Nice post that helps clarify one of the confusing in-flux aspects of Nim. The transition from mark-sweep to ref-counting may not affect most uses of Nim, but if you’re doing multi-threading, or bindings to other languages, it makes a big difference. I’m really excited about ARC/ORC; it’s not just a change of GC algorithm, but as the article describes, it brings in a lot of modern performance features like move semantics and lifetime detection.
I’m glad to see that the 1.4 release is supposed to be out RSN — that one is promised to provide stable ARC/ORC so we can focus on the future (and the docs can be updated to be consistent about the GC features.) Once that’s out I’ll feel more comfortable evangelizing Nim to everyone I know :) I’ve been using C++ for ages, and I know Rust is the new hotness here & elsewhere, but honestly I feel so much more comfortable & productive with Nim.
Nice post that helps clarify one of the confusing in-flux aspects of Nim. The transition from mark-sweep to ref-counting may not affect most uses of Nim, but if you’re doing multi-threading, or bindings to other languages, it makes a big difference. I’m really excited about ARC/ORC; it’s not just a change of GC algorithm, but as the article describes, it brings in a lot of modern performance features like move semantics and lifetime detection.
I’m glad to see that the 1.4 release is supposed to be out RSN — that one is promised to provide stable ARC/ORC so we can focus on the future (and the docs can be updated to be consistent about the GC features.) Once that’s out I’ll feel more comfortable evangelizing Nim to everyone I know :) I’ve been using C++ for ages, and I know Rust is the new hotness here & elsewhere, but honestly I feel so much more comfortable & productive with Nim.