rowan VS logos

The probably most well thought out approach for handling complicated mutable trees (with parent-references) in Rust is Rowan with red/green nodes as popularized by the Roslyn project, though its particular implementation has a few rough edges (doesn't use the type system to denote whether a node is mutable or not). The core idea is to have "green nodes" that form an ordinary immutable tree to store actual data, and on-demand "red nodes" that serve as a kind of cursor and offer a richer view, for example tracking parent–child relationships as the tree is traversed.

Make a Language by Luna Razzaghipour is a fantastic series. I especially like that she uses Rowan for building her syntax tree. While this makes your compilation step harder, you get to see how rust-analyzer does syntax trees, which I think is great.

Since you mentioned IDE tooling, I think it would be useful to have a CST representation and associated conversion from tree-sitter. [rowan](https://github.com/rust-analyzer/rowan) would be a good representation for the CST.

This is mighty impressive! I've been trying to get some motivation for the mythical rewrite of the proc macro in Logos, and this might just do it for me :D. I'll have a proper look later today and see if any of your findings have something that can be generalized. Also really surprised to see aarch64 doing better than x86_64 since the latter is what I've been optimizing for!

Rust is a very nice langage for implementing compilers, and has a nice ecosystem for it (logos, rust-peg, lalrpop, astmaker -- this one is mine --, etc...).

rust-langdev has a lot of libraries for building compilers in Rust. Perhaps you could use these to make your implementation easier, and revisit it later if you want to build things from scratch. I'd suggest logos for lexing, LALRPOP / chumsky for parsing, and rust-gc for garbage collection.

Thanks! For compile times you might find the CLI version that Andrew Hickman contributed useful, it's undocumented still mostly I fear but shouldn't be hard to use, see original PR: https://github.com/maciejhirsz/logos/pull/248

For my lexer generation purposes, I tend to use https://github.com/maciejhirsz/logos, as it not only generates an easy to use lazy lexer, but the result is also exceptionally fast!

Hi, I'm using rowan to create a parser and want to print more useful error messages with position in the text/file. I'm using logos (https://crates.io/crates/logos) to generate the lexer. Is there a way to get the starting and ending positions of a SyntaxToken? If not I thought of adding my own wrapper struct around the SyntaxTokens.

Is there a way for a lexer created with the logos crate (https://crates.io/crates/logos) to get the starting and ending positions for the tokens?

I wouldn't recommend Logos at this point. This recent bug is quite nasty and seems easy to hit, and the maintainer is unresponsive. Last commit was half a year ago. At this point I consider Logos abandonware, though it would be great if its development continued, or if it were forked.

logos uses "ridiculously fast".