topic: self
45 modules sharing the self name family (derived from the tree's prefix discipline).
The 'self' topic family in the Nishi sovereign ecosystem focuses on validating the self-hosting capabilities of the NishiLang runtime, ensuring it can correctly interpret and execute low-level operations. Modules like _self_host_li_test verify accurate handling of large constants, while _self_host_mem_test confirms proper memory management through mmap and store operations. _self_host_or_shift_test investigates bitwise operations, ensuring expressions involving shifts and OR function as expected within the self-host environment.
auto-narrated by the local model from this topic's module headers; links verified against the wiki index.
narrated overview -- maintained by the narration lane, module links verified against this wiki.
| module | description | lines | funcs |
|---|---|---|---|
| _self_host_lane_binops_test.nx | one smoke each. | 65 | 1 |
| _self_host_li_boundary_test.nx | exercises corner cases at | 59 | 1 |
| _self_host_li_test.nx | Test that the self-host can load a 49-bit constant correctly. | 10 | 1 |
| _self_host_mem_test.nx | Test that mmap+store+read works in self-host. | 14 | 1 |
| _self_host_neg_li_test.nx | Test that li of large negative value works through self-host. | 9 | 1 |
| _self_host_or_shift3_test.nx | 9 | 1 | |
| _self_host_or_shift_test.nx | Bisect: does `expr | (expr << N)` work in self-host? | 10 | 1 |
| _self_host_pack4_neg_test.nx | Test pack4_i16 with negative first arg + vreduce_min. | 22 | 2 |
| _self_host_pack4_oneline.nx | Same pack4 but body on ONE line. If this passes but multiline | 13 | 2 |
| _self_host_pack4_repro.nx | Exact reproduction of pack4_i16 to isolate the bug. | 17 | 2 |
| _self_host_sat_trio_test.nx | exercise 3 new saturating ops | 62 | 1 |
| _self_host_shift32_test.nx | 9 | 1 | |
| _self_host_shift_arith_matrix.nx | Regression matrix: every shift + funct7-discriminated arith op | 86 | 7 |
| _self_host_simd_i32_test.nx | i32x8 set: 5 ops verified. | 46 | 1 |
| _self_host_simd_i64_test.nx | i64x4 closes the 4-width spectrum. | 63 | 1 |
| _self_host_simd_i8_test.nx | i8x32 set: 6 ops, 6 assertions. | 52 | 1 |
| _self_host_simd_minmax_sat_test.nx | 3 ops in one smoke. | 49 | 2 |
| _self_host_simd_shift_test.nx | vsll/vsrl/vsra for i16x16. | 55 | 1 |
| _self_host_simd_sum_bcast_test.nx | vreduce_sum + vbroadcast. | 58 | 1 |
| _self_host_stmt_boundary_regression.nx | Regression for at_stmt_boundary refinement (2026-05-16). | 48 | 4 |
| _self_host_vdot_full.nx | full vdot [1..16] . [1..16] via self-host. | 31 | 1 |
| _self_host_vdot_simple.nx | minimal sanity: dot([1,1,..1] . [1,1,..1]) = 16. | 20 | 1 |
| _self_host_vdot_test.nx | proves the self-host can emit | 39 | 2 |
| _self_host_vreduce_min_test.nx | Minimal smoke for vreduce_min: just call it and return modulo 100. | 14 | 1 |
| nx_self_audit.nx | substrate-native self-awareness primitive. | 184 | 3 |
| nx_self_build.nx | the 3-tier SELF-BUILD POLICY + the reversible-modify | 365 | 14 |
| nx_self_build_cycle_test.nx | THE CAPSTONE: a full autonomous SELF-BUILD CYCLE, | 109 | 5 |
| nx_self_build_test.nx | THE NET, PROVEN BEFORE THE TRAPEZE. | 186 | 6 |
| nx_self_compose.nx | AUTONOMOUS compose+derive (the rung up: the team composes its OWN | 154 | 14 |
| nx_self_heal_conductor.nx | the CONDUCTOR's autonomous self-heal BEAT. The Conductor | 62 | 2 |
| nx_self_heal_conductor_test.nx | prove the CONDUCTOR's Warden-gated self-heal beat on | 56 | 5 |
| nx_self_host_candidate_verify.nx | the TEAM's capability to drive codegen | 74 | 1 |
| nx_self_host_load_codegen_repro.nx | MINIMAL REPRO for the | 52 | 2 |
| nx_self_measure_cost.nx | BUILD OUR OWN cost data, on whatever chip the seed | 53 | 6 |
| nx_self_model_lint.nx | the LIBRARIAN's validator for the self-model registry, so a bug like a cap | 48 | 6 |
| nx_self_model_regen.nx | the LIBRARIAN regenerates the self-model count FROM the registry (data), instead | 25 | 4 |
| nx_self_model_regen_test.nx | prove the registry-driven self-model: a SMALL loop writes CAPREG lines | 55 | 5 |
| nx_self_reliance_gate.nx | HONEST self-reliance map (operator: "total self reliance, self sufficient, building | 57 | 3 |
| nx_self_repair_cycle_test.nx | the self-build cycle UPGRADED to self-REPAIR, | 125 | 4 |
| nx_self_similarity.nx | box-counting fractal dimension primitive. | 253 | 6 |
| nx_self_similarity_test.nx | smoke for box-counting fractal dim. | 113 | 1 |
| nx_self_sufficiency.nx | the team models its OWN full-loop self-sufficiency (operator: "what is our | 50 | 5 |
| nx_self_sufficiency_test.nx | honest map of the full loop. Two cases: | 45 | 3 |
| nx_self_tune.nx | LOCAL-ONLY DRIVER SELF-TUNE (X-DRV-F3). | 164 | 12 |
| self_parse_test.nx | feed parse.nx's OWN source through the | 49 | 1 |