Best way to cast unsigned int to int?

I have an unsigned int variable that I want to cast to an int, as it may store a -1 value.
What is the best way to do this?

I tried this:

Code:

---

lastCharIndex_ui = static_cast<int>(lastCharIndex_ui);

---

as well as:

Code:

---

(int)lastCharIndex_ui;

---

and other things as well, but it doesn't appear to work since the compiler still complains about the comparison between a unsigned int and a signed int expression.

Why not post a larger snippet of what you tried, including the portion that the compiler warned about?

Quote:

---

Originally Posted by laserlight

Why not post a larger snippet of what you tried, including the portion that the compiler warned about?

---

This is the only relevant code:

Code:

---

lastCharIndex_ui = static_cast<int>(lastCharIndex_ui); //cast the unsigned int to an int since it may now store a -1 value
//get the index of the last character of the current enumerator:
lastCharIndex_ui = findIndexOfLastChar(filestreamInBuffer_s.c_str(), currentEnumeratorName_s.c_str());

if (lastCharIndex_ui == -1) {
    return false; //since something funny occurred...
}

---

The bolded expression is what the compiler complains about.

Well, duh. You're only converting the contents of lastCharIndex_ui to a signed int, but the type itself is still an unsigned type.

Oh, then that cast achieved nothing. lastCharIndex_ui is presumably an unsigned int, and of course it remains an unsigned int.

Rather, what you can do is:

Code:

---

//get the index of the last character of the current enumerator:
lastCharIndex_ui = findIndexOfLastChar(filestreamInBuffer_s.c_str(), currentEnumeratorName_s.c_str());

if (lastCharIndex_ui == static_cast<unsigned int>(-1)) {
    return false; //since something funny occurred...
}

---

~0U is an unsigned number where all the bits are set. You could replace -1 with that instead.

Quote:

---

Originally Posted by Elysia

Well, duh. You're only converting the contents of lastCharIndex_ui to a signed int, but the type itself is still an unsigned type.

---

Well, how do I change the type then?

Quote:

---

Originally Posted by laserlight

Oh, then that cast achieved nothing. lastCharIndex_ui is presumably an unsigned int, and of course it remains an unsigned int.

Rather, what you can do is:

Code:

---

//get the index of the last character of the current enumerator:
lastCharIndex_ui = findIndexOfLastChar(filestreamInBuffer_s.c_str(), currentEnumeratorName_s.c_str());

if (lastCharIndex_ui == static_cast<unsigned int>(-1)) {
    return false; //since something funny occurred...
}

---

---

Somehow that doesn't seem right...
Since the cast in your code is after where the variable might be assigned a -1 value (i.e. because of the return value of findIndexOfLastChar), rather than before, then it seems like casting a -1 to an unsigned int (which value I have no idea what it would be, btw) and comparing it with whatever value was stored in the unsigned int variable "lastCharIndex_ui" (since it couldn't have been a -1 seeing as its an unsigned int) makes no sense at all.
So I think I have the right idea of doing the cast before the function call which might return a -1 value. I just don't know how to change the type of the variable instead of casting the value stored in the variable to a int type.

Quote:

---

Originally Posted by iMalc

~0U is an unsigned number where all the bits are set. You could replace -1 with that instead.

---

Yeah, but is that what the unsigned int variable lastCharIndex_ui would be given too if the function returned a -1? Otherwise the comparison wouldn't work.

Quote:

---

Originally Posted by Programmer_P

Since the cast in your code is after where the variable might be assigned a -1 value (i.e. because of the return value of findIndexOfLastChar), rather than before, then it seems like casting a -1 to an unsigned int (which value I have no idea what it would be, btw) and comparing it with whatever value was stored in the unsigned int variable "lastCharIndex_ui" (since it couldn't have been a -1 seeing as its an unsigned int) makes no sense at all.

---

Ah, but conversion from signed to unsigned is well defined. For -1, the result will be the maximum value for the unsigned integer type, which is why iMalc's suggestion will work (although it is arguably a little less readable: most readable might be something like std::string::npos, which is -1 cast to std::string::size_type, which is an unsigned integer type).

Quote:

---

Originally Posted by laserlight

Ah, but conversion from signed to unsigned is well defined. For -1, the result will be the maximum value for the unsigned integer type, which is why iMalc's suggestion will work (although it is arguably a little less readable: most readable might be something like std::string::npos, which is -1 cast to std::string::size_type, which is an unsigned integer type).

---

And what value is the maximum value for the unsigned integer type? I'm guessing its probably not equilivant to a -1... :D

Quote:

---

Originally Posted by Programmer_P

Well, how do I change the type then?

---

You cannot change the type of an existing variable! That would break the foundations C++ was built upon.

Quote:

---

Somehow that doesn't seem right...
Since the cast in your code is after where the variable might be assigned a -1 value (i.e. because of the return value of findIndexOfLastChar), rather than before, then it seems like casting a -1 to an unsigned int (which value I have no idea what it would be, btw) and comparing it with whatever value was stored in the unsigned int variable "lastCharIndex_ui" (since it couldn't have been a -1 seeing as its an unsigned int) makes no sense at all.

---

It is right. The signed value returned will be converted to an unsigned value.
Logic then tells me that if we take that same signed number and convert it to unsigned, it should be the same as what was assigned to that unsigned number, yes?
That is what this logic builds upon and is perfectly defined and right.

Quote:

---

So I think I have the right idea of doing the cast before the function call which might return a -1 value. I just don't know how to change the type of the variable instead of casting the value stored in the variable to a int type.

---

You can't and shouldn't.

Quote:

---

Originally Posted by Programmer_P

And what value is the maximum value for the unsigned integer type? I'm guessing its probably not equilivant to a -1... :D

---

Assuming 4 bytes, 0xFFFFFFFF, or a little over 2 * 10^9.

Quote:

---

Originally Posted by Programmer_P

And what value is the maximum value for the unsigned integer type?

---

UINT_MAX, or if you prefer, std::numeric_limits<unsigned int>::max().

Quote:

---

Originally Posted by Programmer_P

I'm guessing its probably not equilivant to a -1.

---

Obviously :)

Quote:

---

Originally Posted by Elysia

It is right. The signed value returned will be converted to an unsigned value.
Logic then tells me that if we take that same signed number and convert it to unsigned, it should be the same as what was assigned to that unsigned number, yes?
That is what this logic builds upon and is perfectly defined and right.

---

Ahh...ok. I get it. That does make sense. So essentially, we've got two -1 values that have become the same something else (neither of which are actually -1). I got it.
Ok, that's what I'll do then.