Hi Madhu,
nice to see a lifesign of you! :-)
On 11.06.2016 05:02, Madhu Macaque Labs wrote:
I am glad to see Rust support being added to genode. How far is this support expected to proceed ? Is a complete Rust based genode being contemplated ? Norman and I had a brief discussion on this a while ago. After that we have started developing a simple rust based microkernel to figure out Rust's suitability as an OS implementation language. I know it can be done but nothing like working code to validate all assumptions ! Going forward we would prefer converting genode to rust rather than doing our own Mk. Converting some components to rust as shown in. 16.05 is the right way to go but was just wondering about the end goal.
The addition of Rust support should not be mistaken as a strategic move. It was suggested and contributed by Waylon Cude, not by Genode Labs.
It goes without saying that I welcome ways to eliminate classes of bugs (i.e., memory-safety issues) by design. On the other hand, I dislike throwing out the baby with the bath water and dismissing C++ altogether. The new Genode API (introduced in Genode 16.05) and the programming style that comes with it shows pretty well how C++ can be leveraged to attain robust components.
In the long term, I wish to guarantee the absence of bugs in the security/safety-critical parts of Genode but I cannot predict the means to get there yet (Formal verification? Use of higher-level languages?). The added Rust support gives us the chance to wet our toes with one possible direction.
Cheers Norman