quickjs
quickjs copied to clipboard
OSS-Fuzz targets improvements
I was experimenting with quickjs fuzzing through libFuzzer, when I recognized some inefficiencies. I've uploaded a patch set to OSS-Fuzz to fix them, where I was advised to ask a review from the quickjs maintainers and possibly move the implementation of the fuzz targets into the main codebase (here). So, I'm doing it in this PR.
The patches should be straightforward: the first one is a copy of the current version of the fuzz targets from oss-fuzz, while the others incrementally apply independent fixes/improvements.
If and when you find these commits acceptable, then I'll adapt the infra in oss-fuzz to use the new code. Plus, if requested, I can enable to mirror the found bugs to the GitHub issue tracker as well.
WAT
May I ask for a review from someone?
Gentle ping
@chqrlie Do you want these here or in NG? I might be able to help with either.
@saghul Since only this original repository is registered into oss-fuzz (allowing to publicly keep track of the performance improvement), my first priority is to merge these changes here. But, since libFuzzer can be executed locally too, NG could also gain profit from these changes, so I'm not against merging there, too.
I hope that I addressed or answered all the requests and questions.
LGTM as a first check-in! @chqrlie PTAL!
Very interesting feature. I am waiting for Fabrice's approval to merge this in. Good work!
LGTM as a first check-in! @chqrlie PTAL!
Very interesting feature. I am waiting for Fabrice's approval to merge this in. Good work!
Thank you @renatahodovan for your contribution. Fuzzing is an enticing path to find software flaws. You seem to have an advanced understanding of the domain, do you participate in conferences such as FOSdem in Brussels where I followed several sessions on the subject? I would be interested to follow the progress and discuss research ideas on this.
Best regards
Chqrlie
Hi @chqrlie ! It's always a pleasure to meet people with similar interests! 😊 I haven't attended FOSDEM in person yet, but some of my colleagues have been a few times. Perhaps next time... My next conference target is ISSTA in September. It's a bit further away, but that shouldn't stop us from discussing research ideas. I'm available via email, Gitter, or any other suitable platform.