@@ -1392,3 +1392,19 @@ there on both binaries, for reasons of the environment (an unset `MCPP_HOME`;
13921392the machine's default toolchain), so the differential read them as unchanged.
13931393A test that fails on both sides measures nothing about the change; its reading
13941394has to come from an environment in which it passes, which here is CI.
1395+
1396+ ** #722 , completed after the first CI run.** The first pass split the seven named
1397+ functions and three more the gate found (10 findings at the base, 6 after it).
1398+ The second pass split the six that remained (` step6_provision_host_tools ` ,
1399+ ` loadVersionDep ` , ` phase11_scan ` , ` step1_target_and_static_overrides ` ,
1400+ ` phase2_define_toolchain_resolver ` , ` step9_kernel_abi_interfaces_and_requirements ` ),
1401+ each with a phase-local context struct where several of its steps share locals.
1402+ Readings on the final tree: the gate reports no function over 400 lines; the
1403+ seven fixtures are byte-identical to the pre-split binary; AddressSanitizer
1404+ reports nothing over the fixtures and a 20-script e2e subset; the unit suite
1405+ passes. The gate runs in ` ci-linux.yml ` 's LLVM toolchain job after mcpp builds
1406+ itself with llvm@20.1.7, with clang-tidy from ` xim:llvm-tools ` at the version
1407+ that wrote the compile database, because it reads that clang's BMIs. The merge
1408+ of the first pass took ` plan.cpp ` to 2,505 lines, over the file gate; the
1409+ records half of P13 (` mcpp.lock ` , ` resolution.json ` ) moved to ` records.cpp ` ,
1410+ byte-identical on the same fixtures.
0 commit comments