Ben Sheldon
@bensheldon.bsky.social
I help small teams do big things and help big teams have big impact. And I build GoodJob, Day of the Shirt, Panlexicon, and Brompt too.
Reposted by Ben Sheldon
If you know of any open-source program using Ractors, let me know, I'd be interested to try it with ractor-shim.
October 30, 2025 at 9:01 PM
If you know of any open-source program using Ractors, let me know, I'd be interested to try it with ractor-shim.
How short lived to make passing it to a ractor be worth it? That's the like real examples I'm looking for.
October 30, 2025 at 9:56 PM
How short lived to make passing it to a ractor be worth it? That's the like real examples I'm looking for.
Anything you see moving to Ractors right now? I can't think of any large manipulations that fit into the constraints in the context of the products I work on. Like maybe a ViewComponent but getting the data into a state to ractorize seems exhausting. Or how to invert flow of child queues on parent.
October 30, 2025 at 6:51 PM
Anything you see moving to Ractors right now? I can't think of any large manipulations that fit into the constraints in the context of the products I work on. Like maybe a ViewComponent but getting the data into a state to ractorize seems exhausting. Or how to invert flow of child queues on parent.
1.1 I think it was ratomic gem that surprised me: github.com/mperham/rato...
1.2 my real question I guess was: what's changed in Ruby 3.5 that makes this possible? (Or easier)
1.2 my real question I guess was: what's changed in Ruby 3.5 that makes this possible? (Or easier)
GitHub - mperham/ratomic: Ractor-safe mutable data structures for Ruby
Ractor-safe mutable data structures for Ruby. Contribute to mperham/ratomic development by creating an account on GitHub.
github.com
October 30, 2025 at 2:22 PM
1.1 I think it was ratomic gem that surprised me: github.com/mperham/rato...
1.2 my real question I guess was: what's changed in Ruby 3.5 that makes this possible? (Or easier)
1.2 my real question I guess was: what's changed in Ruby 3.5 that makes this possible? (Or easier)
Ooh! Questions:
- the last time we looked at this Ractors required a c extension to work. Did that all get into Ruby 3.5?
- Say I already have calculated fibonacci.... what else might I use this for?
- the last time we looked at this Ractors required a c extension to work. Did that all get into Ruby 3.5?
- Say I already have calculated fibonacci.... what else might I use this for?
October 30, 2025 at 12:56 AM
Ooh! Questions:
- the last time we looked at this Ractors required a c extension to work. Did that all get into Ruby 3.5?
- Say I already have calculated fibonacci.... what else might I use this for?
- the last time we looked at this Ractors required a c extension to work. Did that all get into Ruby 3.5?
- Say I already have calculated fibonacci.... what else might I use this for?
that's good stuff! post more bangers as you find them.
October 25, 2025 at 6:17 PM
that's good stuff! post more bangers as you find them.
Yep. to_i on a AS::Duration does the same math you did. Just a little more wordy
October 22, 2025 at 12:52 AM
Yep. to_i on a AS::Duration does the same math you did. Just a little more wordy
You could do: + 7.days.to_i
Also TIL. Also, wouldn't you want it to end at the same local time? (Like is the test assertion itself wrong)
Also TIL. Also, wouldn't you want it to end at the same local time? (Like is the test assertion itself wrong)
October 21, 2025 at 8:34 PM
You could do: + 7.days.to_i
Also TIL. Also, wouldn't you want it to end at the same local time? (Like is the test assertion itself wrong)
Also TIL. Also, wouldn't you want it to end at the same local time? (Like is the test assertion itself wrong)
There is no Microsoft stack. MS does not care what GitHub's stack is. A few decisions have influence (Azure dropping their Ruby SDK) but GH can trivially work around that.
imo, GitHub is same as Jean described about Shopify: the eng leaders they have (ic & mgt) simply want to do something else
imo, GitHub is same as Jean described about Shopify: the eng leaders they have (ic & mgt) simply want to do something else
October 20, 2025 at 1:52 PM
There is no Microsoft stack. MS does not care what GitHub's stack is. A few decisions have influence (Azure dropping their Ruby SDK) but GH can trivially work around that.
imo, GitHub is same as Jean described about Shopify: the eng leaders they have (ic & mgt) simply want to do something else
imo, GitHub is same as Jean described about Shopify: the eng leaders they have (ic & mgt) simply want to do something else
... and obvs given this thread that didn't happen 🫠
Though now that's all in my rearview and I feel ok, given what I knew at the time and my skills and abilities and so forth.
Though now that's all in my rearview and I feel ok, given what I knew at the time and my skills and abilities and so forth.
October 20, 2025 at 4:01 AM
... and obvs given this thread that didn't happen 🫠
Though now that's all in my rearview and I feel ok, given what I knew at the time and my skills and abilities and so forth.
Though now that's all in my rearview and I feel ok, given what I knew at the time and my skills and abilities and so forth.
💛
And a majority of GitHub's engineers work in Ruby; the largest language cohort internally. And my intention (and sorta my role, though with downleveling who knows) was to turn that into a cohesive *program* (both in concrete resources but also as a conceptual narrative) across and outside the co
And a majority of GitHub's engineers work in Ruby; the largest language cohort internally. And my intention (and sorta my role, though with downleveling who knows) was to turn that into a cohesive *program* (both in concrete resources but also as a conceptual narrative) across and outside the co
October 20, 2025 at 3:59 AM
💛
And a majority of GitHub's engineers work in Ruby; the largest language cohort internally. And my intention (and sorta my role, though with downleveling who knows) was to turn that into a cohesive *program* (both in concrete resources but also as a conceptual narrative) across and outside the co
And a majority of GitHub's engineers work in Ruby; the largest language cohort internally. And my intention (and sorta my role, though with downleveling who knows) was to turn that into a cohesive *program* (both in concrete resources but also as a conceptual narrative) across and outside the co
Which isn't to be defensive about it. But to explain the perspective from inside the company.
It's not *just* a communications problem, but I think examples of what it looks like to be contributing to Open Source internally were really hard to get written and published through the company blog.
It's not *just* a communications problem, but I think examples of what it looks like to be contributing to Open Source internally were really hard to get written and published through the company blog.
October 20, 2025 at 1:18 AM
Which isn't to be defensive about it. But to explain the perspective from inside the company.
It's not *just* a communications problem, but I think examples of what it looks like to be contributing to Open Source internally were really hard to get written and published through the company blog.
It's not *just* a communications problem, but I think examples of what it looks like to be contributing to Open Source internally were really hard to get written and published through the company blog.
I think GitHub contributes a lot; they do conf and meetup sponsorships and employees do work upstream and there is a team that focuses on enabling that (Ruby Architecture, my former team).
And I'm ashamed by everything that I wasn't able to achieve. And publicizing it was its own challenge.
And I'm ashamed by everything that I wasn't able to achieve. And publicizing it was its own challenge.
October 20, 2025 at 1:12 AM
I think GitHub contributes a lot; they do conf and meetup sponsorships and employees do work upstream and there is a team that focuses on enabling that (Ruby Architecture, my former team).
And I'm ashamed by everything that I wasn't able to achieve. And publicizing it was its own challenge.
And I'm ashamed by everything that I wasn't able to achieve. And publicizing it was its own challenge.
accepts_nested_attributes_for. And probably some name/value buttons to differentiate the different actions in a controller. That's about it.
October 19, 2025 at 8:14 PM
accepts_nested_attributes_for. And probably some name/value buttons to differentiate the different actions in a controller. That's about it.
Good show? I'm going Thursday.
October 15, 2025 at 4:47 AM
Good show? I'm going Thursday.
A little bit of context in here: github.com/rails/rails/...
Whenever `helper` or `helper_method` is used, the controller ends up with its own custom set of classes for holding the custom set of helper methods.
Whenever `helper` or `helper_method` is used, the controller ends up with its own custom set of classes for holding the custom set of helper methods.
Avoid creating controller _helper modules until modification by jhawthorn · Pull Request #40204 · rails/rails
In applications which use :all helpers (the default), most controllers won't be making modifications to their _helpers module.
In CRuby this created many ICLASS objects which could cause a larg...
github.com
August 19, 2025 at 1:53 AM
A little bit of context in here: github.com/rails/rails/...
Whenever `helper` or `helper_method` is used, the controller ends up with its own custom set of classes for holding the custom set of helper methods.
Whenever `helper` or `helper_method` is used, the controller ends up with its own custom set of classes for holding the custom set of helper methods.
I finished it a few weeks ago. Really good! I'm almost done with Martha Well's entire catalog. So good.
February 1, 2025 at 6:25 AM
I finished it a few weeks ago. Really good! I'm almost done with Martha Well's entire catalog. So good.
Unfortunately not: github.com/hotwired/tur...
Feature request: Ability to target multiple frames · Issue #475 · hotwired/turbo
Currently Turbo frames is amazing at handling HTML updates back to the UI when specifying a single frame target to update on the page. In a lot of cases this is all you need however in more complex...
github.com
January 26, 2025 at 11:07 PM
Unfortunately not: github.com/hotwired/tur...
Ooh ooh. Maybe I'll go subscribe 😄
I've been looking for someone else doing a good job of this: github.com/hotwired/tur...
I've been looking for someone else doing a good job of this: github.com/hotwired/tur...
Replace two different turbo-frame from one response · Issue #56 · hotwired/turbo
Situation: There is a profile form where the user can update his own name. There is, in the same page, a navigation bar where the username is shown. When the user submit the profile form, the respo...
github.com
January 26, 2025 at 6:28 PM
Ooh ooh. Maybe I'll go subscribe 😄
I've been looking for someone else doing a good job of this: github.com/hotwired/tur...
I've been looking for someone else doing a good job of this: github.com/hotwired/tur...
Oooh! Are you using streams to do the Refresh All?
January 20, 2025 at 5:37 AM
Oooh! Are you using streams to do the Refresh All?