I've been testing web software that I work on with Servo and Ladybird. I'm pretty impressed!
Servo has some areas where it performs very well, though support for things like CSS Grid are still experimental and off by default (but working pretty well), and it's missing a few important APIs still. (I haven't checked in a while but last I looked it was missing AbortController. That's pretty unfortunate and probably breaks a lot of random stuff across the web.
I think Ladybird in Nixpkgs still doesn't have fully working hardware acceleration, so I'm probably not seeing the full smoothness Ladybird can offer in an optimal setup. Still! It runs pretty well, and it supports a vast majority of the web platform. Right now for my own development, the only major two omissions I've noticed are Origin Private File System, which is pretty new, and OffscreenCanvas, which actually is "implemented" but only as a stub. Throwing Web Workers and heavy WebAssembly workloads at Ladybird seems like no problem at all, and it renders a very reasonable subset of modern CSS almost perfectly.
(Tangent: Ladybird Web Workers are actually separate processes, which is kind of funny. If one runs out of control, you can literally SIGKILL it. The browser copes with this seemingly fine!)
Unfortunately, a lot of the web uses libraries and frameworks that are happy to eagerly adopt web standards and technologies that are not supported by browsers from a year or two ago. While I realize a lot of these standards were created specifically to solve problems that real developers have, I've been unimpressed watching this play out. Often new technologies increase complexity to solve relatively minor already-manageable issues with the web platform, and I feel like it is counterproductive to fatten up the web platform just for that. I mean really. There eventually has to be a point where we all agree that there is simply too much CSS technology and there is little justification to add more, but yet it just continues to grow uncontrollably. (I'm not saying it's all bad, though. Personally, I think the text-wrap additions were actually pretty nice things that really do need to be part of the layout engine.)
Servo has some areas where it performs very well, though support for things like CSS Grid are still experimental and off by default (but working pretty well), and it's missing a few important APIs still. (I haven't checked in a while but last I looked it was missing AbortController. That's pretty unfortunate and probably breaks a lot of random stuff across the web.
I think Ladybird in Nixpkgs still doesn't have fully working hardware acceleration, so I'm probably not seeing the full smoothness Ladybird can offer in an optimal setup. Still! It runs pretty well, and it supports a vast majority of the web platform. Right now for my own development, the only major two omissions I've noticed are Origin Private File System, which is pretty new, and OffscreenCanvas, which actually is "implemented" but only as a stub. Throwing Web Workers and heavy WebAssembly workloads at Ladybird seems like no problem at all, and it renders a very reasonable subset of modern CSS almost perfectly.
(Tangent: Ladybird Web Workers are actually separate processes, which is kind of funny. If one runs out of control, you can literally SIGKILL it. The browser copes with this seemingly fine!)
Unfortunately, a lot of the web uses libraries and frameworks that are happy to eagerly adopt web standards and technologies that are not supported by browsers from a year or two ago. While I realize a lot of these standards were created specifically to solve problems that real developers have, I've been unimpressed watching this play out. Often new technologies increase complexity to solve relatively minor already-manageable issues with the web platform, and I feel like it is counterproductive to fatten up the web platform just for that. I mean really. There eventually has to be a point where we all agree that there is simply too much CSS technology and there is little justification to add more, but yet it just continues to grow uncontrollably. (I'm not saying it's all bad, though. Personally, I think the text-wrap additions were actually pretty nice things that really do need to be part of the layout engine.)