nx_dwline_stmt_gate.nx

nx_dwline_stmt_gate.nx -- DOES STATEMENT GRANULARITY ACTUALLY BEAT FUNCTION GRANULARITY? DDR-006 rung 2. Rung 1 emitted ONE .loc per function. This gate does not merely assert that the new compiler emits line info -- rung 1 did that too, and a gate that both versions pass measures nothing. It runs the SAME probe through BOTH compilers and measures the DIFFERENCE: OLD _offc/nx_cc_sovereign.elf (blessed 3bd47652, function granularity) -> lines {5, 9} NEW ../nx_compile_x86.elf (statement granularity) -> must ALSO carry 6 and 10 runtime/nx_probe_dbgtiny.nx: L5 func dbgtiny_a() -> i64 { L6 return 1 L7 } L9 func main() -> i64 { L10 return dbgtiny_a() L11 } so lines 6 and 10 are STATEMENT lines that function granularity CANNOT produce, and lines 5 and 9 are the function lines it already produced. THE OLD COMPILER IS THE NEGATIVE CONTROL, and it is what makes this a measurement of the improvement rather than a restatement of the feature. Deliberately reads the ".s" rather than the ELF: the claim under test is what the COMPILER emits. nx_dwline_e2e_gate already proves the assembler and decoder carry it through to a real ELF, so duplicating that here would test the wrong subject. license_tier: ORIGINAL expect_exit: 0