@samruby

https://nitter.cf/t.co/Vfbxo6yzUw Rails Specialist. Open Web Communities. he/him

 Raleigh, NC
ruby.social/@samruby
Joined March 2009
  • Tweets1.6K
  • Following9
  • Followers3.5K
39Media
Include

Only show posts containing:

Exclude

Hide posts containing:

Time range
-
Minimum likes
Seeing the same thing with Roundhouse. Also, I'm dogfooding Campfire/Roundhouse/Spinel: github.com/rubys/roundhouse/…

CI capacity: investigate queue delays and request a GitHub Actions concurrency increase · Issue...

Sam @rubys — could we confirm the account's effective GitHub Actions concurrency allowance and, if appropriate, request an increase? Queue delays are postponing validation of merged main, not j...

github.com
Yukihiro Matz
@yukihiro_matz
4h
Rails World 効果で Spinel への PR と issues がものすごい勢いで増えてる。がんばれ、Claude Code.
16
890
Sam Ruby retweeted
Nate Berkopec
@nateberkopec
6h
.@yukihiro_matz appears to have achieved hard takeoff. He's merged 200 PRs in the last 24 hours to spinel and 1000 this week
7
7
4
173
16,510
Replying to @Cufe_Haco
I remember the discussion, but I remember the details of your app differently, so perhaps I didn't follow. Specifically, I don't remember any out of process execution like you are describing here, which sounds interesting.
1
1
31
Not built yet, but that would let LLMs write Rails apps that define their own DSLs and still compile. Where a DSL falls outside that subset, the compiler could say why.
5
198
Roundhouse already expands one narrow kind of app-defined Rails macro at compile time. Generalizing that to any DSL whose inputs are known at compile time is part of what Writebook needs (it defines two).
chad fowler
@chadfowler
Sep 30
And they can ship with a compiler that gives the agents immediate actionable feedback.
1
21
1,532
Replying to @EstebanTrejosQ @nateberkopec
That was before the days of LLMs, so yes.
1
17
Replying to @sshconnection
All cool; I could have made that more explicit up front. There are others that advocate for leaving it up to LLMs, that's where I part ways. For me, I'm using Rails as the spec language, but I'm open to alternatives.
15
Replying to @alihfadel
Agents welcome: github.com/rubys/roundhouse/…

roundhouse/AGENTS.md at main · rubys/roundhouse

Rails as a specification; the deployment target is a build flag - rubys/roundhouse

github.com
6
484
Replying to @sshconnection
Sorry, yes compilers. Roundhouse is a deterministic compiler of Rails code. Pass it the same input and you get the same output. 1. Define spec language. 2. Build compiler for that language. 3. Author spec - either by human or LLM.
1
20
Replying to @sshconnection
question; what do you think completed do?
1
31
Replying to @BijanRahnema
no. I tell the agent what tests to pass
1
3
565
100% written by a human. Porting and on-shotting are anecdotes. Yes, Rust is faster than CRails. Duh. But is that really the interesting question? What happens when your needs change and you are faced with a requirement than you never anticipated before? intertwingly.net/blog/2026/0…
2
5
1
53
10,602
This is the right discussion to be having, though it is not an either or. Keep your Rails app, transpile it to the deployment target that meets your app's needs. Offline first, shared nothing, and other architectures are more important than runtimes.
Irina Nazarova
@inazarova
Sep 30
Tech is expanding as normal people build apps with agents - for their team at work, friends, family. The last thing they need is Rust. They value simplicity, security, batteries-included frameworks. They can occasionally read an algorithm or calculation to double check. They should pick idiomatic Ruby, batteries-included Rails, and plain SQLite.
1
1
29
2,660
Replying to @jdkornac
Sorry, but no. It actually is the AI generated drivel that is the interesting thing. Oddly, it reads OK to me, but apparently not to everyone. I welcome others to fill the gap that I'm leaving.
1
2
82
Replying to @timtilberg
Nokogiri is a wrapper over libxml, which is written in C, so you already have what you need. If your hot spot includes your code, you can compile your code with Spinel, write a small wrapper and call that from Ruby.
1
2
41
Replying to @th1agofm
It is a good idea to infer it. OCaml and Haskell have been doing it for decades. Why not Ruby and Rails?
1
32
Replying to @largo @nateberkopec
Made a fix. Should look better now.
1
1
51
Replying to @aashishkhullar1 @yukihiro_matz @dhh
rubys.github.io/roundhouse/b…
16
Replying to @timtilberg
You write your application in a Spinel subset of Ruby (no eval or method_missing) and the whole application (not just the hot paths) gets compiled, first to C, then to an executable.
1
2
129
Replying to @th1agofm
Why maintain the types when you can infer them? rubys.github.io/roundhouse/d…
1
9
405
Load more