jahboater wrote: ↑
Sun Oct 25, 2020 9:33 am
Python has its 2/3 problems, and my last experience of Rust was that the repository version did not work and the nightly build had to be used. Once a strong standard appears (that everyone adheres to), stability improves and progress slows (except for C++ perhaps).
I'm curious to know what that was about.
As far as I can tell this has held true. Apart from obscure cases of code that depended on a compiler bug that was subsequently fixed. Rather like code that depends on UB in C and then finding that what the compiler does with that UB changes one day.
Presumably you were trying to build some code written to a nightly build, basically a beta of a future release, which used features not yet available in the stable release.
In short, unlike Python and more like C/C++, Rust will not suddenly stop compiling your old code.
What I wonder about the idea of an ISO standard or whatever for Rust is: If Rust has open governance and an RFC process in place and so on, then why would it need to get ISO to rubber stamp it? What benefit would it have apart from getting a prestigious standards body name attached to it and the privilege of then having to pay to get a copy of the standard
No, what Rust needs is a formal specification, so that anyone could build a conformant compiler from it. Which folks are working on. Rust needs a "Rust Foundation" preferably based outside of the USA.