typedef a primitive type

This topic is 1705 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

Recommended Posts

I often see typedefs on a primitive type.

For example

    typedef signed   char int8;
typedef unsigned char uint8;

typedef signed   short int16;
typedef unsigned short uint16;


What's the point of this?

Share on other sites

Just as an addition, I always prefer using the C99 header stdint.h (also available as cstdint in C++) and the types uint8_t and similar and not handling this mess myself. In one company I worked for, this was not done and during a switch to 64 bit I had to update a whole bunch of header files with lots of #ifdefs to take care of this. Had the project used stdint, it would have saved me a whole day of debugging to find the problem and fix it.

I agree that stdint.h should be used, but before that was available I used to write a little program that would check the size of various integer types (using sizeof) and produce the text of a header file with those typedefs. The makefile knew to compile and run this program to generate the header file. I used that without problems for years.

#include <iostream>
#include <string>
#include <climits>
#include <cstdlib>

std::string find_type(int size) {
if (CHAR_BIT * sizeof(char) == size) return "char";
if (CHAR_BIT * sizeof(short) == size) return "short";
if (CHAR_BIT * sizeof(int) == size) return "int";
if (CHAR_BIT * sizeof(long) == size) return "long";
if (CHAR_BIT * sizeof(long long) == size) return "long long"; // OK in gcc
std::cerr << "ERROR: I couldn't find a " << size << "-bit type!\n";
std::exit(1);
}

void define_signed_and_unsigned(int size) {
std::cout << "typedef signed " << find_type(size) << " int" << size << ";\n";
std::cout << "typedef unsigned " << find_type(size) << " uint" << size << ";\n";
}

int main() {
define_signed_and_unsigned(8);
define_signed_and_unsigned(16);
define_signed_and_unsigned(32);
define_signed_and_unsigned(64);
}


Edited by Álvaro

Share on other sites

In this case, others have already explained the specific use here, but more generally typedefs of primitive (and non-primitive) types can be used to provide additional context about their use.

Primitive types convey only two things: Their size (which is platform specific), and their format (which is also platform specific, though less obviously so). But it doesn't convey a purpose or intent as to what it holds. For example, 'typedef float velocity;' and 'typedef float acceleration;' give you added information about what instances of these types (ought to) hold, that would otherwise be described only by the variable name. Now, perhaps unfortunately, typedefs aren't strong -- that is, they don't create new distinct types, they just allow you to call a type by another name -- so you can still assign a 'velocity' variable to an 'acceleration' variable or a plain old float. But the point is that it creates a logical distinction between them, even if its not enforced by the compiler.

Another practical advantage is that if you decide that float is insufficient for your 'velocity' or 'acceleration' types, then you can easily redefine them to be of type 'double' in only a single place, rather than hunting through every single float in your code to determine whether it holds a velocity or acceleration. In this way it can save you time and avoid subtle bugs that arise from missing one of the things you should have changed.

Share on other sites

Thank you all. I now understand it.

So a typedef header cannot be cross platform. Am I right?

Share on other sites

In any case, yes, they can be cross-platform.  That is the point.

L. Spiro

Share on other sites

Primitive types are *not* cross-platform in terms of the number of bits they contain or their representation -- IIRC, paraphrasing the standard, it says only something along the lines of "a 'char' is the smallest addressable unit of storage; a 'short' is at least as big as a 'char'; an 'int' is at least as big as a 'short'; ..." The standard doesn't say that an 'int' is exactly 32 bits (although it is in many platforms), I don't even believe it says that signed numbers must be represented as two's compliment form.

The idea of a typedef header like stdint.h is to define a type (the typedef) which *is* the same number of bits across many platforms, by changing the underlying primitive types appropriately on a platform-specific basis. In other words, the idea is that uint32_t is always a 32bit unsigned integer on any platform, but the underlying primitive type might be different on a 32bit x86 machine running Windows than it is on a 64bit MIPS machine running Unix.

Edited by Ravyne

Share on other sites

Thank you all. I now understand it.

So a typedef header cannot be cross platform. Am I right?

You make these headers be cross platform like:

#if defined(PLATFORM_ONE)
typedef foo int32;
#elif defined(PLATFORM_TWO)
typedef bar int32;
#else
#error "this platform not supported"'
#endif

Share on other sites

Thank you all. I now understand it.

So a typedef header cannot be cross platform. Am I right?

You make these headers be cross platform like:

#if defined(PLATFORM_ONE)
typedef foo int32;
#elif defined(PLATFORM_TWO)
typedef bar int32;
#else
#error "this platform not supported"'
#endif

Oh I see, I didn't think of conditional compilation

Share on other sites

Nice list of the various defines.

Share on other sites

Now, perhaps unfortunately, typedefs aren't strong

Ada does this. It's a nice safety rope but it can also get quite messy. Velocity equals acceleration times time? Nope, not without converting everything to a common base type. It is nice to know that you cannot mix different types by accident, but it also means you cannot mix them intentionally without writing more code. Edited by rnlf

Share on other sites

Ada does this. It's a nice safety rope but it can also get quite messy. Velocity equals acceleration times time? Nope, not without converting everything to a common base type. It is nice to know that you cannot mix different types by accident, but it also means you cannot mix them intentionally without writing more code.

Now I want to see a language that associates units with variables and intelligently checks that they remain valid throughout statements

Share on other sites

And if you don't mind that it will probably your compile times and might have issues with compilers where certain types don't exist, you can use black template magic to get around ifdefs and having to know the different platforms and type sizes in advance.

Essentially you simply stuff all signed, unsigned and floating point types into a type list and let the template recursively go through this list until it finds one with the requested sizeof. Though this is only going to be of any help if a) you can't use the (c)stdint header and b) really don't want to worry about all the potential platforms. It's also going to be messy on exotic machines, where one "byte" can be more than 8bit.


template<bool condition, class A, class B> struct SelectT { typedef A Type; };
template<class A, class B> struct SelectT<false, A, B> { typedef B Type; };

struct NullType;

template<typename A = NullType, typename B = NullType, typename C = NullType,
typename D = NullType, typename E = NullType, typename F = NullType>
struct TypeList
{
typedef TypeList<B,C,D,E,F> Tail;
};

template<size_t size, typename list>
struct TypeOfSize
{
typedef typename SelectT
typename TypeOfSize<size, typename list::Tail>::type
>::Type type;
};

template<size_t size> struct TypeOfSize<size, TypeList<> > { typedef NullType type; };

typedef TypeList<char, short, int, long, long long> SignedType;
typedef TypeList<unsigned char, unsigned short, unsigned int, unsigned long, unsigned long long> UnsignedType;
typedef TypeList<float, double, long double> FloatingType;

typedef TypeOfSize<1, SignedType>::type int8;
typedef TypeOfSize<2, SignedType>::type int16;
typedef TypeOfSize<4, SignedType>::type int32;
typedef TypeOfSize<8, SignedType>::type int64;

typedef TypeOfSize<1, UnsignedType>::type uint8;
typedef TypeOfSize<2, UnsignedType>::type uint16;
typedef TypeOfSize<4, UnsignedType>::type uint32;
typedef TypeOfSize<8, UnsignedType>::type uint64;

typedef TypeOfSize<1, FloatingType>::type float8;
typedef TypeOfSize<2, FloatingType>::type float16;
typedef TypeOfSize<4, FloatingType>::type float32;
typedef TypeOfSize<8, FloatingType>::type float64;


Share on other sites

Bacterius, on 23 May 2013 - 06:12, said:

rnlf, on 23 May 2013 - 04:04, said:
Ada does this. It's a nice safety rope but it can also get quite messy. Velocity equals acceleration times time? Nope, not without converting everything to a common base type. It is nice to know that you cannot mix different types by accident, but it also means you cannot mix them intentionally without writing more code.

Now I want to see a language that associates units with variables and intelligently checks that they remain valid throughout statements

C++ can do that. You can use a template class that has integer parameters indicating the power of kilogram, meter and second. All the class does is wrap a floating-point number. Now define operators that deal with types appropriately.

Or you can use Boost.Units.

EDIT: Fixed the link. This editor and I don't get along... Edited by Álvaro

Share on other sites

Uuhh... nice idea. Haven't heard of this one before, nor of Boost.Units. I will definitely look into this.

Edited by rnlf

Share on other sites

Ada does this. It's a nice safety rope but it can also get quite messy. Velocity equals acceleration times time? Nope, not without converting everything to a common base type. It is nice to know that you cannot mix different types by accident, but it also means you cannot mix them intentionally without writing more code.

Now I want to see a language that associates units with variables and intelligently checks that they remain valid throughout statements

As others have said, its possible to put together a pretty good system in C++ using templates, but some languages do have this kind of notion built in. Haskell and, I think, F# are two examples. I believe its called type annotation or somesuch.

Share on other sites

Now, perhaps unfortunately, typedefs aren't strong

Ada does this. It's a nice safety rope but it can also get quite messy. Velocity equals acceleration times time? Nope, not without converting everything to a common base type. It is nice to know that you cannot mix different types by accident, but it also means you cannot mix them intentionally without writing more code.

Yeah, it does take effort to define the proper relationships, and it can grow into a sort of combinatorial problem. The approach would be to overload operators and/or create classes that represent various units. AFAIK, this is basically what the C++ template libraries do, except that they use template magic to generate the code, perhaps type traits as well in their implementation.

Its a mixed bag whether strong or weak typedefs are preferable, personally I think it would have been nice to have both options: typedef as strong, typealias as weak. C/C++ typedefs are weak mostly because C's notion of type equality is rooted in its physical properties (size and format), rather than its logical ones (its intent). That's a perfectly reasonable choice for a systems language, but its unfortunate that C++ didn't introduce a way to make a new strong type other than classes.

Share on other sites

I actually like this a lot. I am using C++ for more than 10 years now and I'm still amazed by its expressiveness.

Share on other sites

I think that thinking of "cross-platform" with these typedefs is only half the story, and possibly even the less common reason. It is of course indeed the case with the "precise size" typedefs in stdint.h.

For the most part, the like typedefs, are much more about code maintainability and making code forward compatible, not so much about "cross-platform".

Cross-platform, is something that you can give a crap about, 95% of the time. Unless you have to layout a structure so it exactly fits some given structure or you make assumptions such as sizeof(int) == sizeof(void*), you usually couldn't care less about portability. An integer is an integer, and most of the time there's no noticeable difference (overflow due to an odd too-small-size integer type is most unlikely nowadays, this is not the 1970s any more).

On the other hand, consider that you wrote 200,000 lines of code that assume some function (let's say posix_fadvise, or glBufferData) takes an integer argument (or, maybe even worse, returns one: say CreateWindow). Incidentially, your integers are 32bit values, but that's big enough for everyone, isn't it. Integers are the same size as a pointer, too, how lucky is that!

One day the maintainers of that function realize that hard disks can be quite a bit larger than 4GiB these days, and people actually have files larger than 4GiB, too (or vertex buffers >4GiB). So they change the function to take a 64bit integer. That's 200,000 lines of code down the drain for you. The maintainers of the other function realize that integers are not necessarily the same size as pointers (did someone say SetWindowLong?).

Now, the clever maintainer would have used off_t for posix_fadvise, and HWND for CreateWindow, which does not look like there is much of an advantage. Indeed it actually looks quite stupid, so many different types for only the same thing.

However, if one day, it is decided that one has to change the typedef, your 200,000 lines of code just need to be recompiled once, and they will work. No manual editing every occurrence needed.

Edited by samoth

• 11
• 9
• 11
• 9
• 12