Anyone else seeing the following get printed multiple times when anything happens with webpack in Ko...

# webassembly

[eygraber]

Anyone else seeing the following get printed multiple times when anything happens with webpack in Kotlin 2.0

Critical dependency: Accessing import.meta directly is unsupported (only property access or destructuring is supported)

[Robert Jaros]

Yes. Since the latest rc.

[Edoardo Luppi]

Maybe it's because of this change?
Probably it complains because of const importMeta = import.meta;
https://github.com/jetbrains/kotlin/commit/9df561a5c284aeba6a16ca00a8699511de5cfb04

[Robert Jaros]

For reference: https://kotlinlang.slack.com/archives/CDFP59223/p1704457939567069

[Robert Jaros]

And also this: https://kotlinlang.slack.com/archives/CDFP59223/p1715023745951099?thread_ts=1715023533.258459&cid=CDFP59223

[Edoardo Luppi]

Thanks. I'm not sure I fully understand what originates the problem tbh, that is why import.meta is required.
If that code is the "loading" code, isn't it better to output a different version based on the compilation target, instead of checking if it's Node or browser at runtime?

[Robert Jaros]

The code generated by the kotlin/wasm compiler contains several different loading methods, which are used in different environments (e.g. browser, node, deno etc.). I reported the original issue - the paths in the bundle were absolute paths to the local filesystem.

[Robert Jaros]

Currently there are no absolute paths in the bundle but the code still doesn't work in node. I use the workaround with a complex regexp in my gradle plugin.

[Edoardo Luppi]

Ahh damn, so it's a blocker for Node.js?

[Robert Jaros]

It's hard to say, because Kotlin gradle plugin doesn't support webpack for NodeJS.

[Robert Jaros]

I use my own custom task to generate webpack bundle.

[Robert Jaros]

There are no problems e.g. when running tests with node.

[Robert Jaros]

So it's a problem with a use case, which is unsupported anyway 🙂

[Edoardo Luppi]

There is an issue to add Webpack for Node tho. So solving it might be useful anyway.

[Artem Kobzar]

@Ilya Goncharov [JB] ^^