Open DonggeLiu opened 2 months ago
Also, we did not document the exact version of matplotlib
used in fuzzbench in requirements.txt.
Related: https://github.com/google/fuzzbench/pull/1982#issuecomment-2120433754
Adding a solution discovered by @DanBlackwell: https://github.com/google/fuzzbench/pull/1982#issuecomment-2120546061
Just to add a note here; trying to install matplotlib==2.2.5
may be tricky on newer Linux setups, as the pip wheel tries to use platform.linux_distribution
which was removed in python3.8 (according to https://stackoverflow.com/a/64106589)
Hey @DonggeLiu, is this been worked on or can I work on this? Can you please assign this issue to me?
@gustavogaldinoo: Are you working on this by any chance? If not, would now be a good time for @ChaitanyaYeole02 to help?
@DonggeLiu, if there are any tasks, that you think can be taken by me. Please let me know. I am still a beginner with fuzzbench library.
Just a note that in the meantime if anyone wants to run generate_report.py
locally, the following docker setup worked for me:
# load up a docker image that has python3.10; mount fuzzbench directory inside
docker run -v $(pwd):/tmp/ -it ubuntu:22.04
# now you're in docker, let's install the dependencies
apt update
apt install -y python3.10-dev python3.10-venv build-essential libpq-dev
cd /tmp/
apt install -y pip
pip install pyyaml google-cloud-storage google-cloud-error-reporting
make install-dependencies
source .venv/bin/activate
pip install pandas==1.5.2 clusterfuzz scikit_posthocs Orange3 sqlalchemy jinja2 psycopg2
# now we can run the generate_report.py script, note: --from-cached-data required!
PYTHONPATH=. python3 analysis/generate_report.py $EXPERIMENT_NAME -d $REPORT_DATA_DIR --from-cached-data --benchmarks bloaty_fuzz_target --fuzzers libafl aflplusplus
Deprecations: https://github.com/google/fuzzbench/actions/runs/8752698488/job/24020848759?pr=1956#step:8:391
Conflicts: