Amen.
500 contenus te permettront d'être top ranké, mais vraiment, un site qui charge en 12 secondes, avec du renderblocking à en bé-gerr, ca me va pas.
The new renderBlocking Flag in Chrome made this possible -> WebPageTest now indicates whether or not a resource is blocking in its waterfalls glennga.be/3Btvsxo via @tkadlec
WebPageTest now uses Chrome's new `renderBlocking` indicator to signal if a resource blocks display of the page, making it easier to zero in on one of the most common perf bottlenecks.
Wrote a post about the new indicator and what we're doing with it. blog.webpagetest.org/posts/n…
ALT Screenshot of part of a WebPageTest waterfall, showing an orange icon indicating requests are render blocking.
At a certain point it might be easier to have a new option to make them (on purpose) renderblocking. In case you selfhost and preload = why bother with the race conditions and just come with 1 unified branded look. Yes maybe a bit delayed, but no CLS. No annoying swaps... #UX
Results on Lighthouse with simulated 3G and simulated CPU slowdown: Blank page: .8s, blank page with image: .8s, blank page with jquery: 1.7s
It's not about the extra kilobytes, it's about the fact that the extra kb are renderblocking JS.
I guess no more messing around with webfontloader, fontfaceobserver for eliminating FOIT and renderblocking nature of fonts.
Hope @Typekit follows the same....
@zachleat@bram_stein
📢 Announcing Google Fonts will soon support setting `font-display` via a new query parameter. This unlocks controlling web font loading without having to self-host! Huge news #io19
Question for @BrendanEich and @brave: How do you handle slow TCP/IP timeouts when blocking bad scripts on sites with renderblocking scripts? Wouldn't blocking those from loading make the site feel even slower?