search

google / oss-fuzz-gen

LLM powered fuzzing via OSS-Fuzz.
Apache License 2.0
802 stars 93 forks source link

"Project not in OSS-Fuzz (likely only contains a project.yaml file)" when generate a benchmark-yaml. #498

Open aT0ngMu opened 1 month ago

aT0ngMu commented 1 month ago

oss-fuzz-gen is a great job, thank you for making your code open source!!!

When I execute this command

python3 runner.py -o oss-fuzz-10 -t 15 -i https://github.com/memononen/nanosvg -m openai,

it generated a project temp-project-4-cov-test-fuzz-build-0-fuzzgen-0temp-project-4-cov-test-fuzz-build-0-fuzzgen-0.zip.

When I execute this command

run_all_experiments.py -l gpt-3.5-turbo -w bench-projects/gen1 -of bench-projects/oss-fuzz -g far-reach-low-coverage,low-cov-with-fuzz-keyword,easy-params-far-reach,optimal-targets -gp temp-project-4-cov-test-fuzz-build-0-fuzzgen-0,

it will send a request to the api

https://introspector.oss-fuzz.com/api/far-reach-but-low-coverage?project=temp-project-4-cov-test-fuzz-build-0-fuzzgen-0&exclude-static-functions=True&only-referenced-functions=False&only-with-header-file-declaration=True.

And I get the response of Project not in OSS-Fuzz (likely only contains a project.yaml file).

I have a few questions to ask:

  1. How to get the right response?
  2. I didn't find any code about request processing for this API https://introspector.oss-fuzz.com/api/far-reach-but-low-coveragein introspectot's, is this part of the code somewhere else?
DavidKorczynski commented 1 month ago

When I execute this command

run_all_experiments.py -l gpt-3.5-turbo -w bench-projects/gen1 -of bench-projects/oss-fuzz -g far-reach-low-coverage,low-cov-with-fuzz-keyword,easy-params-far-reach,optimal-targets -gp temp-project-4-cov-test-fuzz-build-0-fuzzgen-0 ,

You can't run the targets generated from https://github.com/google/oss-fuzz-gen/tree/main/experimental/c-cpp using run_all_experiments.py just yet. At the moment they're a bit detached, but I have a fix in mind to solve this issue, give me a few days and it should be ready. You can follow along here where I'll reference the interim updates as well including when you can take the output from runner.py and use it in run_all_experiments.py: https://github.com/google/oss-fuzz-gen/issues/450

I didn't find any code about request processing for this API https://introspector.oss-fuzz.com/api/far-reach-but-low-coveragein introspectot's, is this part of the code somewhere else?

https://introspector.oss-fuzz.com/api#api-far-reach-but-low-coverage