You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Yes, I have seen this as well.
It's actually hard to prove that a[i] is initialized using abstract interpretation, since we compute an over-approximation of reachable states. We need to infer that the loop actually iterated over all i from 0 to 9, and not just some of them (you could do i+=2).
I think we just disabled the uninitialized variable analysis in the past, but it was re-enabled for @richardlford's work.
Note that you can use -a to run specific analyses or disable others, e.g -a="*,-uva". See README.md
I think as of right now there's nothing that needs to be done here. The issue here is that one of ikos' defaults tends to throw many false positives. We either leave it as is (close), document it further, or change the default back.
I'm happy closing this as is and re-visiting it if people keep asking about it.
ivanperez-keera
changed the title
Bogus warning 'expression xx might be uninitialized'
Bogus warning expression <xx> might be uninitializedDec 18, 2023
In this program:
a[i-10]
is initialized in the end, but ikos complains that it isn't initialized.Version: 3.1
FreeBSD 13.1
The text was updated successfully, but these errors were encountered: