00:21 <d_bot> <Anurag> @companion_cube httpaf is indeed fast, but there is room for improvement in how IO is handled (using Lwt and Async). As an example there are some benchmarks done in the ocaml multicore project <https://github.com/ocaml-multicore/retro-httpaf-bench/pull/16#issuecomment-914852288>. Multiple httpaf runtimes, running in a single threaded mode but with very different perf characteristics!

00:31 <companion_cube> Hu, interesting.

00:37 <d_bot> <Anurag> That said, all of the httpaf versions look pretty decent to me personally. At some point I gotta look at realistic workloads instead of feeling bad about potentially being slower than Rust :P the increase in productivity with OCaml saves me more time during development and maintaining 😄

01:52 <d_bot> <joohnym> I'm not sure that matters as much. One core thermal throttling is not that different than all cores thermal throttling from the perspective of heat output. It sometimes helps to spread the workload across multiple cores to produce less heat.

02:07 <smondet[m]> rgrinberg: BTW I'm with you not liking when a dumb computation takes over all the cores of my laptop (it's usually a build or a web-browser though). Still, I often wish Emacs knew how to use maybe just 2 cores instead of 1 to avoid all those UI freezes ;)

08:38 <CodeBitCookie[m]> smondet: To get around this problem, I open up vterm, type in a command and right at then end of the command I put a space and the the '&' symbol, for example: 'firefox https://searx.me &'''""'''

08:39 <CodeBitCookie[m]> s/smondet: To get around this problem, I open up vterm, type in a command and right at then end of the command I put a space and the the '&' symbol, for example: 'firefox https://searx.me &'''""'''/smondet: To get around this problem, I open up vterm, type in a command and right at then end of the command I put a space and the the '&' symbol, for example: 'firefox https://searx.me &'. You can also install the 'multi-vterm' package for

08:39 <CodeBitCookie[m]> multiple instances of vterm. It is the fastest terminal emulator. Don't forget the shell side configuration/

08:40 <CodeBitCookie[m]> smondet: You can also make an emacs function to ask you with `ivy` or `helm` (or something else) for a command and automatically add the '&' at the end. Then just run it with vterm and bury the buffer.

08:41 <CodeBitCookie[m]> if you use evil-mode you already have that feature by typing in ':!firefox https://searx.me &' in normal mode.

13:19 <companion_cube> Isn't the fastest alacritty? :)

13:51 <Tardigreat[m]> i'd guess fastest is relative depending on your usecase for a terminal... all i care about is that it starts quickly. GPU rendering is completely useless to me

13:51 <Tardigreat[m]> anything that is spammy i pipe through awk to reduce the noise anyway

13:52 <d_bot> <darrenldl> who doesnt want a 4k 60fps terminal

13:54 <Tardigreat[m]> i'm sure it's nice if you make use of it, i just don't and i have other priorities in a terminal emulator

13:55 <d_bot> <darrenldl> there was meant to be a /s heh

13:55 <Tardigreat[m]> ah :)

13:56 <d_bot> <darrenldl> emacs did freeze on me a fair bit sadly, otherwise pretty nice

15:34 <Leonidas> ehh, I would want an alacritty with ligature support

15:34 <Leonidas> but so far only kitty realistically provides ligatures because VTE support is stuck in development hell

17:02 <d_bot> <Uhuru> Is there any good tools to format the menhir code on emacs,messy code makes crazy now...but ocamlformat seems not work on .mly file😩

21:54 <d_bot> <RegularSpatula> Just saw this on 4.13 release notes…”* #10081: Typecheck `x |> f` and `f @@ x` as `(f x)`”. Wondering if this will affect a weird thing I’ve run into with option types where `maybe |> Option.value_exn` will work but `Option.value_exn @@ maybe` gives an error

22:03 <d_bot> <leviroth> Do you use jane street ppxes by chance?

22:09 <d_bot> <leviroth> If so, the difference is probably due to https://github.com/janestreet/ppx_pipebang, and I believe 4.13 makes that ppx unnecessary, and both `|> Option.value_exn` and `Option.value_exn @@` should work.

22:12 <d_bot> <RegularSpatula> Oh that’s interesting, yeah I use ppx_jane often, and it does include ppx_pipebang