Jacob "kurtextrem" Groß
Senior Performance Engineer @Framer · The INP optimization guy · webperf/UX/security · @kurtextrem on other platforms.
- Chrome 136+ ships a new V8 version that supports a "compile hint", which makes V8 parse and compile all functions in that file eagerly. Great for entry points – where you know all functions will very likely be used. Less so as 'one weird trick to make all your JS faster'.
- An easy way to figure out if this might make sense to add for you: Open Chrome DevTools -> 3 dots -> More tools -> Coverage. The files with the least amount of unused functions right after a page load, will gain the most from this.
- The "Performance" tab of DevTools shows compilation & parsing here - important: disable caching to avoid v8's byte cache! The compile hint changes how early the compilation happens. If the main thread stalls while waiting for compilation to finish, the compile hint can have positive impact.
- Blog post: v8.dev/blog/explici... Status page: chromestatus.com/feature/5100...
- [Not loaded yet]
- anytime! :)
- [Not loaded yet]
- Sorry to hear that, Barry. My thoughts are with you and your family
- [Not loaded yet]
-
View full threadI disagree with this take. See the library measurements from Ariakit. They don't involve any site navigation whatsoever. The important part is React offers utilities to have great UX & INP, but not enough people use it (yet). So it's the knowledge part that is missing in my opinion.
- That said, you can build SPAs with both.. but it doesn't really matter, because the takeaway *should* be you (as a react user) should pay attention to use more of what react offers to achieve good INP - and not "react is bad". A bold headline draws attention to that. Happy to hear feedback though.
- [Not loaded yet]
- I'll defer that to @tunetheweb.com, but personally I disagree with the data I have. Page navigations rarely have any bad INP unless you synchronously re-render the full page without fetching data. @danielroe.dev held a talk about how e.g. Nuxt suffered & fixed this.
- [Not loaded yet]
- sorry, I hope my comparison didn't trip anyone's feet (that wasn't the intention). My intention was to bring attention to INP & React. I also mention it's not only react, but also other frameworks that perform worse than the average
- Did you know that jQuery outperforms React, what? 😳 Sounds like clickbait, but is reality. The HTTP Archive shows 82% of sites that use jQuery have good INP on mobile📱, while only 67% do when they use React⚛️. Not just sites, also libraries. Let's fix that👇#webperf kurtextrem.de/posts/improv...
- [Not loaded yet]
- INP doesn't differentiate between CSR and SSR.
- Part 2 of my "How to improve INP" series is for library authors and devs working on React codebases. We'll take a look at how to: ✨ Ensure great INP & UX after user interactions 🚀 Make hydration concurrent 💡 Get non-blocking useEffects, always so that... your site outperforms jQuery instead.
- [Not loaded yet]
- yes: typescript-eslint.io/rules/return... with the option "in-try-catch"
- is this still true, even for script type=module? 🤯 cc: @chromium.social … can’t for the life of me find the message but thank you to the person at @perfnow.bsky.social who sent this my way ❤️ web.dev/learn/perfor...
-
View full threadYes I did. After writing the module, we received feedback from the V8 team (including that the doc you linked is outdated). The problem is that it fails silently so it doesn't show up in a trace. Tomorrow I can send you all the info I have as I'm not on my laptop. Cc @kurtextrem.de
- Tagging also @addyosmani.bsky.social and @mths.be, as e.g. I know some info in this article: v8.dev/blog/cost-of... is also outdated by now (the streaming thresholds for example). Do we have any public docs regarding script streaming that have info on how it works in Chromium & Node in 2024?
- Let's see! Happy to stand corrected (at least for more recent versions of chrome/V8, I suppose?)
- [Not loaded yet]
- [Not loaded yet]
- likewise! looking forward to chatting to you in the future :)
- that's an awesome idea
- "Everyone experiences P99" - @anniesullie.bsky.social ... at subway stations, or flaky internet. Understanding & monitoring that experience can make sense. blog.chromium.org/2021/04/digg... #PerfNow
- Closing talk day 2 @ #PerfNow: @anniesullie.bsky.social showing how much time improving CWVs saved. I like how Alex's great talk at the closing of day 1 was a wake-up call and now we get to see the positive side of things as contrast.
- Yesterday during my #perfnow talk, I mentioned the Web Performance Slack. If you’d like to join the Slack, use this link before Nov 25: join.slack.com/t/webperf…
- Damn it's been some time since we had an invite link 👀 we should have printed that on a QR
- [Not loaded yet]
- [Not loaded yet]
- Half of the WebPerf slack too 😄🤝🏼
- I joined Twitter when I was between 13-14, and now joining a new platform that is basically Twitter, but different, feels like really getting old. So far I like the feed though.
- [Not loaded yet]
- oh yeah, that's on point!
- I'll be speaking on the #PerfNow pre-conference meetup - performance.sync(). Topic: Collecting UX-impacting metrics your boss understands. Lightning talk, to spark thoughts!⚡ Looking forward to it: lu.ma/z1qt2a63
- [Not loaded yet]
- [Not loaded yet]
- what would be needed for react support?
- Cool stuff! I think Marvin also published this here: github.com/denoland/pre... so... both is useable for plain React, right?
