Jean Boussier

After several months of work, we finally enabled Pitchfork reforking on 100% of Shopify's monolith. ~30% reduction in memory usage, ~9% better latency and more! railsatscale.com/2023-10-23-pit…

@jdxcode @nateberkopec For the record the first name I reached to was forklift, but it was already taken.

Eh it was the other way around… I loved the name so much I was hoping the conflict with a ruby library wouldn’t conflict too much. I dont think I would make the same decision today, on the other hand it’s really hard to find a purely unique name anymore. mise wasn’t unique either

For those of you confused about pitchfork - basically the evil Byroot stole the name from Jeff's daemon manager ;) github.com/endevco/pitchf…

@matheusrich A Rails app.

@_byroot Is this Rails?

This one feels good. No more stupid comments at the top of files. At least in one repo.

@paracycle @Dagnan @vipulnsward To be fair to this article, the very first version of "chilled strings" did return true when `frozen?` was called on it. It was changed a little bit after that post was published.

@Dagnan @_byroot @vipulnsward the title (and some of the content) of that post is quite misleading, as you can see above. I understand the post is trying to say that on 3.4+ strings "act-like" they're frozen by default, but that's not quite true. Is there a way to have the post updated?

@Dagnan It's not, though. It's still opt-in even in 4.0.

@_byroot Didn't realize this was the default since 3.4!

@marcoroth_ About the boot performance, as discussed in Hakodate, I'd be interested to look in what it would take to support precompiling templates with Bootsnap, making parsing performance less important.

Really exciting to see Herb being adopted at GitHub! 🌿 Over the past months, the GitHub team has been running Herb across one of the largest Rails codebases out there where Herb helped uncover real issues missed by existing tooling while also providing invaluable feedback. 🙌

I wrote answer letter: github.com/ruby/rubygems/…

@samsaffron And right now we have no way to specify that in the gemspec, all we can do is raise in your face during boot, which isn't ideal.

@samsaffron Well, the proposed `ban` directive does answer that. But while I didn't list it here, optional dependencies are indeed a long requested feature. E.g. in Active Support, we don't force you to install `dalli` or `redis`, but if you do it has to satisfy a certain minimum version.

I wrote some more as a way to fight jet lag, hopefully it's still readable 😅: byroot.github.io/ruby/bundler/2…

@owaiswiz @fxn P75 CI time is roughly 6 minutes. Simple changes that don’t trigger a full build run in about 2 minutes

@_byroot @fxn Great article. Curious what the CI times at intercom look like now and the size of the suite? Ours is around 18mins for ~3k tests. We paralleize across 80 workers, not 1000+ but already have hit diminishing returns, but i always wonder how bad is 18mins relative to others

With the new job, I didn't get much time or energy to blog lately, but I figured I'd use the various downtime while traveling to Ruby Kaigi to write about some of the things I worked on recently: byroot.github.io/ruby/performan…

@nateberkopec It is! Very conveniently located too.

@_byroot it's a nice hotel!

Happy to share that today I'm starting as senior principal at Intercom!

@morgoth85 No downsides here f that works for your app.

@_byroot I guess now we can remove the comment and put RUBYOPT="--enable-frozen-string-literal" on app docker image build and when running tests. This of course assumes all the dependencies are freeze-safe. Are there any downsides to this approach?

I was recently reminded that not everyone fully understand what the frozen string literal magic comment is about. So I figured it was the occasion for another deep dive. byroot.github.io/ruby/performan…

@hakunin Yes, it's the best approach.

@_byroot Great writeup. When you do want to keep literals mutable in a way that's future-proof, would you call +"string" a good approach? E.g. #L26-L27" target="_blank" rel="nofollow noopener">github.com/maxim/portraya…

I tried to explain why I don't believe the recent accusations toward my former teammates, as well as how the Ruby and Rails Infra team at Shopify operates and why it can be trusted. byroot.github.io/opensource/rub…

Honored to be back on stage at #RailsWorld in Amsterdam next week to help guide a conversation with @tenderlove, @_byroot, and @hsbt. Still a few days before I lock in the topics. What questions would you love to hear them answer?

@seuros @maciejmensfeld #issuecomment-149828828" target="_blank" rel="nofollow noopener">github.com/rails/rails/pu…

@seuros @maciejmensfeld It very much was a mistake and was never necessary. If you follow the PRs you'll see.

🚀Sometimes the best code is no code! @_byroot just removed a 10-year-old mistaken call to Unicode.tidy_bytes from ERB::Util and got a massive 3.89x performance boost in #Rails 📈👏👏 github.com/rails/rails/pu… Follow me for more performance insights and Rails deep dives!