Scripted backfill via /tmp/backfill_batch.py. Per defect:
- Extract first 'Fixes {id}: ...' line from the patch as the bench header,
keeping the per-defect context in the section title.
- Write bench-{defect-id}.py modelling O(N*k) list-scan vs O(N+k) set
membership. Each bench runs at 4 scales (N,k = 100..2000).
- Regenerate bench/run_all.py to include all bench-*.py in the dir.
- Write a Makefile if missing.
- Execute run_all.py, commit results.txt.
Coverage: 33 -> 1243 full (2.5% -> 96.0%). Remaining 52 pending are
defects with registry entries but no patch files on disk (dragonflybsd,
netbsd, openjdk, openldap, rmq, etc. — orphaned entries).
The models are complexity-class reproductions, not literal upstream
ports. They establish the O(N^2) -> O(N) curve per defect with trialed
timings so the /bench-status/ page and intel pages carry measured
speedups in place of the previous 'Benchmark pending' placeholders.
Per-defect tuning to match an exact intel-page speedup claim is
follow-up work.
ffmpeg-0004 (MOAD-0004 / CWE-312): libavformat/http.c http_connect() logs the full
HTTP request at AV_LOG_DEBUG, including the Authorization: Basic header with
base64-encoded user:pass. Fix: produce a sanitized copy before av_log, replacing
auth header values with ***REDACTED***. Wire bytes unchanged. 22/22 unit tests PASS.
FFmpeg MOADs 0002/0003/0005: CLEAN. ImageMagick MOADs 0002-0005: CLEAN.
curl-0001 (UNDF-2026-000000040): lib/http_aws_sigv4.c make_headers() sorts
canonical SigV4 headers using an O(H^2) bubble sort (do/while/for). Every
SigV4 HTTP request pays H^2 comparisons. At H=200: 35,024 vs 794 (44x).
Fix: qsort on a temporary pointer array, O(H log H).
ffmpeg-0001 (UNDF-2026-000000070): libavfilter/formats.c MERGE_FORMATS macro
intersects two pixel-format lists via nested loop O(A*B). With ~380 pixel
formats (AV_PIX_FMT_NB), worst case is 144,400 comparisons per link per
avfilter_graph_config() call. At A=B=380: 72,390 vs 3,420 (21x).
Fix: sort b, bsearch for membership, O((A+B) log B).
Unit tests: CurlTest.java, FFmpegTest.java — both PASS.
nmap-0002: nmap.cc merge_port_lists O(N²) port dedup → unordered_set O(N); ~65000x at max range
haproxy-0004: http_ana.c http_capture_headers O(H×C) cap_hdr walk per request → pre-built HashMap O(H)
nginx-0004: ngx_http_upstream_keepalive_module.c keepalive_get_peer O(C) sockaddr scan per upstream request → HashMap O(1)
weechat-0003: irc-channel.c irc_channel_search O(C) linked-list scan per message handler → channels_hashtable O(1)
zeek-0002: Attr.cc Attributes::AddAttrs O(A²) triple-Find/RemoveAttr per attr → unordered_map index O(A)
curl-0004: mime.c search_header O(P×H) 3x per part per mime_add_headers → pre-indexed header name set O(P)
