00:34 <discocaml> <Ada> any idea if esy will support ocaml 5 soon?

00:35 <discocaml> <Ada> i’m fairly new to the language and the javascripty package management is appealing to me coming from go and nodejs

00:35 <discocaml> <Ada> but i’d rather use eio if i’m gonna write any prod code

04:32 <greenbagels> using ocamlbuild, if i have a source file (say foo.ml) that uses a record from bar.ml, do i have to provide bar.ml in any way for the build system to see record definitions from foo.ml in bar.ml?

04:32 <greenbagels> im getting an unbound type constructor error during compilation when trying to use a type annotation for the record type

04:43 <greenbagels> hm weird, a simpler test case worked fine, ok ill take another look at my code first

04:59 <greenbagels> ah i had an empty .mli file lingering about

04:59 <greenbagels> hahaha

07:29 <discocaml> <chrisarmstrong> Very cool to see - Ill have to check them out given I’ve been experimenting with generating aws bindings from smithy definitions

07:41 <discocaml> <chrisarmstrong> In JS, it used to be one giant package, but now it’s one package per service (sometimes more if there are special helpers like S3)

07:42 <discocaml> <chrisarmstrong> From what I can tell the boto definitions seem to be derived from the smithy ones as they use the same language and type system

13:40 <discocaml> <mbacarella> the ocaml compiler stack overflows if you try to compile .ml files that are O(100kloc)

13:45 <discocaml> <sim642> I'd understand if it was one big thing, but given how many packages it's split into, it's surprising that they'd still be so big

14:34 <discocaml> <mbacarella> only 2 or 3 packages require it. EC2 is pretty huge

17:22 <discocaml> <anmonteiro> Is this the case on OCaml 5? I thought the stack was unlimited since 5?

17:51 <discocaml> <octachron> It is not completely unlimited to be able to catch unbounded recursion but the default stack size has been considerably in OCaml 5 expanded since the OCaml stack is now decoupled from the system stack. Also with OCaml 5.1 switching to TRMC implementation for many list functions there are less source of stackoverflow (which also means more risk of compilation taking a long time due to non-linear algorithm in the code source size).

17:58 <dh`> there's no such thing as unlimited memory, except in turing machines operated by theoreticians

20:57 <discocaml> <vczf> Is the mathjax section rendering for anybody else or is it just my browsers? https://ocaml.org/docs/if-statements-and-loops#if-statements-loops-and-recursions

20:57 <discocaml> <vczf> https://cdn.discordapp.com/attachments/436568060288172044/1130241816235606027/Screenshot_2023-07-16_at_4.56.57_PM.png

21:01 <discocaml> <._null._> Broken for me as well

21:04 <discocaml> <._null._> Forwarded to #ocaml․org

21:07 <discocaml> <hockletock> weirdly it works on a local build of ocaml.org https://i.postimg.cc/JzhMScdR/Screenshot-from-2023-07-16-17-07-12.png

21:10 <discocaml> <._null._> Is it the same code ? Yours seems to be missing comments

21:11 <discocaml> <hockletock> No, it is not