runs a custom fork of opencode, had to make room for some distributed systems (swapped in Postgres for persistence and NATS as a message bus)
feature-parity with regular opencode so any config is inherently compatible, and all versions are compatible (cli, web, desktop etc)
Been exploring running Opencode as an ephemeral worker agent in a distributed sandboxed environment under Kubernetes.
Ideal for an enterprise setting where teams want to run their agents in a controlled environment. Each session is isolated, sandboxes short or long lived depending on the need.
Wrapper built with @EffectTS_ so @kitlangton would approve (not sure about @thdxr
it's kubernetes + effect after all)
Will be looking for some testers soon if you have a cluster and you'd want to try it
runs a custom fork of opencode, had to make room for some distributed systems (swapped in Postgres for persistence and NATS as a message bus)
feature-parity with regular opencode so any config is inherently compatible, and all versions are compatible (cli, web, desktop etc)
@saltyAom@saltyAom do you have by any chance any guides or tips to use Effect ? I have been exploring tooling for APIs recently and getting Elysia and Effect to play along would be great
Building a new open source library for react native that is composed of various reusable blocks and interface scenes. Perfect for quickly bootstrapping and building your apps.
Uses Uniwind for Tailwind styles, and uses primitive components from heroui
@TablePlus Not sure if this idea crosded your mind, but seeing the quality of the apps you make, I'd be among the first to buy a course if you're ever to provide one on how to make native apps 😄
Question to my favorite database gurus on Twitter @aarondfrancis and @tobias_petry 😂
Is there a reason a count(*) might have poor performance on a table with hundreds of millions of rows?
The optimizer does indeed select an optimal index, yet it takes 20 minutes on 500M rows 🤔
@aarondfrancis@tobias_petry That it is using an index, not the primary key but that shouldn’t matter, and the filter is at 100. Running on RDS with 4 vcpus and 32gig of ram.
@tobias_petry@aarondfrancis But somehow I still can’t make sense of these numbers, on a beefy instance, it still takes like 20-30min for a count. Or maybe it is normal and in that case it might make me reconsider mysql as an engine 😬
@arukomp@aarondfrancis@tobias_petry Yes I think this is what I’m trying to evaluate, if MySQL is the right engine for this kind of volume. I surely have considered Clickhouse but was hoping to spare myself the migration
@TablePlus maybe a quick create button? like I happen to create multiple servers (I know I can use different databases on the server, but hey I'm lazy)
I'm imagining it as a button to quickly create a MySQL server on a random available port. Extra points if we can create our own presets.
🚀 At long last, Unfinalize has been released! No longer will you be restricted by those who wish to limit you. They may impose "final" keywords in their packages, but you shalt not fret, and you shalt not waver, as we will win this war. For FREEDOM! ⚔️
https://t.co/hMuexLwHMh
@ste_bau There’s a lot of arguments in favor, and I respect those, however, a major caveat/deal breaker, the use of final is making the assumption that you know better than everyone else, and that you’ve anticipated all of the different use cases, which might not be the case always 🤷♂️
@ste_bau This is spicy 🌶️ and I love it
I have a policy of always open, want to extend something and make it your own? Sure go ahead, I reserve myself the right to not help you if you open a new issue, but at least I give you the freedom to do what you want 💁♂️
@michaeldyrynda@enunomaduro I'm curious to see how you log to Datadog as we used to use it too. So beside using JsonFormatter, what kind of transport do you use? and do you do some kind of parsing in a pipeline? 🤔