Question

I would like the best routine the community can devise which will provide a version number string using only 32 bits. The version number will contain three parts: major, minor, and build number. Bonus karma points for answers that include Java source.

Update 1

Ok, so to clear up some confusion the major and minor fields should be represented with one 8-bit byte each and the build number would use the remaining bits.
Major: 8 bits
Minor: 8 bits
Build: 16 bits
The byte order will be little endian.

The tags are to denote that an answer in java and / or c++ would be preferred over python, ruby, etc..

Update 2

Example: take an integer (4 bytes) like 101020 and run it through an algormithmic routine to produce an output string like so: 0.1.2

Java Solution

    public static final int versionToInt(int major, int minor, int build) {
        int ver = (major << 24) | (minor << 16) | build;
    return ver;
    }

    public static final String versionToString(int ver) {
        return String.format("%s.%s.%s", ((ver & 0xff000000) >> 24), ((ver & 0x00ff0000) >> 16), (ver & 0x0000ffff));
    }

Answer 1

Java:

class Version
{
    public char major;
    public char minor;
    public short build;

    // And the rest of your class definition ...
}

Edit: So you want to do it with a bit shifts and store it in a 32-bit integer (even though using a class or struct would still use the same amount of storage space)?

An example function in C would be as follows

uint32_t MakeVersionNumber( int8_t major, int8_t minor, int16_t build )
{
   return ((uint32_t)major << 24) | ((uint32_t)minor << 16) | ((uint32_t)build);
}

Edit 2:

Riiight I think i get ya. You have a 32-bit number and you want to get a string. So here is an example in C++:

std::string MakeVersionString( unsigned int ver )
{
    std::stringstream ss;
    ss << ((ver & 0xff000000) >> 24) << "." << ((ver & 0x00ff0000) >> 16) << "." << (ver & 0x0000ffff);
    return ss.str(); 
}