Community
Participate
Working Groups
Build Identifier: 8.0.0.201109151620 Code Analysis gives the warning: Catching by reference is recommended 'std::bad_cast', even when there isn't a parameter name to reference. Example: #include <iostream> #include <exception> class B {}; class A: public B {}; int main() { A obj; try { B& other = dynamic_cast<B&>(obj); } catch (std::bad_cast) { std::cout << "Error casting A to B"; } return 0; } I don't think using std::bad_cast& would give any benefit here. Reproducible: Always Steps to Reproduce: 1. See example code
I think it might still make sense to catch by reference in this case. Try the code below and you'll see that an unnecessary copy is made: g++ -Wall test_exception.cpp && ./a.out catch(A&) instead and no copy is made. Before closing this issue, could one of the C++ gurus please comment if there are valid use cases for not catching by reference? #include <iostream> #include <string> struct A { A(void) { std::cout << "constructing an A" << std::endl; s = "content of A"; } A(const A& other) { std::cout << "copying an A" << std::endl; s = other.s; } std::string s; }; int main() { std::cout << "main" << std::endl; try { throw A(); } catch(A) { std::cout << "catch" << std::endl; } std::cout << "end" << std::endl; return 0; }
I don't know of any valid use cases for catching by reference. I think the warning is fine as it is.
(In reply to comment #2) > I don't know of any valid use cases for catching by reference. I think the > warning is fine as it is. I think you meant for *not* catching by reference?
(In reply to comment #3) > (In reply to comment #2) > > I don't know of any valid use cases for catching by reference. I think the > > warning is fine as it is. > > I think you meant for *not* catching by reference? Yes, sorry. I meant "I don't know of any valid use cases for catching by value."
Closing per comment 4.
