• Lena ( lena@gregtech.eu ) 
        link
        fedilink
        English
        arrow-up
        32
        ·
        2 years ago

        I agree that Chrome fucking sucks, but it’s disingenuous to call it unoptimized. Chrome and chromium-based browsers are as fast as or faster than Firefox. Although I agree that manifest V3 is horrible to the web as a whole and shouldn’t have been created.

        • Zagorath ( Zagorath@aussie.zone ) 
          link
          fedilink
          arrow-up
          1
          ·
          2 years ago

          It got added because it worked extremely well on browsers that implemented it, and it solved a problem that was needed on the site in question, which was very difficult to solve otherwise. I can’t blame a site for using an open standard that works for a majority of its users and which makes the development effort significantly less.

      • MonkderVierte ( MonkderVierte@lemmy.ml ) 
        link
        fedilink
        arrow-up
        7
        ·
        edit-2
        2 years ago

        This was due to Firefox at the time not implementing web standards that Chrome did.

        Uhm, yeah, that’s what browsers do. There are somewhere about 150 web standards and some are hard requirement while others are soft. Blink has some implemented that Webkit hasn’t but Gecko has and that’s true for all three. Same for browsers.

        Btw, the one with the most implemented standards is QtWebkit by far. It’s still slower tho.

  • Xylight ( Xylight@lemm.ee ) 
    link
    fedilink
    arrow-up
    23
    ·
    2 years ago

    When developing photon I always end up with more issues on chrome browsers than firefox. and half of those are because of its god awful scrollbar. Please use an overlay scrollbar instead of shifting the stupid page around, chrome.

    • I agree with you that failing to support multiple browsers is an old problem, but I think the cause has shifted.

      Back in the last century, supporting both browsers amounted to sniffing the browser and implementing the same feature twice. document.layers vs document.all for example.

      Nowadays I think the problem is different: we just don’t know what’s going on. The site is transpiled from TypeScript, written on top of React or Vue which drastically switches paradigm (bonus for Tailwind), packed with building tools, and the average dev has little understanding of what actually comes out. It’s a tall stack of leaky abstractions on top of the already tall one of the web. The dev is pretty sure it works on Chrome so they say it does work there, but it was not even a deliberate choice.

      • sunbeam60 ( sunbeam60@lemmy.one ) 
        link
        fedilink
        arrow-up
        6
        ·
        2 years ago

        For most sites it’s a testing matrix issue. Most testing teams look at browser stats and choose how to apply their limited resources based on that. So the dev probably doesn’t even see the bug that exists for an old Firefox version as there’s no testing done on it.

    • Is that http2? Cause http2 allows for reuse of a connection for additional requests.

      This caught me out with envoy reverse proxy doing a few subdomains using a wildcard cert.
      The browser would reuse the connection cause the cert authority and IP was the same, but envoy couldn’t figure out how to route the request correctly. Absolute head scratcher!