fix(migrate): properly handle rule removal and insertion

[Sec-ant]

Summary

• Fixes 🐛 npx @biomejs/biome migrate command crashes #3179. biome migrate can now properly handle the comma separators when removing stale nursery rules from the nursery group.
• When the target rule is already present in the target group, but the source rule is also present in the nursery group, biome migrate now only removes the source rule from the nursery group to avoid duplicate keys in the target group.
• Demote useImportRestrictions, see fix(migrate): properly handle rule removal and insertion #3207 (comment)

Sorry about the diff. I had to rename most of the variables to understand the code.

Test Plan

Added several test cases where the nursery rule to migrate is present in different positions. I also tested the repo provided in #3179.

[ematipico]

The schema seems not updated, for example it still sees useImportRestrictions as a nusery rule. What should I do to update the schema?

I'm not sure what you mean, useImportRestrictions is in nursery, currently.

[Sec-ant]

The schema seems not updated, for example it still sees useImportRestrictions as a nusery rule. What should I do to update the schema?

I'm not sure what you mean, useImportRestrictions is in nursery, currently.

It's defined in the RULES_TO_MIGRATE:

biome/crates/biome_migrate/src/analyzers/nursery_rules.rs

Line 114 in b70f405

("useImportRestrictions", ("style", "useImportRestrictions")),

[ematipico]

Oh, shoot, that's a mistake. It was meant for promotion, discussed it and decided to revert the change. I forgot to revert that too. We should remove it, it's a bug

[Sec-ant]

Oh, shoot, that's a mistake. It was meant for promotion, discussed it and decided to revert the change. I forgot to revert that too. We should remove it, it's a bug

I will fix it in this PR. Anyway it's a good bug because it helps us catch this bug :)

[ematipico]

Left a suggestion. Thank you for fixing this! I admit, I wasn't expecting a case of double rules