Re: [GNUnet-developers] [GSoC] Question on "Rust implementation of GNUnet utils" project

[Jeff Burdges]

There is a project by Andrew Cann <address@hidden> on github here : 
https://github.com/canndrew/gnunet-rs

It's referred to as "GNUnet bindings", but it's really a
reimplementation of GNUnet utils, so that Rust code can speak some
existing GNUnet IPCs.  Just fyi, one cannot make literal "bindings"
because GNUnet code is too unsafe, well it's not even thread safe.

There are a couple things one might work on here : 
1.  Implement more existing GNUnet IPCs in gnunet-rs.
2.  Switch gnunet-rs to asynchronous IO using mio either directly or
indirectly : https://github.com/canndrew/gnunet-rs/issues/1

I'd kinda recommend doing 2 first and then doing 1 if there is time, but
if you're more interested in learning about GNUnet itself, then maybe
just 1 from the beginning. 

Have you used asynchronous IO before?  What sort of model?  
After hacking around inside the GNUnet utils scheduler, I developed a
slight prejudice against basic callbacks, as used in GNUnet.  And
towards something with more structure, like maybe the ADT state machines
of rotor : https://github.com/tailhook/rotor
Do any of rotor, gj, eventual_io, etc. sound/look like something you've
used before?  I'll have a conversation with Christian this week about
what GNUnet protocols look like to try to get a sense if one of these
approaches is more amenable. 

Jeff

p.s.  There are cryptographic bindings in GNUnet utils, but that appears
unnecessary.  I'd envision individual projects that using gnunet-rs
using an orthogonal rust crypto library, either sodiumoxide,
rust-crypto, or ring, well the crypto would always be a project talking
to itself after all.




On Sat, 2016-03-19 at 15:43 +0100, kc wrote:
> Dear GNUnet developers,
> 
> I'm a MSc student currently studying in TU Delft, specialising in
> cryptography and security. GNUnet aligns closely with my interests so
> I'm very excited about this opportunity with GSoC. 
> 
> I'm interested and would like to know about the "Rust implementation of
> GNUnet utils" project. The description implied that there are existing
> Rust implementations, but I couldn't find any *.rs files in your SVN
> repo.
> 
> Suppose there does not exist any Rust code yet, which utils would have
> the highest priority, i.e. that would benefit the most from asyc IO?
> 
> Kind regards,
> Kelong
> 
> _______________________________________________
> GNUnet-developers mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/gnunet-developers

signature.asc
Description: This is a digitally signed message part