Thread: the main

I've seen this coding convention in certain books and i was wondering. Even if it does work correctly isn't kinda wrong

Code:

int main(void)
.......
.......
.....
return 0;

if main returns an integer value to the processor to let it know it has compiled succesful, why then do you need void just to say that it does not accept any arguments? doesn't the compiler/processor knows this by default.


while i'm at it:

if "0" is supposed to say it is compiled succesfully and any other integer can be any type of error. why use that convention seeing as how in C if !0 is seen as true. basically as long as its not zero it is true/fine. seems counter intutive to me and confusing. Its used to represent one thing and then it turns around and is used to represent another.


I know I'm a newbie so take it easy on me with your responses.

The 0 does not tell the compiler or processor that it compiled correctly, it tells the OS that it executed and exited normally. There are times when you need to tell the OS that it did not execute and exit normally.

The reason you put void in the () on function calls that don't have any parameters is so that the compiler knows that any call with parameters is wrong.

The argument still stands if some value not equal to 0 be returned should, like the bool value, represent true and logically a succesful exit.

But its not a simple boolean yes or no. The value returned actually means something to the OS. What each value means is unknown to me, however if you wanted to say single to the OS to reboot and you knew that if you return -345 it would do so, then you could return -345 and the OS would reboot.

0 and 1 just happen to be the most command and standard returns.

I realized this but I'm glad you posted. By allowing an int return value you can have many exit conditions with a variety of meanings.

Rofl, I love those questions. Usally its me doing it to other people and not this way around

Edit: To clarify its usally the OS, but any program that invokes another can (and should) use the exit conditions.

By allowing an int return value you can have many exit conditions with a variety of meanings.

There are. I believe they are OS specific. In MS DOS a return of 0 means "Program exited successfully", 1, 2, and 3 (and probably many more) are different error codes which I am unclear of the meaning. Look it up specifically to whichever OS your programs will be running on. Although without some sort of log these return values are useless, as they don't make the OS behave in a different manner.



Function Prototype:

return data type function name(argument data types and variables passed to function);

For an example see this list of what MsiExec.exe returns.

http://msdn.microsoft.com/library/en...rror_codes.asp

Sample batch file for doing a silent install on a remote machine and saving the result to a log file:

Code:

\\PathtoUpdate\update.exe /q:a /r:n

ECHO Computer: %COMPUTERNAME% >> %LOGFILE%
ECHO Update Result: %ERRORLEVEL% >> %LOGFILE%

%ERRORLEVEL% holds the result of the last program to run.

In C, you can return any int value you like from main. However, the only ones which have defined behaviour on all operating systems are:

0,
EXIT_SUCCESS, and
EXIT_FAILURE.

(Those two macros ae defined in stdlib.h)

0 and EXIT_SUCCESS mean, well, success. EXIT_FAILURE means failure.

The C implimentation is obliged to translate these values into whatever values the underlying OS uses for success and failure.

Returning any other value from main is implimentation defined behaviour... it's outside of the scope of C to say what it's meaning is.

Ian Woods