Hacker News new | past | comments | ask | show | jobs | submit login
The road ahead for Rust (pcwalton.blogspot.com)
54 points by icey on April 30, 2011 | hide | past | favorite | 6 comments



When the author talks about missing features, is he referring to the new self-hosting compiler only? I remember skimming thru the PDF manual when I first heard about Rust, and I'd find it alarming if stuff like deterministic garbage collection and stack growth have been documented but not actually implemented.


Garbage collection isn't implemented; we're relying on reference counting for now. The self-hosted compiler has no cyclic data structures, so we're fine.

Stack growth was implemented at one point, but we disabled it because the way it was implemented was a proof-of-concept hack (it just rewrote anything that looked like a pointer on the stack). I suspect we're going to switch to a more Go-like model, in which stack growth happens by calling through a stack-switching thunk.


What kind of of macros do you think about?


Too early to say - this is probably the one part of the language that has had the least design work - but I suspect they'll feel like Dylan's macros. That is, Lisp/Scheme-inspired, but tailored to a C-like syntax.


Ah very very cool. I was think about saying something about dylan (the PLOT language by Dave Moon is something too look at too). Dylan has a BASIC like syntax do you think this will be easy to adopt to C like syntax? Is the C syntax set?

I really like what I see with Rust. Keep up the good work.


I don't foresee many problems with adopting Dylan-like macros to the C syntax (they're both Algol-derived anyhow), but again this is all up in the air :)

The C syntax is pretty much set. Switching at this state would do more harm than good, and we think that in the long term syntax is less important than semantics.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: