Result: partial (6 of 11 assigned fixed; the rest open, see Needs)
Pushed: 282dc867 fa7ba00d 5fa2a7a1 9dd789bc 5162370e 2d6c1344 (develop, fast-forward)
Tests: tests/compiler –both-settings 348/348 (898, 936), then 351/352 (909, 917, 931; the one failure was my own test using print, fixed to println); guard set + lent_arguments:: green; borrowed_subject:: 9/9 after 2d6c1344; clippy -D warnings clean; pre-push-check ok; lower-std at its fixpoint
Failed unfixed: 898 NK1102; 936 NK1131 x3; 909 rustc E0308; 917 rustc E0271; 931 NK1245 x2
Deviation: commits made with git commit and hunk staging (git apply –cached) rather than scripts/commit.nika, because emit_expr.nika/.rs, lowered.rs, bounds_walk.rs and check_numbers4.rs also held other lanes’ unfinished hunks; only my hunks went in
Changelog 936: a text field of a borrowed subject handed to a parameter whose contract keeps nothing of it is lent, not a move (no NK1131); a keeping parameter still refuses it
Changelog 898: a proved ref T? parameter reads as ref T; an option bound by reference out of a lent subject reads through .as_ref().unwrap() and goes back into a ref T? slot as Option::as_ref(x) (narrowed_reads bit 32)
Changelog 909: a tuple literal handed to a kept generic parameter is checked part by part; a lent part is NK1102 with the .clone() help
Changelog 917: the element/key/value argument of push/insert/contains/remove/get/contains_key/entry asks an open integer for the receiver’s type; only Vec/set contains needed it; the toolchain’s lowered code now writes : i64 on such locals
Changelog 931: sum/product of a walk over a list of number literals is that open number (asked by let total: i64, else the default i64, written on the let)
NK codes: none used (NK1308-1311, NK2706-2707 still free)
Found 900: lane T’s file. text_tiers.nika tier_struct_of has no Expr::MethodCall arm for .clone(), so let mut c = a.clone(); c.name = t leaves the field &'static str (NK1105). Fix: one arm, clone with no args -> tier_struct_of(receiver)
Found 2d6c1344: 282dc867 (#936) turned develop red: borrowed_subject’s takes(s: String) ignored s, so it is now a reader; the test now keeps s (the new behaviour is right)
Found laneT-1: let mut sg = s.get("b") ?? return; sg.push_str("?") does not reproduce at 5162370e (prints p! q?); recheck against lane T’s tree
Found laneT-2/#980: [[], ["q"]] into Vec[Vec[String]] lowers as Vec<Vec<&’static str»: a push through lists[0] is not linked in text_tiers (source_of/place_of lack Expr::Index; #980’s own body). Tier pass, not check_core_args owned_texts; lane T’s
Found p980 (lane E’s probe): bind(t, w) where bind keeps name into w.items (a mixed-tier Vec) is lowered either(&t) on the path before return; lending a local into a kept mixed slot that outlives it needs a decision on when check_core_call_lent may lend to a kept position. Not fixed
Found 927: an index read of an owned list (out.push(rest[i]), String) is not covered: check_core_a_lent_place_given_away returns for unlent roots and for all text; telling owned from literal-only text needs the tier result. Not started
Needs: #927 #591 #919 #677(step 1) #968 #907 #947 #938 #896(step 4) #627 #980 are not started (turn budget); a fresh lane C can take them from here
Gate: green
Files: docs/briefs/lane-c.md.report.md
Commit: lane C report: #936 #898 #909 #917 #931 and the borrowed_subject fix
Cache: target/nikaia-project-tests 2.1G; disk / 9.3G free
