00:33 <companion_cube> so tired of recompiling whole switches because apparently OCaml has changed (!)

06:59 <discocaml> <._null._> This hasn't happened to me for a while, what version does this concern ?

08:28 <reynir> It can affect any version where its opam file is modified https://github.com/ocaml/opam-repository/commits/master/packages/ocaml-base-compiler'

08:29 <reynir> oops, the ' is not part of the URL

08:30 <reynir> I think for example this commit that removes post-messages will trigger a rebuild of the compiler which is annoying and a bit silly https://github.com/ocaml/opam-repository/commit/058cc373d548aa074576ed0bba8b30e14c6fa8a6 (please correct me if I'm wrong)

08:33 <discocaml> <hannes6838> according to https://github.com/ocaml/opam-repository/pull/26585#issuecomment-2358644559 this shouldn't trigger a rebuild. but there are several opam versions in the wild, so it is frustrating, and the question is how to move there (maybe a question for the opam-repository maintainers -- maybe compiler packages shouldn't be updated unless e.g. dune is released (which triggers recompilations anyways)?

08:42 <reynir> Oh cool they actually do a lot of checks to see if it's necessary

08:43 <discocaml> <peuk> Is there any way to put `ocamlformat` in opt-out mode? In the sense that, by default, it does nothing (or say, only indent) so that we can progressively opt-in new actions.

08:46 <discocaml> <peuk> I tried to apply it on an existing project and I see more places where it makes the code harder to read than before than the opposite.

09:24 <discocaml> <emiletrotignon> It is not possible to use `ocamlformat` only to indent because it does not indent, it parses the code into an extended ast which it then prints from scratch.

10:25 <discocaml> <andrew.02907> I’ll help 20 people on how to earn $70k or more within 72hours from the Crypto Market, but you’ll pay me a 10% of your profit when you receive it. Ask me how!!

10:25 <discocaml> <andrew.02907> If interested, send me a direct message. For more information…contact via.. WhatsApp

10:25 <discocaml> <andrew.02907> ‪+1 (424) 424‑6672‬

10:55 <discocaml> <Kali> @admin

11:11 <discocaml> <reynir1396> Oh, is spam the reason why I get unread channels and then there's nothing new when I open them??

11:11 <discocaml> <reynir1396> It's been driving me crazy

11:15 <discocaml> <octachron> Indeed, unfortunately I don't know any options to ban spams and restore channel state to unread.

11:29 <discocaml> <reynir1396> No worries. I'm just glad to have an explanation (discord gas lighting me)

11:57 <companion_cube> It's probably a discord bug because they don't want to recompute the notification status for every channel or something

12:06 <discocaml> <pmetzger> It’s not a bug, it’s a feature! But you can always report this as an unwanted feature to them.

12:18 <companion_cube> :D heh

12:58 <discocaml> <peuk> There is a bot named Not a Bot that can stop spammers before it reaches too many channels.

13:01 <discocaml> <peuk> https://github.com/DigitalPulseSoftware/NotaBot

13:17 <discocaml> <eduardorfs> There is any good way for a ppx to inject an error on the LSP and the compiler?

13:17 <discocaml> <eduardorfs> Such as a node with an attribute and a proper error message?

13:21 <companion_cube> [%error "ohno"] works iirc

13:29 <discocaml> <eduardorfs> right, it's in the compiler, but I couldn't find the docs for it tho

13:47 <discocaml> <peuk> Sad 😦

13:47 <discocaml> <peuk> Thanks for the answer

14:02 <discocaml> <emiletrotignon> There is ocp-indent though to do indentation !

14:14 <discocaml> <peuk> Yes, this is what we use currently. But :

14:14 <discocaml> <peuk> - it does not integrate well with `dune` so we ended up with a dedicated script to call it on the entire project,

14:14 <discocaml> <peuk> - there are other formatting actions that are interesting (for example spaces around `:` in signatures, etc).

14:14 <discocaml> <peuk> Reason why I had a look to ocamlformat but I am pretty sure that I cannot convince all the team to move to ocamlformat since it removes blank lines in MLIs, sometimes change the code to oneliners, etc.

14:16 <companion_cube> after some years of not using ocamlformat I eventually found a config I liked enough

14:16 <companion_cube> but it's a pain for sure

14:29 <discocaml> <emiletrotignon> It is going to do things like that indeed. It gives a huge benefit of not having to care about formatting. Most people I have heard who do not like it already care about formatting so ocamlformat is never going to be exactly like what they would want.

14:29 <discocaml> <emiletrotignon> I started using ocamlformat before having strong tastes in formatting so I think its great (then I got to look at the codebase a little and I do not think it is great anymore but thats another story)

14:30 <companion_cube> I'd hate to have to change my config, but with it ocamlformat is good enough for me

14:30 <companion_cube> and I care less about things that used to upset me

15:32 <discocaml> <darrenldl> i just worry ocamlformat will stop supporting some options should i rely on it

15:33 <discocaml> <darrenldl> i still call ocp-indent after dune via a makefile

17:28 <discocaml> <deepspacejohn> For new projects I just use the default ocamlformat. Sometimes I end up changing a couple of rules if there's something specific that bugs me (usually whatever rule that puts line breaks after ; sequences) but only if that thing comes up often enough to warrant me caring about it.

17:29 <companion_cube> is the default one the one that puts `;` at the beginning of lines?

17:29 <discocaml> <deepspacejohn> no, not the default profile.

19:04 <discocaml> <anmonteiro> check `Builtin_attributes.error_of_extension`

19:04 <discocaml> <anmonteiro> not documentation, but it's something to read at least

19:10 <discocaml> <eduardorfs> yeah yeah, I saw this when CC mentioned the [%error]

19:21 <Anarchos> where to add a type variable in " type d = [<`A| `B];;" to be correct ?

19:26 <discocaml> <anmonteiro> usually something like ``type 'a d = [< `A| `B] as 'a``

22:04 <eilvelia> /save