01:41 <mrm> bremner: Oh, thanks for updating backports to 7.1 :-)

02:08 <aidalgol> technomancy, single-handedly taking on LSP.

02:15 <technomancy> my ocaml codebase is hilariously broken for no good reason because at the time there was no way to lock my dependencies to the correct version, and everyone in #ocaml was like "what's the big deal; it'll be fine; it probably won't break."

02:16 <technomancy> NARRATOR VOICE: it was not fine.

02:16 <aeth> technomancy: but if it compiles it works

02:16 <technomancy> I was getting like ... linker errors?

02:17 <technomancy> I kind of know what a linker is

02:22 <aeth> well I think that's part of the compiler so at least it lives up to the promise

02:22 <aeth> technomancy: how is ocaml? Is it worth studying for language design ideas?

02:23 <technomancy> I've barely scratched the surface really

02:23 <technomancy> it was surprisingly easy to learn the language, and unsurprisingly tedious and difficult to learn the build tooling

02:24 <technomancy> I published a short codebase for a dev tool designed to help with clojure development in ocaml and I got a couple contributors sending me stuff who had never used ocaml before but were able to pick it up well enough to make changes

02:24 <technomancy> (just over 500 lines)

02:28 <aidalgol> technomancy: Wow... that makes ocaml package-management worse than JavaScript's.

02:28 <jcowan> There is only one build tool that is any good, and that is Prolog, er, I mean `make`.

02:28 <aidalgol> Hint: If the *JavaScript* ecosystem is better than you at something, you're doing it *really* badly.

02:30 <aidalgol> jcowan: The only reason I do not use make for my web stuff is that so many tools I need in the pipeline are *only* available as @*$#ing webpack plugins, with no CLI counterpart.

02:32 <aeth> why use make when you can use bash?

02:32 <aidalgol> aeth: Um, dependency resolution?

02:33 <aeth> Technically, not doing dependency resolution automatically satisfies technomancy's criteria of locking dependencies to the correct version. Just probably not in a fun way.

02:35 <jcowan> If you try to build Chibi Scheme by typing ./configure, it says "Autoconf is an evil piece bloatware encouraging cargo-cult programming. Make, on the other hand, is a beautiful little prolog for the filesystem. Just run 'make'."

02:37 <jcowan> When you start work at Google they make sure your desktop is all set up before you get there. When Stu Feldman, the author of `make`, arrived for his first day, they had given him a keyboard with the Tab key missing.

02:53 <aidalgol> aeth: Er, different kind of dependency resolution.

03:06 <aeth> jcowan: Just C-q C-i if you need to tab in Emacs.

03:06 <jcowan> I don't use Emacs, remember.

03:13 <brendyyn> jcowan: what do you use?

03:14 <jcowan> I am an 'ex' troglodyte, with excursions to vi-mode for parenthesis bouncing

03:14 <brendyyn> i have no idea what that means

03:23 <notzmv> jcowan: are you the author of the CLL? :o

03:23 <jcowan> That's me

03:23 <notzmv> wow

03:24 <jcowan> brendyyn: vi has two modes, a screen editor and a command line editor called ex. I mostly use the latter, but when I need to match parentheses in Lisp code I switch to vi mode

03:30 <greghendershott> Well that took some very circuitous searches for me to finally figure out that "the CLL" is "The Complete Lojban Language" (as opposed to various diseases and law firms). :)

03:30 <greghendershott> Very cool.

03:40 <friscosam> haha I was reviewing some R7RS stuff in the past couple of days and along the way have run into 2-3 unrelated jcowan references

03:41 <friscosam> The most recent was looking at some figlet font

05:12 <technomancy> damn; I thought CLL was Common Lisp the Language

06:10 <friscosam> CLtL iirc

06:10 <friscosam> :p

09:29 <aidalgol> Do we have any Matrix users in our midst? I'm (verrry slooowly) writing a Matrix bot in Racket.

09:30 <aidalgol> https://gitlab.com/aidalgol/racket-matrix-sdk/

09:31 <aidalgol> (named "SDK" for consistency with the other Matrix libs)

13:16 <bremner> mrm: welcome. For once I didn't just forget about, it actually took a long time to get it into debian testing.

16:46 <avemi> if I create a DSL with Racket, would it be possible to embed the DSL interpreter into a C program and to have some kind of a standard library for that DSL, but the library functions would not be implemented in the DSL, but just be C callbacks into the main program? I skimmed through <https://docs.racket-lang.org/inside/embedding.html> without any good clue, and so far I just need to know if what I want is

16:46 <avemi> achievable at all

16:47 <avemi> i.e. given a script written in such DSL, some commands/functions would be executed in the DSL environment, but some would call back to C functions and the information would be passed "up" to the main program

17:23 <bremner> avemi: maybe just using FFI is enough?

18:43 <odanoburu> hello! I have defined a little DSL, and written a test file using it as #lang. running the file returns me some result value. what is the best way to define a function that takes the path to this test file, runs the file, and does something with the result?