[loops] Benchmarks with low coverage

[izlatkin]

apache-get-tag.i.p+nlh-reducer.c - huge benchmark with a lot of conditions apache-get-tag.i.p+sep-reducer.c - huge benchmark with a lot of conditions benchmark09_conjunctive.c diamond_2-2.c n.c11.c n.c24.c string-1.c Error: key \d+ not found string-2.c Error: key \d+ not found terminator_02-2.c Error: key \d+ not found veris.c_OpenSER_cases1_stripFullBoth_arr.c Error: key \d+ not found vogal-1.c MADWiFi-encode_ie_ok.c apache-escape-absolute.i.v+cfa-reducer.c Big benchmark

benchmark28_linear.c benchmark37_conjunctive.c benchmark43_conjunctive.c benchmark46_disjunctive.c gauss_sum.i.p+lhb-reducer.c gsv2008.c.i.v+lhb-reducer.c jm2006.c jm2006.c.i.v+cfa-reducer.c mod3.c.v+cfa-reducer.c multivar_1-1.c n.c40.c nested9.c simple_array_index_value_4.i.v+nlh-reducer.c

I am not sure that there is one feature of benchmarks, but I noticed a few moments witch could affected the results

1. Case when coverage is significantly low: A) TG could not find a condition for a branch that guards several other conditions, so they stay uncovered B) “Key not found” - compiler merge conditions
2. Case when coverage is less than 20% lower than VeriFuzz: A) One of the double conditions is not coverage (for example: x < 100 && y < 100) B) Arrays C) goto D) border cases, like x > LARGE_INT

[izlatkin]

recheck: vogal-1.c

[grigoryfedyukovich]

For vogal-1.c, we have Error: key 22437 not found