https://gcc.gnu.org/bugzilla/show_bug.cgi?id=128852
--- Comment #7 from David Malcolm <dmalcolm at gcc dot gnu.org> ---
****************************************************************************
BenchProctor: 2026.11.19.1 (23a5f7bde67d4ee9f7ebf0994084cb414c6cc2fa)
-fanalyzer: GCC trunk: r17-6248-gf7f84d53720030
****************************************************************************
==============================================================================
CATEGORY-AVERAGED SCORING (macro-averaged)
==============================================================================
Each category is weighted equally regardless of test count.
Each category is scored independently and averaged.
Category CWE TPR FPR Score
--------------------------------------------------
buffer_overflow 121 0.0% 0.0% +0.0%
buffer_overread 126 0.0% 0.0% +0.0%
double_free 415 100.0% 0.0% +100.0%
heap_buffer_overflow 122 0.0% 0.0% +0.0%
intoverflow 190 0.0% 0.0% +0.0%
memory_leak 401 100.0% 0.0% +100.0%
null_deref 476 100.0% 0.0% +100.0%
uninit_variable 457 100.0% 0.0% +100.0%
use_after_free 416 26.7% 0.0% +26.7%
--------------------------------------------------
MACRO AVERAGE 47.4% 0.0% +47.4%
==============================================================================
FLAT AGGREGATE SCORING
==============================================================================
Category CWE TP FP FN TN TPR FPR Score
------------------------------------------------------------------------------
buffer_overflow 121 0 0 30 30 0.0% 0.0% +0.0%
buffer_overread 126 0 0 30 30 0.0% 0.0% +0.0%
double_free 415 30 0 0 30 100.0% 0.0% +100.0%
heap_buffer_overflow 122 0 0 30 30 0.0% 0.0% +0.0%
intoverflow 190 0 0 30 30 0.0% 0.0% +0.0%
memory_leak 401 30 0 0 30 100.0% 0.0% +100.0%
null_deref 476 30 0 0 30 100.0% 0.0% +100.0%
uninit_variable 457 30 0 0 30 100.0% 0.0% +100.0%
use_after_free 416 8 0 22 30 26.7% 0.0% +26.7%
------------------------------------------------------------------------------
TOTAL 128 0 142 270 47.4% 0.0% +47.4%
==============================================================================
SUMMARY
==============================================================================
Total test cases: 540
Vulnerable (ground truth): 270
Safe (ground truth): 270
TP/TN balance: 50/50
Tests with findings: 128
Declared out of scope: cmdi, ldapi, pathtraver, sqli, ssrf, weakcipher,
weakhash, weakrand, xpathi, xxe
Scope: 9 categories scored, 10 left out by the tool
Flat Score: +47.4%
Category-Averaged: +47.4% (macro-averaged)
Score interpretation:
+100% Perfect (catches everything, zero false alarms)
0% No measured discrimination (TPR equals FPR; flag-everything
and flag-nothing both land here)
-100% Perfectly wrong (flags safe, misses vulnerable)