Comment by grantwu | Hacker News Reader

Comment by grantwu | Hacker News Reader

> By far the most important aspect is that we need to keep the new codebase as compatible as possible, both in terms of semantics and in terms of code structure. We expect to maintain both codebases for quite some time going forward. Languages that allow for a structurally similar codebase offer a significant boon for anyone making code changes because we can easily port changes between the two codebases. In contrast, languages that require fundamental rethinking of memory management, mutation, data structuring, polymorphism, laziness, etc., might be a better fit for a ground-up rewrite, but we're undertaking this more as a port that maintains the existing behavior and critical optimizations we've built into the language. Idiomatic Go strongly resembles the existing coding patterns of the TypeScript codebase, which makes this porting effort much more tractable.

--https://github.com/microsoft/typescript-go/discussions/411

I haven't looked at the tsc codebase. I do currently use Golang at my job and have used TypeScript at a previous job several years ago.

I'm surprised to hear that idiomatic Golang resembles the existing coding patterns of the tsc codebase. I've never felt that idiomatic code in Golang resembled idiomatic code in TypeScript. Notably, sum types are commonly called out as something especially useful in writing compilers, and when I've wanted them in Golang I've struggled to replace them.

Is there something special about the existing tsc codebase, or does the statement about idiomatic Golang resembling the existing codebase something you could say about most TypeScript codebases?

> I'm surprised to hear that idiomatic Golang resembles the existing coding patterns of the tsc codebase. I've never felt that idiomatic code in Golang resembled idiomatic code in TypeScript.

To be fair, they didn't actually say that. What they said was that idiomatic Go resembles their existing patterns. I'd imagine what they mean by that is that a port from their existing patterns to Go is much closer to a mechanical 1:1 process than a port to Rust or C#. Rust is the obvious choice for a fully greenfield implementation, but reorganizing around idiomatic Rust patterns would be much harder for most programs that are not already written in a compatible style. e.g. For Rust programs, the precise ownership and transfer of memory needs to be modelled, whereas Go and JS are both GC'd and don't require this.

For a codebase that relies heavily on exception handling, I can imagine a 1:1 port would require more thought, but compilers generally need to have pretty good error recovery so I wouldn't be surprised if tsc has bespoke error handling patterns that defers error handling and passes around errors as values a lot; that would map pretty well to Go.

Most TypeScript projects are very far away from compiler code, so that this wouldn't resemble typical TypeScript isn't too surprising. Compilers written in Go also don't tend to resemble typical Go either, in fairness.

I'm not involved in this rewrite, but I made some minor contributions a few years ago.

TSC doesn't use many union types, it's mostly OOP-ish down-casting or chains of if-statements.

One reason for this is I think performance; most objects are tagged by bitsets in order to pack more info about the object without needing additional allocations. But TypeScript can't really (ergonomically) represent this in the type system, so that means you don't get any real useful unions.

A lot of the objects are also secretly mutable (for caching/performance) which can make precise union types not very useful, since they can be easily invalidated by those mutations.

In the embedded video they show some of the code side by side and it is just a ton of if statements.

https://youtu.be/pNlq-EVld70?si=UaFDVwhwyQZqkZrW&t=323

to be fair, there's not many ways to implement a token matcher.

though looking at that flood of loose ifs+returns, i kinda wish they used rust :)

I’d guess Rust compile times weren’t worth it if they weren’t going to be taking advantage of the type system in interesting ways.

Reply on news.ycombinator.com