donat.nagy added a comment.
For me it's an unfortunate surprise that Clang Tidy is traversing the concrete
//template instantiation// (I would've guessed that it's working on the
original template body), but if that's the observed behavior, then this commit
is a clean and straightforward fix.
I'd suggest adding another testcase that looks like
template <typename T>
int branch_clone_in_template(T t) {
if (t) {
return 42;
} else {
return 42;
}
}
int use_template(int x) {
return branch_clone_in_template<int>(x);
}
to clarify the behavior on cases where the checker encounters a "real" branch
clone that happens to be in a template. I think we can/should merge your change
even if it suppresses these true positives (as it's a self-contained change
that replaces a serious issue with a minor limitation), but in that case we
should mark this testcase with a FIXME comment to highlight this low-priority
followup work. (Of course it's even better if you happen to have a quick idea
which ensures that the checker can report "real" branch clones even if they
happen to appear in templates.)
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D151133/new/
https://reviews.llvm.org/D151133
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
