00:55 <discocaml> <mirazzzzzz> Hello, we are looking for employees for a cryptocurrency project:

00:55 <discocaml> <mirazzzzzz>

00:55 <discocaml> <mirazzzzzz> * Beta tester ($45/hour)

00:55 <discocaml> <mirazzzzzz> * Moderator ($500/week)

00:55 <discocaml> <mirazzzzzz> * Community Manager ($1000/week)

00:55 <discocaml> <mirazzzzzz> * NFT-artist (negotiable)

00:55 <discocaml> <mirazzzzzz> * Advertisers (negotiable)

00:55 <discocaml> <mirazzzzzz> * Web Developer (negotiable)

00:55 <discocaml> <mirazzzzzz> * Investors

00:55 <discocaml> <mirazzzzzz>

00:55 <discocaml> <mirazzzzzz> If you are interested in collaborating, please send me a DM

00:56 <discocaml> <Ada> wasn’t this guy in here yesterday

00:56 <discocaml> <lhommemagique> yup

06:54 <toastal> With compiler plugins removed, how does one get back the functionality of `ocaml-m17n` in 2024? I was looking to use it to do some teaching, but English & Latin scripts are not the students’ first language.

06:55 <toastal> It would be helpful towards their accessibility if ASCII limitations were lifted.

06:55 <toastal> For reference, the m17n’s Microsoft GitHub project: https://github.com/whitequark/ocaml-m17n

08:22 <octachron> toastal, which script would you need support for?

08:24 <octachron> At some point in the future, it should be relatively easy to extend language support by writing a custom driver for compiler-libs.

08:27 <octachron> While we are not yet here, if you have a pedagogical use case in mind, I could have a look at writing a compiler variant with the missing bit and lexer support.

08:34 <toastal> Thai script, which is not bicameral

08:37 <octachron> So the support for "casing alternative marker" from m17n is mandatory. Noted.

08:41 <toastal> Yes. Even PureScript, Haskell, etc. which can tolerate Greek, Cyrillic, et al. have no way of handling the majority of scripts (which are not bicameral). But these languages I believe are ultimately too complicated for beginners.

08:42 <toastal> The proposed sigil was a concise solution to the problem

08:43 <toastal> (albeit *possibly* conflicting with prefix operators)

08:46 <octachron> Or code using the fact that "x@l" is currently valid.

09:07 <octachron> Anyway, I will try to have a look at an opam switch variant equivalent to m17n this week.

09:40 <toastal> octachron: that would be awesome!

14:39 <discocaml> <idontwantausernamefeckoff> i know dev is english centric but still shocking to me how rare that type of support is

14:49 <discocaml> <octachron> It is bit more than being english centric, unicode is both recent, complicated, and still a work-in-progress. Unicode properties tables are megabytes of pure data that are better kept updated outside of the compiler release cycle for instance.

14:51 <discocaml> <octachron> That's why my current plan for full internationalization support is to have the core compiler support minimal utf8+unicode and have extended version of the compiler that switches the lexer, the module name normalization functions and the user display function to ones that support full internationalization.

14:53 <discocaml> <bluddy5> Unicode has gotten ridiculously bloated

14:53 <discocaml> <idontwantausernamefeckoff> you’re telling me you’re not enthused by seven modifiers each on 500 emoji

14:55 <discocaml> <idontwantausernamefeckoff> or characters that are both visually identical to ascii and allowed in urls

15:04 <discocaml> <octachron> From my point of view those are fine, once one integrate the rule that if you are not doing the graphical rendering you don't try to parse user input, you only transport them as black boxes. Ok, it would be nice to be able to compute a graphical length of a string, and I am still waiting for support for the hieroglyphic control characters anywhere, but surely that will happen some day?

15:10 <discocaml> <idontwantausernamefeckoff> any questionable unicode decision is sort of outweighed by the fact that browsers and the like ever importing this huge, intentionally all-encompassing character set didn't check it over first

15:57 <discocaml> <darrenldl> would it be reasonable to have a pinned post on discuss.ocaml.org that discourages use of ocamlwiki.com?

17:02 <discocaml> <idontwantausernamefeckoff> i have never heard of that site before

18:09 <discocaml> <functionalprogramming> Do expect tests need to be inline? or can they be in a test directory?

18:09 <discocaml> <regularspatula> They can be in a test directory or separate library too

18:10 <discocaml> <regularspatula> (If you’re not testing internal implementation at least)

18:12 <discocaml> <functionalprogramming> right

18:13 <discocaml> <functionalprogramming> okay, so i cant use a test stanza, it has to be library

18:16 <discocaml> <regularspatula> Separate library is what I do

18:16 <discocaml> <functionalprogramming> how do i get the inline_tests_runner script?

18:17 <discocaml> <regularspatula> I generally do not use it directly…rather use dune to run tests

18:18 <discocaml> <functionalprogramming> ah cool! runtest works on the library stanza

18:18 <discocaml> <functionalprogramming> didnt expect it to

18:18 <discocaml> <functionalprogramming> sorry to bother you, but do you know a convenient way to replace the file with the expect output file?

18:18 <discocaml> <regularspatula> Use dune promote

18:19 <discocaml> <functionalprogramming> thanks

18:19 <discocaml> <functionalprogramming> god this is so nice

18:19 <discocaml> <regularspatula> You can also use auto promote with watch mode if you are still exploratory

18:19 <discocaml> <regularspatula> (Or adventurous)

18:19 <discocaml> <functionalprogramming> hehe

18:19 <discocaml> <functionalprogramming> thanks

18:19 <discocaml> <regularspatula> No problem

18:20 <discocaml> <functionalprogramming> it is curious how promote is not even mentioned on the ppx_expect readme

18:23 <discocaml> <regularspatula> Oh is it not…hmm that’s odd

18:25 <discocaml> <regularspatula> Btw there is a thing in dune docs about ppx_inline_test dependency not allowed to be a with-test dependency. But if you put the inline tests in another library than it can be in my experience

18:47 <discocaml> <christopherfujino> does anyone talk about reasonml here, or is it forbidden?

18:55 <discocaml> <undu> It's not forbidden, just not much interest

18:55 <discocaml> <christopherfujino> interesting

18:56 <discocaml> <christopherfujino> i'm not interested in using it, but i'm curious if it's more effective in adopting new users than regular ocaml

18:56 <discocaml> <christopherfujino> i guess they have their own discord server though

19:07 <discocaml> <idontwantausernamefeckoff> when i started learning ocaml reason seemed like a dying project

19:21 <discocaml> <bluddy5> Reason started out on a massive upward trajectory. When we started this discord, they were outnumbering us 3 to 1 (which was great -- people were being introduced into the ecosystem). They then lost a critical component of the ecosystem (Bucklescript), and that caused many people and companies to lose confidence in the language. At this point OCaml has more online people than the Reason server, which tells you a lot.

19:30 <discocaml> <functionalprogramming> so you have the tests, and you have the library with the inline tests, and you depend on that library in your tests?

19:34 <discocaml> <Et7f3 (@me on reply)> Melange try to give confidence back (and reason is first class citizens)

19:39 <discocaml> <bluddy5> Correct. But it'll be tough. Once you've betrayed people's trust, it's very hard to get it back.

20:11 <discocaml> <regularspatula> I will generally have “real library” like A then another library Test_A. Then Test_A requires A. A knows nothing about Test_A

20:11 <discocaml> <idontwantausernamefeckoff> i never quite understood reason

20:12 <discocaml> <idontwantausernamefeckoff> the syntax isn't that scary and if i'm going to learn a new way of programming (as ocaml was since the only functional code i'd written before that was array processing in javascript) i would rather learn a new syntax to go along with the new concepts

20:14 <discocaml> <idontwantausernamefeckoff> anyway curly spaces isn't any more readable than parentheses and `then`-style statements. i like typescript and i've still written a lot of ruby that i find more readable

20:14 <discocaml> <regularspatula> (Adjust capitalization as required for lib/module/etc). I the test lib I will have a mix of ppx_expext, ppx_inline_test and quick check tests. Then a separate dir for cram tests usually

22:57 <mizlan> What is the difference between @[<b> and @[<hov> ? The docs say the former is "demonstrating indentation" but I don't get what that means

22:57 <mizlan> Since you can put indentation on <hov>

22:59 <mizlan> in Format module btw^

23:08 <octachron> with b boxes, the formatting engine also breaks line when this lead to a change in indentation

23:24 <companion_cube> Hu, I didn't even know about b boxes :o