05:36 <d_bot> <Bluddy> Copilot uses GPT-3. It doesn't really understand what you want -- it just picks up on (complex) patterns.

05:44 <d_bot> <.kodwx> at least match the programming language though

05:44 <d_bot> <.kodwx> maybe using the file extension if the current file has no code

06:02 <d_bot> <Bluddy> it's a black box. you can't force it to do anything.

06:03 <d_bot> <Bluddy> and I assume the problem is that there just isn't enough OCaml data in github for it to refer to.

11:09 <d_bot> <dario_teixeira> It's been a few years since I've used Sqlite from OCaml. Back in the day I used the `ppx_sqlexpr` syntax extension, but it seems it hasn't been updated to work with more recent OCaml compilers. What are kids using these days? There's a tonne of Sqlite related packages in OPAM... (Thanks in advance!)

11:16 <reynir> I use caqti. It doesn't have any fancy syntax for writing queries though

11:21 <d_bot> <dario_teixeira> There's always also the option of using the low-level `sqlite3` bindings, but I'd rather use something higher-level, that would take care of preparing statements, (de)serialisation, etc. But I'll take a look at `caqti`.

12:04 <d_bot> <orbitz> I need to release my SQLite library. I've been lazy

12:13 <d_bot> <leviroth> I’ve never used it, but caqti has `ppx_rapper` if you want fancy syntax.

12:17 <d_bot> <dario_teixeira> Oh yeah, I had totally forgotten about `ppx_rapper`. The combination `caqti` + `ppx_rapper` should satisfy my requirements for a maintained, high-level interface to Sqlite. Thanks!

12:53 <companion_cube> I have my little wrapper around sqlite3 bindings, too

12:53 <companion_cube> with heterogeneous lists

14:34 <d_bot> <Drup> @octachron Nice work on the "detailed write up" for 4.14

14:39 <d_bot> <octachron> Thanks!

14:43 <d_bot> <Drup> You're going to be busy to do the same for 5.0 😄

14:46 <d_bot> <octachron> There is probably some personal bias visible through the level of detail on changes on which I have worked on or reviewed. But that will be hopefully an incentive for making the writing more collaborative for 5.0.

14:46 <d_bot> <octachron> Or otherwise, I will be very busy asking questions on the finer detail of ARM64 assembler instructions for 5.0...

14:47 <d_bot> <octachron> And I am not completely sure which option is the more realistic one.

17:53 <sleepydog> the release notes were excellent, and the additions themselves are great. it's really a great way to end the 4.x branch