test(server): prune tests that guard no behavior
Removes 299 of 2178 test declarations (13.7%) that guard no behavior, plus the test-only production code they kept alive. Coverage holds within 0.1 point, and the CPU tier loses its shared-temp-file flake.
The tests removed fall into these patterns:
- tests that rebuild the code under test inside the test body;
- call-shape checks on recording fakes;
- copied tables and inventories;
- environment variables nothing reads;
- prints with no assertion;
- duplicate invocations of one contract;
- tests whose only job was keeping a test-only seam alive.
Every deletion names the test that still catches the same bug, or says why there is no contract. Review the ledgers, not the diff (151 files): the reasoning per test is there.
Numbers
| Before (origin/main 4af911b7) | After | |
|---|---|---|
| Test declarations, C++ / Python / shell+cmake | 2178 (1853 / 316 / 9) | 1879 (1628 / 244 / 7) |
C++ CPU tier (ctest -L cpu, GPUs hidden) | 1474/1475 pass, 1 flake | 1274/1274 pass |
| C++ host coverage, lines / branches | 29.29% / 26.31% | 29.26% / 26.24% |
Python CI tier (-m "not server and not model and not slow"), lines / branches | 273 pass, 51.89% / 43.61% | 245 pass, 51.67% / 43.48% |
| Production code | 28 files, −244 / +31 lines |
Declarations removed per lane:
| Lane | Before → after |
|---|---|
| test_server_unit.cpp (3 lanes) | 657 → 562 (−95) |
| KVFlash / PFlash | 382 → 337 (−45) |
| Python, shell, CI | 316 → 244 (−72) |
| DFlash / speculative | 227 → 205 (−22) |
| Serving infra (cluster, slots, admission, IPC) | 184 → 162 (−22) |
| Kernels | 114 → 102 (−12) |
| DeepSeek4 unit | 113 → 102 (−11) |
| Qwen / GDN | 111 → 97 (−14) |
| DeepSeek4 rest | 74 → 68 (−6) |
The target was 20%. This stops at 13.7% because every remaining declaration has a named contract in its lane's Retain summary; about 25 medium-confidence deletions were skipped rather than cut on a guess.
Per-file coverage drops over 5 points are all explained:
- forwarding getters reached only through an unused constructor (
slot_kv.h); msg_type_name, which only builds log text;- gcov counting inline copies (
backend_ipc.h); - production code that was deleted (
glue_fusions.h,model_capabilities.h); quantize_draft_q8.py, whose keeper runs the CLI in a subprocess that coverage does not trace. Six mutations to the lost lines all turn that keeper red.
Production code removed (test-only seams)
Each had no caller outside tests (repo-wide grep, rechecked by the reviewers):
ModelBackend::compress_batchdefault and the cluster forwarder; the per-backendcompress_batchstays as the helper eachcompress()calls.supports_kvflash/supports_mixed_backend_layer_splitgetters on the backend and layer-split adapter headers.QueryCaptureSlice(pflash_compress.h); the vector overload ofselect_inline_evict_victim(prefix_cache), moved into the test file.floor_to_chunk,serialized_kv_k_type(kvflash_pager.h);kvflash_turn_cut(prefill_helpers.h).truncate_verified_path(ddtree);single_replay_inherited_reason(gdn_single_commit.h).deepseek4_dspark_dump; the always-trueenabledparameter of the DeepSeek4 roctx callbacks.- Glue-fusion
on/inert_whyand the twin gate'sinert_why, whichmmq.cualready covered.
Fixed along the way
- CPU-tier flake. Test processes running in parallel shared one fixed path in /tmp.
test_tmp_pathnow returns a per-process directory. - Weak tests rewritten so they can fail (the
Fentries in the ledgers). Examples:gdn_scalar_replay_bitsset an env var nothing reads; it now pins the scalar kernel and asserts one launch.- The Bailing router test now computes the router on CPU and checks the selected experts.
test_moe_streamused a fill pattern with a 256-byte period, so every expert's bytes were identical and a wrong-expert bug passed.
test_ds41_mxfp8registration restored. fb696713 had dropped it.
Preservation review
Six fresh reviewers, one per area, compared what was deleted against what remains. They found 7 contracts that had lost their only test, plus 4 assertions that could not fail or were too weak. Each is restored, and each restore was mutation-checked (keeper red with the mutation, source restored byte for byte):
| Contract | Restored where | Mutation it catches |
|---|---|---|
| Ladder evicts nothing within budget | evict_stops_when_only_protected_remain | evict before the budget check (prefill_helpers.h) |
| PFlash release default 16384 | test_launch_policy.cpp | threshold 16384 → 32000 (launch_profiles.h) |
| PFlash IPC refuses a negative query width | formatter_fails_closed_on_invalid_values | < 1 → == 0 (pflash_drafter_ipc.cpp) |
| Same-identity models of one process share the disk store | test_disk_cache_save_async_queue_and_free_space | drop the identity refcount (disk_store.cpp) |
| MMQ twin type membership (perf) | tile_choice_table | remove IQ4_XS from the 32-row twin set; add Q3_K to the 64-row twin set |
| DeepSeek V4 completed-turn render keeps the effort line (turn checkpoints extend the prompt) | test_deepseek4_render_system_only_gen_prompt | effort line only with a generation prompt (chat_codec.cpp) |
. is not a call: sentinel | test_parse_call_verb_glued_to_word_rejected | add . to the sentinel class (tool_parser.cpp) |
The reviewers also fixed:
test_rms_norm_scalenow skips only when the fusion never launched. A new skip had turned off a real check on gfx1201.- The kernel-1 bound in
test_flashprefill_kernelsis tightened from 1e-2 to 1e-3. - A carried
CompressResult none{}value-initialized the struct, so it passed with or without theok = falsedefault. It is now default-initialized, as production builds it. test_moe_stream's fill pattern is the 256-byte-period bug listed above.
Not run here: GPU tests
The GPU tests compile on HIP gfx1201 and CUDA sm_86 but were not run. Before merging, run the GPU tier on both vendors. These rewritten GPU tests have expected values worked out by hand or from source:
test_qn_split_fusionsargmax_ties_gpu(V=33, V=2)test_gdn_scalar_replay_bitstest_moe_stream(now registered as a ctest, CUDA only)test_flashprefill_kernelstest_rms_norm_scaletest_deepseek4_*HC{512,4}test_kvflash_gpu_mass
Judgment calls
- MMID ctests dropped. The six
deepseek4_mmid_grouped_cuda_*sub-ctests (two cmake scripts) were removed. They exercise the test binary's own CLI guards, skip code and benchmark loop. The parity ctest itself stays. - Silent row in
test_qwen35_ar_input_order. Its CUDA row is now left out silently when no GPU is present, where it used to report SKIP. - chat_template.cpp guards kept. Three guards whose tests were deleted stay, although production cannot reach them.
Follow-ups (not in this PR)
- CI gaps: these GPU tests run in no CI job:
q5_mmid_grouped_cudareturns 77 except on sm_86, and the 3090 runner is commented out.- The WMMA paged-attention comparator: both routes run in AMD CI, but nothing compares their dumps.
- The gfx1151 HIP parity tests,
gdn_single_*,gdn_scalar_replay_bitsandtest_moe_streamhave never run on the sm_121 runner, and lucebox3 has no step-time headroom.
- More host-only tests to audit for
cpu(pass with GPUs hidden, not yet checked against the CUDA hosted build):test_image_input,test_image_resize,test_image_decode,test_gpu_page_pool,thinking_budget,test_tool_checks,test_ddtree_path,test_chain_spec_shapes,test_draft_swa,rocmfp4_reference,rocmfpx_reference. - Test-only seams still in place:
PrefixCache::confirm_inline_snap(about 50 tests seed state with it);SeqSlotManager(SlotKv &, int).
- Untested writer:
write_shared_payload_segments(backend_ipc.cpp) has no direct test. - Loose bound:
test_flashprefill_kernelskernel 4 needs larger Q/K before a tighter bound means anything. - Pre-existing flake, untouched here:
test_concurrent_prefix_budget_allowance_clampingreads live available memory twice and compares the results, so it can fail underctest -j. It failed once in 5 full runs and passes alone.
Follow-up commits
These answer the open questions. A Fable 5.2 review decided each one from the source.
-
Delete
src/pflash/anchor_scan.{h,cpp}andtest_anchor_transitive.cpp. They never had a production caller: cefa3caf only included the header, and c2569a0a deleted that includer. -
Delete
gpu_sampler_microbench_when_enabled. It asserts onlyCHECK(true)behindLUCE_SAMPLER_BENCH, which nothing sets. -
Label host-only tests
cpuso the hosted job runs them:- DS4V and DS4.1V image and vision;
- DS4 engram, dmix and mix-converter;
- deepseek4 and qwen35 roctx;
- qwen35 image and split-tree guard;
- recurrent snapshot;
- qwen4exp MTP and MTP dispatch;
server_cli_model_names. This last one addsluce_servertocheck-cpu.
All of them pass in a CUDA
check-cpurun in docker with no GPU and thelibcudastub, set up like the hosted job (1318/1318).qwen4exp_prefix_cachestays unlabelled: on CUDA its KVFlash pager records events the stub refuses. Drop any label the hosted Build still turns red.
Proposed test-ownership rules
These come from mistakes this campaign found repeatedly. Candidates for a server/AGENTS.md or the Tests section of DEVELOPER.md:
- Assert on the production function, never on a copy of its loop or table in the test body.
- No test-only overloads, getters or parameters in
server/src. Test through the real entry point, or move the helper into the test. - Expected values come from an independent reference or a literal golden, never from the code under test.
- A GPU test that cannot exercise its path exits 77 (SKIP). It never passes vacuously, and never skips when the path does run.
- Temp files go under
test_tmp_path, never a fixed path in /tmp. - An env var a test sets must have a reader. Grep for it before relying on it.
The ledgers (one per lane: every deletion with its keeper, the carried assertions, the applied list), the six review reports and the coverage diffs are attached to LUC-284.
Part of LUC-284
🤖 Generated with Claude Code