Feature: sort imports access level handling

[Brett-Best]

Attempts to resolve #6374.

[SwiftLintBot]

	1 Warning
⚠️	This PR may need tests.

	19 Messages
📖	Building this branch resulted in a binary size of 26873.88 KiB vs 26873.84 KiB when built on main (0% larger).
📖	Linting Aerial with this PR took 0.73 s vs 0.78 s on main (6% faster).
📖	Linting Alamofire with this PR took 1.06 s vs 1.04 s on main (1% slower).
📖	Linting Brave with this PR took 6.86 s vs 6.87 s on main (0% faster).
📖	Linting DuckDuckGo with this PR took 22.59 s vs 22.65 s on main (0% faster).
📖	Linting Firefox with this PR took 10.95 s vs 10.96 s on main (0% faster).
📖	Linting Kickstarter with this PR took 7.71 s vs 7.67 s on main (0% slower).
📖	Linting Moya with this PR took 0.46 s vs 0.42 s on main (9% slower).
📖	Linting NetNewsWire with this PR took 2.29 s vs 2.28 s on main (0% slower).
📖	Linting Nimble with this PR took 0.67 s vs 0.64 s on main (4% slower).
📖	Linting PocketCasts with this PR took 7.22 s vs 6.99 s on main (3% slower).
📖	Linting Quick with this PR took 0.44 s vs 0.41 s on main (7% slower).
📖	Linting Realm with this PR took 3.23 s vs 3.27 s on main (1% faster).
📖	Linting Sourcery with this PR took 1.78 s vs 1.77 s on main (0% slower).
📖	Linting Swift with this PR took 4.31 s vs 4.31 s on main (0% slower).
📖	Linting SwiftLintPerformanceTests with this PR took 0.33 s vs 0.3 s on main (10% slower).
📖	Linting VLC with this PR took 1.1 s vs 1.1 s on main (0% slower).
📖	Linting Wire with this PR took 17.24 s vs 17.24 s on main (0% slower).
📖	Linting WordPress with this PR took 11.4 s vs 11.35 s on main (0% slower).

Generated by 🚫 Danger

[SimplyDanny]

This makes sense to me, especially if the rule behaved like that before the rewrite. Apparently, there didn't exist enough tests to cover that. Thanks for the fix!

We'd need violating examples/tests that proof the right ordering after correction (in addition to the non-triggering ones).

[Brett-Best]

Thanks, I thought I had an edge case I couldn't solve for when the access level and attributes were combined but I couldn't remember it.

This should be a good improvement to merge in nonetheless 🙏🏻

I think I have included sufficient coverage in the examples and corrections now.

Let me know!

PS: As a side note, we'd appreciate even a pre-release of SwiftLint since we use the artifactbundle it is a pain for us to use commits that are untagged. Thanks for the consideration!

[Copilot]

Pull request overview

This PR reinstates handling of access level modifiers in the sorted_imports rule, addressing issue #6374. When the grouping: attributes configuration is enabled, imports are now sorted by attributes first, then by access level modifiers (public, package, internal, fileprivate, private), and finally by module name.

Key Changes

• Added access level modifier detection and sorting logic for Swift imports when using the attributes grouping mode
• Imports are sorted with more restrictive access levels (public) appearing before less restrictive ones (private), with plain imports appearing last
• Added comprehensive test examples covering various combinations of attributes and access level modifiers

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
Source/SwiftLintBuiltInRules/Rules/Style/SortedImportsRule.swift	Adds modifier field to Import struct and implements modifier extraction and comparison logic for access-level sorting
Source/SwiftLintBuiltInRules/Rules/Style/SortedImportsRuleExamples.swift	Adds non-triggering, triggering, and correction examples demonstrating the new access level modifier sorting behavior
CHANGELOG.md	Documents the bug fix in the Main section with proper credit and issue reference

[Brett-Best]

Providing CI passes and copilot is happy this is good for review 🙏🏻

[SimplyDanny]

PS: As a side note, we'd appreciate even a pre-release of SwiftLint since we use the artifactbundle it is a pain for us to use commits that are untagged. Thanks for the consideration!

There you go! 🎁