Question

I have the following code segment:

std::list<reply_t> l = m[index];
for (std::list<reply_t>::iterator it = l.begin(); it != l.end(); it++) {
    // do something
}

and

std::list<reply_t> *l = &(m[index]);
for (std::list<reply_t>::iterator it = l->begin(); it != l->end(); it++) {
    // do something
}

m is just a map with int index and list as values, ie std:map<int, std::list<reply_t> >.

The two versions ONLY differ in how the list is being referred to, ie one by pointer one by object. Everything else in my code is exactly identical. HOWEVER, when I run with these two versions, version 1 consistently fails tests and version 2 consistently succeeds. The code is used in a multi threaded context, if that helps.

I’m new to c++ and this is just completely bizarre to me. Anyone with an possible explanation?

Answer 1

The two examples differ in that the first creates a copy of the list returned by m[index], then iterates over that copy.

To match the second version, you need to use a reference to the list. Try this instead:

std::list<reply_t>& l = m[index]; // notice extra '&' character.