build(deps): track ASAPPlanner main

[milindsrivastava1997]

Switches the ASAPPlanner git deps (asap-types, asap-aware-mapping, asap-frontend-promql, asap-frontend-sql) from the pinned rev 029ff2fe to branch = "main" (lock now at 8b90f04c). The other git deps (asap_sketchlib, asap-precompute-rs) were already on main.

Planner main added DataType::{Date, Interval} and ScalarValue::Interval. The ClickHouse relational adapter and backfill reader now reject them as unsupported (interval table-population literals are rejected in TablePopulation::validate, alongside nulls). Workspace builds.

⚠️ Two control_plane tests fail with Planner main (both pass on origin/main); see the comment below.

🤖 Generated with Claude Code

[milindsrivastava1997]

@zzylol could you help with two control_plane tests that fail after bumping ASAPPlanner from 029ff2fe to main (8b90f04c)? Both pass on origin/main, and both look like Planner selection no longer choosing a summary.

1. physical::compiler::tests::raw_per_entity_state_carries_explicit_isolation

panicked at control_plane/src/physical/compiler.rs:4752: post-ASAP selection: NoLegalCandidate

Fails on sum_over_time(m[1m]) / count_over_time(m[1m]). select_summary_with_evidence (planner_selection.rs:455) calls SketchAlgorithmStrategy::with_models_and_evidence(...).replacements(&TargetSubDAG::new(&root)) with EpsilonDelta { 0.01, 0.01 }, DefaultAccuracyModel, EqualSplitAllocator, and ControlPlaneCostModel; it now returns zero replacements.

2. physical::compiler::tests::issue_701_702_temporal_workloads_have_warm_candidates

panicked at control_plane/src/physical/compiler.rs:4145: no warm candidate for quantile_over_time(0.9,data[5m])/quantile_over_time(0.5,data[5m]): []

Every candidate compiles, but none has a materialization, i.e. search_workload_with_targets + global_selection (planner_selection.rs:377, strategies at :348) appears to pick KeepPreAsap at EpsilonDelta { 0.01, 0.01 }.

Both involve a binary op over two range-vector functions, so they may share a cause. Is this an intended Planner behavior change, or a regression?