C++: flow through output iterators with user-defined operator= and operator* #4468
Conversation
| this.hasName(["front_inserter", "inserter", "back_inserter"]) and | ||
| this.getNamespace().hasName("std") |
There was a problem hiding this comment.
Why not use hasQualifiedName("std", [...])?
| *i14++ = source(); | ||
| sink(v14); // tainted [NOT DETECTED by IR] |
There was a problem hiding this comment.
I'd like to see some prefix-++ and some -- too.
This reverts commit 28fa266.
|
I've rewritten most of the QL changes. It now works with just the previously existing nodes and some additional flow edges to the return value input for |
jbj
left a comment
jbj
left a comment
There was a problem hiding this comment.
Thanks for making these changes! The code looks much better now that the library-specific details have been moved into the model classes. I have a few more review comments.
Co-authored-by: Jonas Jensen <[email protected]>
Accept test changes for unnamed elements
|
I've started https://jenkins.internal.semmle.com/job/Changes/job/CPP-Differences/1511/. Feel free to self-merge if there are no performance or result regressions. |
|
@rdmarsh2 The slowdowns on Linux and Kamailio are too big to ignore, so please investigate. They occur in queries that don't use data flow, so perhaps they're inside the models library. |
Resolve merge conflict in tests
jbj
left a comment
jbj
left a comment
There was a problem hiding this comment.
The re-run of CPP-Differences doesn't have the same performance outliers, so perhaps it was just noise.
This PR adds general flow through
operator*inAddressFlow.qllandFlowVar.qll, and adds iterator-specific flow throughoperator=inDataFlowUtil.qll. Those combine to support flow through iterators returned by thestd::back_inserterfunction.A few existing tests are affected by the
operator*changes. I think those are mostly positive.