User:Sahilvaidya/CQL/Log

From BRL-CAD

Development Log

2/17

  • Found exactly how to create a test

2/19

  • Found best files to begin trying to fuzz

2/24

  • Made basic programs using zzuf and afl to begin fuzzing

2/27

  • Tried to integrate AFL with BRL-CAD tests

3/2

3/4

  • Tried to link LibFuzzer target to BRL-CAD library
  • Struggled with the compilation of BRL-CAD (required installing libraries not needed before)
  • Struggled with compiling target so it had BRL-CAD libraries

3/9

  • Experimented with compiling target and consulted Shikhar, a PHd student at UT
  • Found the BRL-CAD header file and the exact syntax needed to fully compile my target
  • Built the target, ran it in various ways tinkering with how the data was input into bu_sort, checking if the output was correct, adding seeds to the input
  • Used this compilation statement

clang++ -g -O1 -fsanitize=fuzzer -Iinclude -Ibuild/include -Ibuild/include/openNURBS/ target.cc -Lbuild/lib -Wl,-rpath -Wl,build/lib -lrt -lbu

3/23

  • Learned Jenkins CI and successfully built BRL-CAD
  • Started creating a build to test BRL-CAD immediately

3/25

  • Fixed testing BRL-CAD in Jenkins
  • Began creating infrastructure for a "make fuzz"

3/30

  • Looked at other regression tests cmake files to find what exactly is required to compile
  • Tried to mimic these to compile fuzzing tests

4/1

  • Switched strategies for compiling using cmake, spent most of my time reading documentation to better understand how exactly cmake works and how that can be applied to fuzzing