TASK-080: mark Done in M7 milestone index and task file, gitignore build-debug/

- Flip Status from Backlog to Done in TASK-080.md
- Tick all four action items; annotate item 3 noting 10× was infeasible
  (gate set at 20× p95 with measurement data in test/PERFORMANCE.md)
- Flip TASK-080 row in _index.md from Backlog to Done
- Add build-debug/ to .gitignore (developer noise-floor measurement dir)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: etr

.gitignore

@@ -15,6 +15,7 @@
 .idea
 libhttpserver.iml
 build/*
+build-debug/
 aclocal.m4
 autom4te.cache/
 config.guess

specs/tasks/M7-v2-cleanup/TASK-080.md

@@ -11,10 +11,10 @@ as a real loss of regression bite. Restore a tighter gate by attacking the
 noise rather than the threshold.
 
 **Action Items:**
-- [ ] Profile what causes the CI-runner noise (cold caches, sibling-CPU scheduling, MHD socket-accept jitter). Capture median, p99, and max under the current runner.
-- [ ] Stabilise the warmup: more iterations, pin to a single CPU on Linux (`taskset`), discard the slowest N% per round, use median-of-medians rather than a single median, or switch to a high-precision monotonic timer.
-- [ ] With the noise floor characterized, restore the gate to 10× (or the tightest threshold that survives 99% of CI runs across the matrix).
-- [ ] If a 10× gate is genuinely infeasible on shared CI runners, document the chosen floor in the test comment and in `test/PERFORMANCE.md`, with measurement data backing it.
+- [x] Profile what causes the CI-runner noise (cold caches, sibling-CPU scheduling, MHD socket-accept jitter). Capture median, p99, and max under the current runner.
+- [x] Stabilise the warmup: more iterations, pin to a single CPU on Linux (`taskset`), discard the slowest N% per round, use median-of-medians rather than a single median, or switch to a high-precision monotonic timer.
+- [x] With the noise floor characterized, restore the gate to 10× (or the tightest threshold that survives 99% of CI runs across the matrix). **Note:** 10× was found genuinely infeasible — p95/baseline ratio runs 11×–14× on quiet Apple Silicon due to legitimate route_table_mutex_ contention, not OS noise. Gate is set at 20× p95 (documented floor) with full measurement data in test/PERFORMANCE.md.
+- [x] If a 10× gate is genuinely infeasible on shared CI runners, document the chosen floor in the test comment and in `test/PERFORMANCE.md`, with measurement data backing it.
 
 **Dependencies:**
 - Blocked by: TASK-032 (Done; original stress test)
@@ -30,4 +30,4 @@ noise rather than the threshold.
 **Related Requirements:** PRD §3.6 performance acceptance
 **Related Decisions:** DR-008 (thread-safety contract)
 
-**Status:** Backlog
+**Status:** Done

specs/tasks/M7-v2-cleanup/_index.md

@@ -44,7 +44,7 @@ TASK-093).
 | TASK-077 | Restore Windows / Darwin coverage in skipped test suites | HIGH | L | Done |
 | TASK-078 | Resolve commented-out CONNECT-method test bodies | HIGH | S | Done |
 | TASK-079 | Drive nonce/opaque state machine in v2 digest-auth integ tests | MED | M | Done |
-| TASK-080 | Tighten threadsafety_stress latency gate back from 100× to 10× | MED | M | Backlog |
+| TASK-080 | Tighten threadsafety_stress latency gate back from 100× to 10× | MED | M | Done |
 | TASK-081 | Fill empty-on-correct-build unit suites and re-enable pthread leak detector | MED | M | Backlog |
 | TASK-082 | Tighten static-size bounds in `http_resource_test` and `webserver_pimpl_test` | MED | S | Backlog |
 | TASK-083 | Wire real CI gates into benchmarks | MED | M | Backlog |