CareerCup

2nd/3rd options invalidate main's prototype.
IMO, the first one will compile because it is possible to cast * to int.

thinking about it more, i think #2 will also work

thinking about it more, i think #2 will also work

Tried in code

with gcc version 4.1.2 20080704 (Red Hat 4.1.2-44)..

and to my surprise, all worked though with a warning in 2 cases :(

yes Anonymous, all 3 are working. But Why? I am not geeting any clue why it is working.
Initially before executing I thought, 1st one will compile through but 2nd and 3rd will not. my reason was that, in 1st case we are not overloading main, so it will compile. But rest of the cases we are trying to oveload main function. So I though it wll not compile. But all 3 get compile through. Don't kw why? anyboby can explain?

You have just compiled using gcc -c or its linked properly also ?

@ Anonymous

I think, its moreover compiler dependent stuff.
But as per language point of view, the main function should have the below properties

1: first argument of 'main' should be `int'
2: second argument of 'main' should be 'char **'
3: return type of 'main' should be int

Please try the test item with gcc's -Wall option
$gcc test.c -Wall

I think it's compiler dependent feature. When compiler compiles, it doesn't know the exact prototype of main which will be called from startup routine [part of the loader domain I guess]

even u can see , main(a,b,c,d,e,f,g) - working and all the variables are stacked properly.

compilerdoesnt really check for th prototype maybe, but it should give a runtime error. Did anyone check for runtime exception???

I wrote a printf statement in the main and the code thrice... Each time with different definition of main. Each time I got the same error

timberlake {~} > gcc test.c -o test
test.c: In function âmainâ:
test.c:7: warning: return type of âmainâ is not âintâ

Then when I ran the code to see if there was any run time bug, to my surprise there was none