wasefire
wasefire copied to clipboard
Continuous fuzzing
trafficstars
We could have simple continuous fuzzing by leveraging scheduled workflow and caches. Every week, we run all fuzz targets for 10 minutes each and cache their corpus. Next week, we repeat starting from the cached corpus.
There should be a ci-fuzz action that loads the cache and runs regression testing (i.e. just runs on the corpus). This action should have an equivalent ./scripts/ci-fuzz.sh script.