Optimize slice comparison utility

[slick-daddy]

The existing SliceCompare performed repeated linear membership scans which is O(n*m) and is hot in edit paths that compare aliases/images/tags, so lowering complexity will reduce CPU on those paths.

Description

• Replace repeated Includes linear scans in SliceCompare with prebuilt set maps via a new makeSet helper to achieve O(n) lookups and reduce overall work in comparisons in pkg/utils/slice_compare.go.
• Add small maps (addedSet and missingSet) to suppress duplicates while preserving the original output order semantics of SliceCompare.
• Add a benchmark BenchmarkSliceCompare in pkg/utils/slice_compare_test.go exercising two 1,000-item slices with partial overlap to guard future regressions and measure improvements.
• Keep Includes available for callers that rely on it while centralizing set construction in makeSet for reuse and readability.

Drafted with Codex.

Edits usually don't include many objects, so practical change can be limited.

[InfiniteStash]

Thanks! I removed the benchmark and tweaked a bit.