Re: GNUstep Projects and Cambridge

[Dr. H. Nikolaus Schaller]

Hi David,

Am 25.06.2013 um 10:39 schrieb David Chisnall:

> On 25 Jun 2013, at 03:21, "Lundberg, Johannes" 
> <johannes@brilliantservice.co.jp> wrote:
> 
>> Hi
>> 
>> This sounds exciting but what does it really mean, put Objective-C in the 
>> FreeBSD kernel? 
> 
> We plan on defining a slightly restricted subset of Objective-C (although I'm 
> not sure exactly what we'll choose to sacrifice - maybe not much), and a 
> runtime that better meets the constraints required for the kernel, then 
> modify clang and LanguageKit to emit code compatible with this ABI.
> 
>> What is the goal?
> 
> To investigate the costs of putting an introspective high-level language in 
> the kernel, which has traditionally been solely the domain of low-level 
> languages.  We'd also like to explore bridging the Objective-C object model 
> with the ad-hoc C-based KObj object model that the rest of the kernel uses 
> and see whether it's possible to seamlessly integrate the two. There's a 
> reasonable amount of code in the kernel that is not CPU-bound.  I'd also like 
> to explore using message sending at the language level as an alternative way 
> for userspace code to interact with the kernel.  If it all works, I'd like to 
> be able to use something like the demo that I gave at FOSDEM to have an 
> environment where I can do live editing on Smalltalk code running in the 
> kernel.

I find this interesting from a theoretical / structureal point of view, but why 
do you want to run any code in the kernel space? Do you expect a performance 
benefit or more security (I would expect less)? Or would any kernel component 
become easier to develop or test?

BR,
Nikolaus