Charusso marked 6 inline comments as done.
Charusso added a comment.
Thanks for the review! The build-bots will fire with that `QualType` fix (1028
TU on its own). I will look into the exploded-graph-rewriter.py after GSoC to
fix every stuff like that patch and also invoke my HTML simplification idea.
================
Comment at: clang/lib/StaticAnalyzer/Checkers/CastValueChecker.cpp:83-93
+static QualType getRecordType(QualType Ty) {
+ Ty = Ty.getCanonicalType();
+
+ if (Ty->isPointerType())
+ return getRecordType(Ty->getPointeeType());
+ if (Ty->isReferenceType())
----------------
NoQ wrote:
> Charusso wrote:
> > NoQ wrote:
> > > NoQ wrote:
> > > > Most of the time we should know exactly how many pointer/reference
> > > > types we need to unwrap. I really prefer we hard-assert this knowledge
> > > > instead of blindly unwrapping as many pointer/reference types as we
> > > > want. Because every time we have an unexpected number of unwraps it's
> > > > an indication that something went horribly wrong. So it's good to have
> > > > the extra sanity checking.
> > > I think this one is still forgotten. Maybe a FIXME?
> > Now I pattern-match that rule in the `evalCall()`, therefore we cannot try
> > to evaluate nested references. Would you like to see more checks?
> Oh, i see, you removed the recursion. Great.
I was dumb enough to do not publish my comment six times, sorry.
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D66325/new/
https://reviews.llvm.org/D66325
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
