C什么是long unsigned int的简短形式

Question

when compiling my program with GCC I get the following warning:

format ‘%d’ expects type ‘int’, but argument 2 has type ‘long unsigned int

Now just by playing around I realize %lo fixes the warning. However I don't really understand what I am doing.

Is there a naming convention to get the short form of a type? For example, int is %d, why??

Thanks!

Answer 1

long unsigned int = unsigned long

You use %lu to print those.

And similarly, %llu is for unsigned long long .

---

%d    -   int
%u    -   unsigned
%ld   -   long
%lld  -   long long
%lu   -   unsigned long
%llu  -   unsigned long long

Answer 2

No, there is convention, you simply need to look at the documentation .

And the formatting specifiers are not a 1:1 mapping to the type, they control the output format (hence the name). There can be many output formats for a single input value type.

For instance, %d is "decimal", since it prints an integer as a decimal number. There is also %i ("integer") that does exactly the same thing.

For unsigned int values, you have both %x (print in hexadecimal) and %u (print in octal).

Answer 3

The definition of int, long, etc are specific to the target. If the size of int and long match then a

printf("%d",var); 

wont complain, but if long is say 64 bits and int is 32 bits as an example, then you will get the warning/error you describe. One of two solutions:

printf("%ld",var);
or
printf("%d",(int)var);

For the latter of course you have to insure that the compiler associates int with the %d size, if you get yet another warning then adjust accordingly.

EDIT:

The compiler is trying to help you out, by worrying about C library stuff which is not really the business of the compiler. printf() uses a variable number of arguments, which hopefully you properly matched to your format string. When printf sees a %d on say a 32 bit system it likely will only grab the next 32 bit argument. But if you had put a 64 bit integer in as a parameter it may grab one half of that integer, and use the other half as the next item in the format string. For example

unsigned long ul;
float f;

f=3.4;
ul=0x3F9DF3B612345678;
...
printf("%X %f\n",ul,f);

Depending on your system, endianess, etc, a 32 bit system you should not at all be surprised if the above code produced this output:

12345678 1.234000

because that is what you told it to to. you told it to take the lower 32 bits of ul and print that as hex (%X) and the upper 32 bits of ul and print that as float (%f) and putting f in the function call was a waste you didnt provide any formatting to use the floating point value.

Now depending on the compiler on the target system on a particular day you may have the above code work as desired, take a system/compiler where unsigned long is interpreted as a 32 bit and %X is interpreted as 32 bit, then you get a warning about the 64 bit assignment but the printf makes a little bit more sense.

Because of this pain, compilers like gcc bother to try to make your life better by assuming that when you use a function called printf() you are using the standard C one and they parse through your format string looking for these types of common mistakes.

Answer 4

However I don't really understand what I am doing.

What you're doing is telling the printf function how to display the data that you provide following the format string. You'll probably find that %lo doesn't really print the results you expect -- it prints the right data, but in octal (base 8) format.

Is there a naming convention to get the short form of a type? For example, int is %d, why??

The "short form" is called a format specifier. There's no convention that lets you derive the appropriate printf format specifier. You just have to look them up in an appropriate reference .

Answer 5

%lo = unsigned long integer printed in octal digits. yes, docs are a good place to start.. but somethings have a pattern, like o for octal, x for hex, l for anything long etc. Getting used to many such format specifiers will help you get the pattern wherever there is.