working in a big c++ codebase is learning to accept that "yeah that memory just randomly corrupts sometimes?" is a valid resolution to a bug
4
8
2
55
It's one of those things that's so bad that it might wrap around back to good, in the sense that it's probably "just" hardware failures
1
3
Not sure we can say "probably", but there's certainly a nonzero expected rate of hardware-related crashes for hashtables used that heavily.
1
... & that are that large. Given that we did eventually find real bugs around the old stuff, this may be real bugs too.
1
Replying to @davidbaron @Gankro
But I don't actually remember if any of the real bugs we found dented the crash rates. (Pretty sure I never checked for most recent round.)

Oct 18, 2017 · 4:05 AM UTC

1
1
Replying to @davidbaron
I'm willing to go up to "probably" at this point given all the data I've seen, my own testing and review, and the nature of the code.
2