[blocklist]: replace O(N²) slice scans in updateInternal with map lookups

[zalegrala]

What this PR does:

Optimizes updateInternal (tempodb/blocklist/list.go). The add loop called slices.ContainsFunc(final, ...) on every incoming block — scanning up to 100K existing entries per block added. With ~1K new blocks per compaction cycle, that's 100M UUID comparisons per tenant per poll.

All four slices.ContainsFunc scans (remove, compactedAdd, compactedRemove, and the final dedup check) are replaced with map[backend.UUID]struct{} lookups built once before the loops, reducing the add phase from O(N·M) to O(N+M).

Benchmark (BenchmarkUpdateInternalLargeAdd — 100K existing blocks, 1K new adds):

# Before
BenchmarkUpdateInternalLargeAdd-16    51    231565779 ns/op    811008 B/op    1 allocs/op
BenchmarkUpdateInternalLargeAdd-16    51    241285117 ns/op    811117 B/op    1 allocs/op
BenchmarkUpdateInternalLargeAdd-16    50    238569548 ns/op    811008 B/op    1 allocs/op

# After
BenchmarkUpdateInternalLargeAdd-16    2493    4238611 ns/op    3175962 B/op    258 allocs/op
BenchmarkUpdateInternalLargeAdd-16    2352    4509954 ns/op    3176071 B/op    258 allocs/op
BenchmarkUpdateInternalLargeAdd-16    2415    4553389 ns/op    3175943 B/op    258 allocs/op

~53× faster per call. The alloc count increase reflects map bucket overhead for the 100K-capacity finalIDs map; memory stays under 4MB per call.

Checklist

• Tests updated
• Documentation added
• CHANGELOG.md updated - the order of entries should be [CHANGE], [FEATURE], [ENHANCEMENT], [BUGFIX]

[Copilot]

Pull request overview

Replaces four slices.ContainsFunc linear scans in List.updateInternal with pre-built map[backend.UUID]struct{} lookups, turning the regular-blocks add phase from O(N·M) to O(N+M). The benchmark (BenchmarkUpdateInternalLargeAdd) demonstrates ~53× speedup on a 100K-existing + 1K-add workload. Logic for the compacted-blocks branch is similarly converted, preserving the original semantics (compacted-add still dedupes against the pre-existing set only, matching prior behavior).

Changes:

• Build removeIDs, compactedAddIDs, compactedRemoveIDs, and finalIDs maps once and use map membership checks in both the regular and compacted update loops.
• Drop the slices import and the hasID / hasIDC closure helpers.
• Add BenchmarkUpdateInternalLargeAdd to guard against future regressions.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
tempodb/blocklist/list.go	Replace O(N²) slice scans in updateInternal with map-based lookups; logic remains equivalent to the prior implementation.
tempodb/blocklist/list_test.go	New benchmark exercising the large-add path (100K existing + 1K adds) to demonstrate and protect the perf improvement.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

[electron0zero]

lgtm