I love VS code, but insisting that its parsing is even nearly as good as anything from Jetbrains is laughable. Working with Rust in VS code is no where near as good as IntelliJ for example, and their Javascript `intellisense` is best in class.
Agreed. Working with rust in e.g. CLion (or even the free PyCharm CE with Rust plug-in) beats vscode by a large margin, it's nowhere near comparable, especially when working with larger projects often requiring major refactoring etc.
Some people install RLS because they haven't heard about rust-analyzer. Anyway, debugging is out of scope, but try filing issues for any other problems you run into.
Having used IntelliJ for a long time I would agree in all cases except rust and flutter. Flutter IntelliJ is not bad but rust analyser in vs code is miles ahead of IntelliJ rust.
Have you tried using rust-analyzer with VS Code? I find it to be a little rough around the edges but mostly pretty great. I used the Rust plugin for IntelliJ a couple of years ago and remembered being disappointed.
They've historically done a lot of things on their own for JS. So perhaps VS Code + TS language server isn't quite on par with their JS tooling (which works a lot better with TypeScript type definitions, though, so perhaps it's just mostly the same as well).
However, what I personally hated was that there's so many squiggly underlined things ... in red, yellow, green and what other colors, each of which (regardless of color) could be either ignored or an actual error. Granted, that's how easy it is to properly check JS code for whether it does something sensible, but I didn't really find it helpful, especially because issue categorization and accuracy was so random.
> Granted, that's how easy it is to properly check JS code for whether it does something sensible
> especially because issue categorization and accuracy was so random
Stockholm syndrome is strong with this one. =)
the reason you didn't find the IDE suggestions helpful is exactly because of how difficult it is to reason about dynamic language code. the tooling is basically just guessing and expecting you to deal with the noise on the off chance it found something.
I've ended up deciding for myself to never write JS and instead using TypeScript. So far it's working well. I just couldn't really see all that random guessing by the IDE as an advantage. If the only way to figure out whether the code is broken is to run it, then it's pointless to guess potential problems on every single expression because dealing with those takes more time than just testing the code – or results in the developer completely ignoring all diagnostics.
