Using stl's vector.size() with printf

Hi guys,

I'm using this code:

Code:

---

vector<int> v;
v.push_back(123);
printf("size: %u\n", v.size());

---

And it retuns the result you would expect.
But if you compile it you get this warning:

Code:

---

warning #181: argument is incompatible with corresponding format string conversion
                        printf("size: %u\n", v.size());
                                            ^

---

Is there any better solution than just writing that

Code:

---

(int) v.size()

---

Can I use a different format string?

That does not make sense: v.size() returns a std::vector<int>::size_type, which is an unsigned integer type. %u should be an acceptable format specifier.

I know. But I just saw that the error only appears if I use Intel's icc.
With g++ it works fine.

hmm... if it is feasible, consider using C++ I/O streams instead.

yeah, would be a workaround. But my output line is quite complex so I would prefer a printf version.
This warning annoys me since years and I want to know why it doesn't work like that.. in principle.
Is it just an error by Intel's Compiler crew?

Quote:

---

Originally Posted by manuels

But my output line is quite complex so I would prefer a printf version.

---

You could consider using Boost.Format.

Quote:

---

Originally Posted by manuels

Is it just an error by Intel's Compiler crew?

---

You could ask them, but they may be right: after all %u is meant to work with an unsigned int, but std::vector<int>::size_type is only specified as an unsigned integer type.

I guess you could see if the compiler supports:

Code:

---

printf("size: %zu\n", v.size());

---

But I don't think most C++ compilers support the %z modifier, and I'm not sure if vector<int>::size_type can be treated the same as a size_t.

Quote:

---

Originally Posted by laserlight

That does not make sense: v.size() returns a std::vector<int>::size_type, which is an unsigned integer type. %u should be an acceptable format specifier.

---

unsigned integ[ral|er] type doesn't mean unsigned int type. so %u might be the wrong format specification.

edit: hmm you said the correct way... not really sure what you think now :p

Quote:

---

Originally Posted by robwhit

unsigned integ[ral|er] type doesn't mean unsigned int type. so %u might be the wrong format specification.

---

Read post #6.

What if you try %ul instead? Maybe Intel thinks it's an unsigned long.

Quote:

---

Originally Posted by bithub

I guess you could see if the compiler supports:

Code:

---

printf("size: %zu\n", v.size());

---

But I don't think most C++ compilers support the %z modifier, and I'm not sure if vector<int>::size_type can be treated the same as a size_t.

---

Cool, %zu seems to work with Intel and GNU compiler.

Thank you, folks!

Size is actually a size_t type which can be easily cast to an unsigned int to remove the warning.