more noisy though, hm...

04:29 <d_bot> <cemerick> "you'd be better off writing a new syntax that has one obvious formatting" >> lisp transliteration, someday 😈

04:30 <d_bot> <cemerick> @darrenldl jsx is definitely the bright spot of reason syntax, the rest is IMO a big downgrade

04:30 <d_bot> <darrenldl> racket dsl for writing ocaml : D

04:31 <d_bot> <darrenldl> yeah, i think i prefer ocaml syntax at a lot of spots

04:31 <d_bot> <darrenldl> hmmmm

04:31 <companion_cube> @cemerick the {} are an upgrade if you consider how bad match/if is in OCaml

04:31 <companion_cube> also I don't hate `let x = y;` nor the concise lambdas

04:33 <d_bot> <darrenldl> containers re when : D

09:31 <d_bot> <Et7f3> Dune use refmt to convert to binary ast (the same output of ppx) then feed it to regular ocaml compiler so they are compatible

09:32 <d_bot> <Et7f3> No need for containers re

10:21 <d_bot> <Azure Plasma> How do I know the scihub search extension isn't malicious? I've been asked to download it to use the search functionality

11:54 <namkeleser> Is there anyway to debug camlrunm binary data file?

11:54 <namkeleser> ./challenge: a /usr/bin/camlrunm script executable (binary data)

11:55 <namkeleser> Like using GDB

11:55 <namkeleser> Or any other tool

11:55 <namkeleser> for moscow ML

11:55 <namkeleser> This format is very niche

12:02 <d_bot> <darrenldl> > No need for containers re

12:02 <d_bot> <darrenldl> well it was more about companion_cube being a fan of re syntax

12:06 <namkeleser> darrenldl is that to me?

12:12 <d_bot> <darrenldl> oh sorry, to @Et7f3

18:40 <companion_cube> not a fan, just a bit jealous of some aspects of it :p

19:03 <d_bot> <mbacarella> is the ocaml github action workflow `windows-latest` a cygwin environment? or something else?

19:13 <sim642> The readme says Cygwin

19:13 <sim642> If you mean setup-ocaml

19:17 <d_bot> <mbacarella> oh. indeed. ok

19:17 <d_bot> <mbacarella> my build is failing on that one because termios.h can't be included, and the solution is to install "cygwin-devel" apparently? though i'd expect that to be installed by default.

19:18 <d_bot> <mbacarella> trying it anyway

19:31 <d_bot> <mbacarella> Package 'mingw64-x86_64-cygwin-devel' not found.

19:31 <d_bot> <mbacarella> huh.

19:44 <d_bot> <mbacarella> maybe i should ... try not including termios.h on cygwin

21:24 <d_bot> <Et7f3> List of package is here https://github.com/ocaml/setup-ocaml/blob/03f74ebd2896e6f07de1993216dbdfa29ccf7b20/src/opam.ts#L180

22:48 <bartholin> how do I get nice error informations in angstrom?

22:49 <bartholin> Like, line number, character number...