t27b: lower std.debug.print by a plan in t27 (reel_particles_mux, zig_open_slice) #10449

Follow-up to #10423 / PR #10431, t27b coverage epic #6063.

Boundary

• specs/tri/t27b/debug_print_plan.t27 (the decision, new) and its generated gen/rust/tri/t27b/debug_print_plan.rs (t27c gen-rust).
• specs/tri/t27b/builtin_plan.t27 and gen/rust/tri/t27b/builtin_plan.rs: the assert and eval-branch-quota decisions that lower.rs held inline move here.
• specs/tri/t27b/conformance/discard_value.t27 (conformance cases; no new file).
• cli/t27b/src/lower.rs glue only, net <= 0 lines.
• docs/reports/t27b_expectations.json rows and .trinity/seals/ for the specs that move.
• Out of scope: running the print (stderr is not observable to a verdict), std.debug.print with a non-literal format, any other std.* call.

User Scenarios

• Given a spec whose main or helper fn calls std.debug.print("{d}:{d}\n", .{ i, f(i) }) (specs/compiler/zig_open_slice.t27, specs/t27f/reel_particles_mux.t27 main), when t27b compiles it, then it does not refuse with ExprCall(std.*); in a fn Zig analyzes the arguments run in order, once each, and in a fn nothing analyzed reaches nothing runs.
• Given an analyzed fn whose format does not fit its tuple (a count mismatch, {d} of a bool, {s} of an integer), when t27b compiles it, then it refuses with a named reason, as the reference does not compile it.

Requirements

• FR-001: t27b MUST skip a std.debug.print in a fn Zig never analyzes, without lowering its arguments.
• FR-002: In an analyzed fn t27b MUST lower the arguments of std.debug.print(<string literal>, .{ ... }) left to right, once each, and drop the values, once debug_print_plan.t27 check accepts the format against the argument kinds.
• FR-003: t27b MUST refuse any other shape or a format that does not fit, with a name and detail from the plan.
• FR-004: The choice and the measured format matrix MUST be recorded in debug_print_plan.t27 with tests.

Success Criteria

1. t27b test --blockers no longer lists ExprCall(std.*) for specs/t27f/reel_particles_mux.t27 or specs/compiler/zig_open_slice.t27.
2. t27b test --check agrees with t27c test-report on specs/tri/t27b/debug_print_plan.t27, specs/tri/t27b/conformance/discard_value.t27 and specs/tri/t27b/builtin_plan.t27.
3. docs/reports/t27b_expectations.json moves zig_open_slice to pass and max_not_pass falls from 75 to 74.
4. Seals are re-saved and the hand-written cli/t27b/src/lower.rs is net <= 0 lines (git diff --numstat).