Looks like SQLite used the sanitizers, and a new tool still found plenty UB.
I believe the SQLite auther verified the compiler output for the UB found, and saw it had no impact in practice, so there was no critical bug yet. Then he fixed the UB anyhow.
Last paragraph was from vague memories, not from the article, BTW. I would welcome better links.
https://blog.regehr.org/archives/1292
Looks like SQLite used the sanitizers, and a new tool still found plenty UB.
I believe the SQLite auther verified the compiler output for the UB found, and saw it had no impact in practice, so there was no critical bug yet. Then he fixed the UB anyhow.
Last paragraph was from vague memories, not from the article, BTW. I would welcome better links.