run compute_simhash and collapse_urls in the thread executor

[liquidsec]

Summary

• waf_bypass submits compute_simhash on full HTTP response bodies through the process pool. xxhash releases the GIL and compute_simhash truncates its input to 3 KB internally, so the pickle-to-worker IPC dominates the actual work.
• More importantly, when a worker submitted this way wedges, the only recovery is helper._reset_process_pool, which fires 300 s after the stall. On a real 14 h scan against a large target set this pattern was observed to stall the pipeline for the full timeout window before the reset recovered. bevigil.collapse_urls was on the same pattern at smaller scale.
• Neither operation needs process isolation. Both move to run_in_executor_cpu (the thread pool). kreuzberg PDF extraction and badsecrets stay on run_in_executor_mp -- crash isolation from adversarial input is the point there.

Changes

• bbot/modules/waf_bypass.py: two run_in_executor_mp(compute_simhash, ...) sites moved to run_in_executor_cpu
• bbot/modules/bevigil.py: run_in_executor_mp(collapse_urls, ...) moved to run_in_executor_cpu

[github-actions]

📊 Performance Benchmark Report

Comparing dev (baseline) vs simhash-cpu-executor (current)

📈 Detailed Results (All Benchmarks)

📋 Complete results for all benchmarks - includes both significant and insignificant changes

🧪 Test Name	📏 Base	📏 Current	📈 Change	🎯 Status
Bloom Filter Dns Mutation Tracking Performance	4.34ms	4.43ms	+2.1% ⚪	✅
Bloom Filter Large Scale Dns Brute Force	17.90ms	18.91ms	+5.7% ⚪	✅
Large Closest Match Lookup	369.30ms	369.11ms	-0.1% ⚪	✅
Realistic Closest Match Workload	195.58ms	194.34ms	-0.6% ⚪	✅
Event Memory Medium Scan	1402 B/event	1402 B/event	+0.0% ⚪	✅
Event Memory Large Scan	1527 B/event	1527 B/event	+0.0% ⚪	✅
Event Validation Full Scan Startup Small Batch	410.45ms	405.62ms	-1.2% ⚪	✅
Event Validation Full Scan Startup Large Batch	549.71ms	546.59ms	-0.6% ⚪	✅
Make Event Autodetection Small	26.26ms	26.33ms	+0.3% ⚪	✅
Make Event Autodetection Large	271.82ms	270.11ms	-0.6% ⚪	✅
Make Event Explicit Types	11.86ms	11.86ms	-0.0% ⚪	✅
Excavate Single Thread Small	4.425s	4.335s	-2.0% ⚪	✅
Excavate Single Thread Large	10.069s	10.103s	+0.3% ⚪	✅
Excavate Parallel Tasks Small	4.621s	4.513s	-2.3% ⚪	✅
Excavate Parallel Tasks Large	6.657s	6.677s	+0.3% ⚪	✅
Intercept Throughput Small	881.05ms	889.13ms	+0.9% ⚪	✅
Intercept Throughput Medium	970.32ms	984.13ms	+1.4% ⚪	✅
Dns Throughput Quiet	2.852s	2.851s	-0.0% ⚪	✅
Dns Throughput Loaded	2.020s	2.020s	+0.0% ⚪	✅
Dns Throughput Inherited	2.379s	2.458s	+3.3% ⚪	✅
Is Ip Performance	2.30ms	2.32ms	+0.7% ⚪	✅
Make Ip Type Performance	248.79µs	244.81µs	-1.6% ⚪	✅
Mixed Ip Operations	2.43ms	2.44ms	+0.3% ⚪	✅
Memory Use Web Crawl	1.0 MB	1.0 MB	+1.0% ⚪	✅
Memory Use Subdomain Enum	28.3 MB	28.3 MB	+0.0% ⚪	✅
Memory Use Deep Chain	5.0 MB	5.0 MB	+0.0% ⚪	✅
Memory Use Parallel Chains	14.3 MB	12.7 MB	-11.1% 🟢🟢	🚀
Scan Throughput 100	2.937s	3.091s	+5.3% ⚪	✅
Scan Throughput 1000	22.634s	23.688s	+4.7% ⚪	✅
Typical Queue Shuffle	5.44µs	5.40µs	-0.8% ⚪	✅
Priority Queue Shuffle	26.51µs	26.13µs	-1.4% ⚪	✅

🎯 Performance Summary

+ 1 improvement 🚀
  30 unchanged ✅

🔍 Significant Changes (>10%)

• Memory Use Parallel Chains: 11.1% 🚀 less memory

---

🐍 Python Version 3.11.15

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90%. Comparing base (3cf28a4) to head (aff3a5d).
⚠️ Report is 6 commits behind head on dev.

Additional details and impacted files

@@          Coverage Diff          @@
##             dev   #3331   +/-   ##
=====================================
- Coverage     90%     90%   -0%     
=====================================
  Files        450     450           
  Lines      46285   46285           
=====================================
- Hits       41547   41537   -10     
- Misses      4738    4748   +10     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[liquidsec]

#3329