The traits just had an overhaul , they seem to have a good handle on
how to handle OO vs type classes. They tried basic classes with type
classes but after a reasonable amount of code there was little code
using  them so they cut them and expanded traits .  Hence what you
probably cant do well  is just trivially port C++ code , though the
nice thing is you can call C functions  as part of the language ( not
sure about calling C++)  .

Its pretty hard to get the information as the documentation is a out
of date as its changing pretty fast and the standard libs are not
written for the new traits . Best info is the tutorial (about 6 months
out of date)  and  on the blogs of some of the authors.

http://pcwalton.github.com/blog/2012/08/08/a-gentle-introduction-to-traits-in-rust/
http://smallcultfollowing.com/babysteps/blog/2012/10/04/refining-traits-slash-impls/
http://pcwalton.github.com/

There is something to be said for a more mature product , tests ,
installers , multiplatform  , greater use etc ..  Rust seems to have a
full time team paid for by Mozilla and its active eg
https://github.com/mozilla/rust/graphs/contributors . That said i
liked the Bitc constraints on types better and the  natural types
though these could be added to Rust..Bitc syntax was more modern ,
rust has a c "style".  To me the killer is a JIT if you want a fast
JIT and cant live with cached compiles than i don't think Rust will do
the job as the type system makes it slow to compiler and LLVM doesnt
hel. Its worth noting that with a really good GC , common  rust code
may end up more similar to bitC using more shared pointers but  decent
GCs are rare for static languages outside of Java / .NET   and a
significant effort .

Anyway worth a look and id thought id let you know that its
progressing / changing. At the very least its worth getting updates on
there lessons / pain points  .

Ben

On Sun, Jan 6, 2013 at 9:09 AM, Jonathan S. Shapiro <[email protected]> wrote:
> I need to go look at Rust again in its current form, but based on what I
> know about it's foundations, I'm dubious. Given LLVM, and the fact that we'd
> be removing the hard parts from BitC, I'm not convinced about your time
> estimate.
>
> But as I say, I should go look.
>
_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev

Reply via email to