Closed Strke closed 8 months ago
Hello! Sorry for late.
Did you put UTopia generated seed in the queue?
UTopia
supports automated seed generation, you should put it in a specific directory, for example 'queue'.
And then run fuzzer with the queue. ./fuzzer queue
.
We will add this explanation into README.md
file.
Thanks!
Closing inactive issue. Feel free to reopen if you have more to discuss.
I would like to ask, how is coverage information calculated. I first executed all the programs and obtained the. framework files for each program separately. Then, llvm-profdata merge --sparse=true -o default.profdata * Last, llvm-cov show -format=html -output-dir=coverage_report -instr-profile=default.profdata ../../../../../exp/assimp/output/libassimp_profile.a
However, the result I received was an error: libassimp_profile.a: Failed to load coverage: No coverage data found
What is the form of the compilation command for each fuzzer, and why is there a situation where the coverage cannot be found. I'm sorry to ask such a silly question. I'm a beginner in fuzzy testing and I hope to get your help.
Hello, may I ask how the testing section in the paper is implemented? I put in a lot of effort but still couldn't get the code coverage in the paper, Here is my testing process:
Assuming that a, b, and c here represent three different drivers under libhtp
llvm-profdata merge --sparse=true a.profraw b.profraw c.profraw -o foo.profdata
llvm-cov report -install-profile=foo.profdata -object a b c
But the code coverage results of most of the libraries I obtained are far from those in the paper How can I deal with these problems? Thanks!