Convert to ESM module (BREAKING). Target: end of beta cycle.

[gadicc]

Rationale
Sindre Sorhus: Get Ready For ESM

Implementation
Sindre Sorhus: Pure ESM package

Sugested timeline

• Target: before first non-beta release (planned around July)
• When things get quiet on other stuff, perhaps we can release with an rc tag.

[gadicc]

Actually we can do the "Dual" approach from the Rationale link above, then it's not a breaking change. We'll make the library ESM but also have a build step that transpiles a CommonJS fallback.

[advaiyalad]

😆 I was about to open an issue for this! You beat me to it! Yes, I agree. ESModules will a lot better for the rest of the ecosystem.

Suggested Timeline

• Start planning important/helpful/required breaking changes
• Incorporate this into that list
• Release new major version with the changes when the list's length is long enough

Dual

I would say that we should not stick to the dual approach. We can start with it, but we should remove the CJS fallback at the end of April.

[gadicc]

Thanks, @PythonCreator27, glad to hear others are on board for this 🙏

Yes, I don't want to support dual moving forward. But I think we can give people a bit more time to get ready if it doesn't cost us anything. Maybe 6 months with a big fat warning with anyone loading the CJS fallback.

From what I understood, as long as we provide the fallback, it won't be a breaking change and should be transparent to the user. It will only be a breaking change / new major version when we drop the fallback.

[gadicc]

🎉 This issue has been resolved in version 1.11.0 🎉

The release is available on:

• npm package (@latest dist-tag)
• GitHub release

Your semantic-release bot 📦🚀