KDE Plasma 5.5 Beta Presents New Features

[Ansla]

Interestingly nobody asked for such an approach before Chrome did it that way.

Because nobody (talking about regular users on forums, right?) even knew it's possible?

Since then it's the glorified solution, because Google does it that way.

And it also works as advertised!

I personally think it's a very bad choice - especially for a web browser. Rendering a web site should not cause crashers, as that is always a potential remote execution vulnerability. By making only the tab crash, I think priorities are moved the wrong way: away from fixing crashers to oh at least the experience is good.

In theory that sounds fine. But lets seee what happens in practice with browser crashing bugs in a browser that doesn't have this feature, say konqueror: https://bugs.kde.org/show_bug.cgi?id=294707. That's a bug from February 2012 that still gets duplicates to this day. Why? Well, mostly because Qt 4 turned into pretty much abandonware even before Qt 5 was officially released.

So now we have a "comunity supported" QtWebkit 2.3 branch for Qt 4 while official releases continue to ship with a broken QtWebkit. And if a distribution ships a standard Qt 4 then KDE SC 4 on that distribution will be a crashfest over there and users will blame you.

Why are you so sure that a bug in Qt 5 will not make plasmoids randomly crash and it will remain unresolved for the lifetime of Plasma 5? And what's so bad about taking measures to minimize the impact of such a bug? Yes, it's a workaround, but it's the best you can do as long as you don't maintain the entire code base used by all plasmoids in the world.

[mgraesslin]

Why are you so sure that a bug in Qt 5 will not make plasmoids randomly crash and it will remain unresolved for the lifetime of Plasma 5? And what's so bad about taking measures to minimize the impact of such a bug? Yes, it's a workaround, but it's the best you can do as long as you don't maintain the entire code base used by all plasmoids in the world.

Oh I cannot be sure that such crashers get fixed. But I don't think that moving plasmoids into dedicated processes would improve the quality. Because the base scene would still be rendered by QML. Now if QML has a crasher which makes the plasmoid crash, why should the base desktop not be affected by it? If such a crash exists we have not won anything by introducing a complicated architecture. And of course the time we would need to implement such an architecture we could also spent on fixing crashers in Qt (which we do).

[Ansla]

Now if QML has a crasher which makes the plasmoid crash, why should the base desktop not be affected by it?

Because there will always be a plasmoid out there that uses an obscure feature that nothing else uses.