Hi all, according to the following issue:
Continuing the discussion from Uninitialized value used but not pointed out:
Do analysis cross translation units possible?
Or there is always a limitation there?
Hi all, according to the following issue:
Continuing the discussion from Uninitialized value used but not pointed out:
Do analysis cross translation units possible?
Or there is always a limitation there?
Hi @deron.chen,
Cross translation-units analysis is possible for some rules. For example, the rule about unused functions.
Yet cross translation-units analysis is considerably more difficult to perform for path-sensitive issues, like the one you mentioned. Especially at scale.
To answer your first question: At the moment our analysis for C, C++, and Objective-C does not support path-sensitive issues across translation units. It is possible for some simpler (AST-based) rules.
To answer your second question: This limitation might be lifted one day. We are considering ways to extend path-sensitive analysis across translation units.
Let me know if the answer is not clear or you have further questions.
P.S. I apologize for taking so long to respond.
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.