Class field initialization false positives?

[ziacik]

Hi, is this a bug?

This

and this

is syntactically equivalent, isn't it?

It should not give error in the first case, should it?

[lukastaegert]

Yes, but the plugin is quite old and badly maintained, which is why advanced logic for class fields is missing.

[ziacik]

@lukastaegert

Thanks for your reply.

I'd be happy to try providing a PR, but I noticed this particular case is explicitly set as invalid in tests.
I'm unable to identify a reason for this decision. Could you please offer an explanation why this is considered a possible side effect?

eslint-plugin-tree-shaking/tests/lib/rules/no-side-effects-in-initialization.js

Lines 771 to 780 in c424400

	{
	code: "class x {y = ext()}",
	parser: PARSER_BABEL,
	errors: [
	{
	message: "Cannot determine side-effects of calling global function",
	type: "Identifier",
	},
	],
	},

[lukastaegert]

Because ext is an unknown function with unknown side effects

[ziacik]

Yes, ext can have side effects, but it is only called at class construction, so it shouldn't be considered side-effected unless the class is globally constructed, no?

[ziacik]

I've created a PR for this. If you agree with my points, I'd appreciate if you could review and merge the PR. Thank you.