[Development] Qt 5.9's new garbage collector documentation? + root_ptr

[Thiago Macieira]

On sexta-feira, 14 de julho de 2017 10:08:13 PDT Phil Bouchard wrote:
> > Please read up a little on a subject before you make such an outlandish
> > suggestion.
> 
> Yes sorry that was just a quick guess but in Linux you can run Linux
> containers which do the same but with minimal overhead.

A container is not the same as a virtual machine. There is such a thing as 
virtualised containers (see [1]), but they don't boot as fast as a regular 
container. We've got it down to less than 200 milliseconds.

Browsers already use process separation and other things to avoid escalation 
from JS. I do believe they also use the same technique as containers for 
namespace separations. But that's for running a JS VM, not for running direct 
native code that can make system calls and trigger hardware errata.

[1] https://clearlinux.org/features/intel%C2%AE-clear-containers

> > And second, standardising on one API (OpenGL ES) has benefits for game
> > makers and other application deveopers, since they can use the very same
> > on both desktop and mobile. It's probably of less value for gamek
> > developers since the interaction is going to be very different from a
> > desktop and a mobile phone or tablet, but other applications can benefit
> > a lot from OpenGL ES.
> 
> I definitely agree OpenGL ES needs to be standardized.

It has always been a standard! That's what the Khronos Group is for in the 
first place.

> Anyway I'm just trying to help here. But Microsoft probably already
> embraced and is extending this strategy as we speak now ;)

No, they went their own route and designed DirectX, though TBH Khronos was 
dysfunctional at the time. WebGL is actually the best thing that happened to 
the standard route because it forced Microsoft to implement OpenGL ES support 
in their browsers in the form of WebGL.

-- 
Thiago Macieira - thiago.macieira (AT) intel.com
  Software Architect - Intel Open Source Technology Center