Change what you get warnings for in Visual studio

I am starting to get tired of some of the warnings I get when I compile. Is there a way to make the compiler not print out so many warnings?

I dont want warnings when converting from float to int or when I compare an unsigned int to a signed int.

Or am I doing something wrong and should fix my code so I dont get the warnings at all?

Well generally you actually want to take note of those warnings. They could actually help you out with a logic error some day.

I dont want to remove all warnings, just the once who are stupid(or I think is stupid)

Like I get a warning for this:
for(int i = 0; i < someSTDvector.size(); i++)

Saying that I compare a signed and a unsigned int. I cant see why that would be a problem?

It is a problem and will bite you someday if you don't heed it. Make both types the same type. In this case, you can make i unsigned. Size can never be negative anyway.
Take heed to these warnings and know how to fix them instead of suppressing them.

You should cast a float if you convert it to int. That way, the compiler will accept it happily.

You should not compare unsigned types with signed types, as that can lead to "interesting problems" if the unsigned value is in the "negative" range of the signed number. E.g. a short of 40000, is that bigger or smaller than 32767?

Warnings are there, mainly, because the compiler programmer thinks it's something that is likely to cause a problem.

For example, a float can be much bigger than an int (approx 1E10 vs 1E37), so it's valid for the compiler to warn about it. And as I mentioned above, mixing signed and unsigned values is also causing "range problems".

--
Mats

Quote:

---

Like I get a warning for this:
for(int i = 0; i < someSTDvector.size(); i++)

---

It should be:

Code:

---

for (std::vector<T>::size_type i = 0; i < someSTDvector.size(); ++i)

---

where T is the value type for someSTDvector. Making i an unsigned int as Elysia suggested should also work, but pedantically speaking is not correct.

Oh yes, of course, it's not always a good thing to mix types (especially not without a cast). If something returns a DWORD, then use a variable of DWORD to store that and not uint32_t or something else.
And, as laserlight mentions, use std::vector<T>::size_type for the vector since that's what it returns. It guarantees that there will be no problem with the type, wherever the code will compile.

Quote:

---

Originally Posted by laserlight

It should be:

Code:

---

for (std::vector<T>::size_type i = 0; i < someSTDvector.size(); ++i)

---

where T is the value type for someSTDvector. Making i an unsigned int as Elysia suggested should also work, but pedantically speaking is not correct.

---

And if you use it often, you could do:

Code:

---

typedef std::vector<T>::size_type STDsize;

---

--
Mats

Point taken. I will change my code, not Visual Studio

Good student :D