OJ Kwon

Same as https://github.com/swc-project/swc/issues/5654#issuecomment-1229584285 . Supporting dynamic (not specific to js) configuration is something core team wanted to have in a long term definitely, but we also would like to avoid...

Related: https://github.com/swc-project/swc/issues/5535 (Not the same issue, but related with error handling) I agree hard panic is not something desirable as there's no way consumers to workaround / ignore failures.

@syrusakbary I was few days faster https://github.com/swc-project/swc/pull/5456 😅 However, we are having some regressions in current main branch so until we figure it out we can't really upgrade unfortunately.

@syrusakbary We tried wasmer@3 and reverted it via #5521. To isolated upgrade, we created new `@swc/core` release contains 2 changes - one for rkyv bump (https://github.com/swc-project/swc/pull/5507) and another for wasmer...

Unfortunately there's no easy steps to reproduce, you have to generate specific host binary with wasm. I'm afraid it's not easy to create reproducible, isolated codebase as well.

I am aware isolated repro case is key to debug, howeve this is very convoluted and we (me and kdy) both failed to correctly dig further to create those. The...

This is inherited behavior from legacy `@swc/cli`. Not saying I am in favor of this or something, just saying this was introduced to ensure feature parity to existing cli. We...

Yes, there are some behavioral differences due to how rust detects stdin. Just saying why the current code works in this way. In any case, I agree we'd like to...

I think upstream will eventually pick 3, but I can't promise when, unfortunately. For the design I wouldn't expect we'll take too long though. Probably I'll accept some breaking changes,...

In my 5c, eventually we'll need some build systems like bazel. I would recommend bazel if it works seamlessly with cargo-based ecosystem, but in my understanding it's not yet. Or...