refactor(data-plane): drop the counter-delta branch a selected program never takes

[zzylol]

RawDagProgram::uses_counter_delta returned a constant false:

pub fn uses_counter_delta(&self) -> bool {
    // Planner represents rate computation as an explicit upstream operator.
    false
}

Three call sites still branched on it, so the counter-delta side was unreachable whenever a program was present.

Site	Before	After
worker.rs:549	program.map_or_else(legacy_rule, |p| p.uses_counter_delta())	legacy_counter_delta(state) — the legacy rule only applies with no program
worker.rs:609	same	same
window_builder.rs:172	if program.uses_counter_delta() { reset_aware_counter_delta(...) } else { Some(value) }	applied directly; program is a required parameter there, so the branch was dead outright

The previous map in window_builder.rs existed only to feed reset_aware_counter_delta and goes with it.

Behaviour

Unchanged. Every removed path was already unreachable, which is why this is a deletion rather than a rewrite.

cargo test -p data_plane --lib 934 passed, cargo test -p control_plane --lib 433 passed, cargo fmt --check and cargo clippy --all-targets with -D warnings clean.

Context

Found while checking, after #774, which backend abstractions the shared asap-physical-operators crate had made redundant. RawDagProgram itself is not redundant: it already delegates accumulation to asap_physical_operators::factory, and what remains is the backend's own binding glue. This constant method was the one genuinely dead piece.

🤖 Generated with Claude Code