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

[Phil Bouchard]

On 07/15/2017 11:35 PM, Phil Bouchard wrote:
> On 07/15/2017 01:18 PM, Phil Bouchard wrote:
>> On 07/15/2017 12:59 PM, Thiago Macieira wrote:
>>> On sábado, 15 de julho de 2017 09:39:20 PDT Phil Bouchard wrote:
>>>> Yes of course, I should have anticipated that. So one option left would
>>>> be to:
>>>> - compile the Javascript file for each architecture / platform
>>>> - link that "jex" to a portable dynamic library API
>>>> - run native containers (Hyper-V on Windows I think)
>>>> - run the "jex" executable in that native container
>>>
>>> Hyper-V is not a container.
>>>
>>> Containers are not completely secure. That's why we spent the
>>> engineering
>>> effort of making the Clear Containers, so that the processor's virtual
>>> machine
>>> protections kick in.
>>>
>>> And you could investigate NaCl.
>>
>> I will but you get the idea.
>
> Like I was saying, there was a bug in my code and perhaps it's
> impossible for Javascript to run without a GC indeed.
>
> But that doesn't mean that we couldn't run executables in one of those
> specialized containers. And although I like the concept of these quick
> function callbacks in Javascript, I do not like the 'escape analysis'
> thing and I think this should be trashed because the ratio costs /
> benefits isn't profitable enough.
>
> Perhaps we could derive a new language from Javascript and C++ which
> could be run inside these specialized containers. I already have a
> parser and only minor changes are necessary. If people want speed then
> there is no other way.

And if I remove that run-time stack thing then the generated code will 
be even more efficient.


-Phil