mohitsaxena019

Hi all,

I am playing with pow() function using gcc compiler. I have got some intresting error while playing. Here I am writing the code as well as the problem

main()
{
int i,j,k;
j=10;
k=2;
i=pow(j,k);
printf("i=%d",i);
}
It is the Complete code no math directory included.
when i compile this using the following command:
gcc test.c
It gives the following error
/var/tmp//ccUP0AVk.o(.text+0x3f): In function `main':
: undefined reference to `pow'

when i compile this using the following command:
gcc -lm test.c
It gives no error and correct output

Now I had make some changes in the program

main()
{
int i,j,k;
j=10;
k=2;
i=pow(10,2);
printf("i=%d",i);
}

when i compile this using the following command:
gcc test.c
It gives no error and correct output

Q1:-) Is it not necessary to include files in the code.
Q2:-) Why it is not giving error in the second code.

Plz help me out.

zacs7

Q1. Because of the default prototype presumed by the compiler for "pow()". It's a bad idea to not include the prototype. See the C89 standard for more information.

Q2. Probably left over object files and linkage. Delete all the generated files (including ones in tmp) then you'll most likely get the error again.

You might want to use the code tags next time you post code. See << !! Posting Code? Read this First !! >>

You should be using prototypes, and you shouldn't be declaring main implicitly. It returns int and takes either void arguments, or int, char **. So declare it as such.

__________________
"I.T. gets the chicky-babes" - M. Kelly
bakefile | vim

mohitsaxena019

Sorry i did'nt think that these are the answers of my questions

Fordy

mohitsaxena019 - why did you report this to the moderators? It wont get you any help and if you do it again I'll close the thread

__________________
"If A is success in life, then A equals x plus y plus z. Work is x; y is play; and z is keeping your mouth shut."
Albert Einstein (1879 - 1955)

Board Rules

mohitsaxena019

Sorry that was an accident

kille6525

ok if your going to use a function you have to declare it before the main loop.

oh and the second time use:

gcc -Wall -Werror -ansi -pedantic main.c -o main.exe

that will probly give you the error look up my vids on youtube for more info about functions

laserlight

That is pretty much what zacs7 already stated, except that it was pointed out that you might not have to declare a prototype, though the results may then not be what you expect.

Anyway, I do not suggest naming a function pow in view of a possible conflict with pow from <math.h>, which mohitsaxena019 apparently is using, considering the -lm option.

__________________
C + C++ Compiler: MinGW port of GCC
Build + Version Control System: SCons + Bazaar

Look up a C/C++ Reference and learn How To Ask Questions The Smart Way

mohitsaxena019

Dear May be both of you do'nt know that pow() is a inbuilt function. It is their in math.h library

mohitsaxena019

Sorry laserlight i am agree with you but this is not the question which I am facing Plz read it clearly

laserlight

zacs7 has already answered your question: you need to forward declare pow, otherwise the signature and return type assumed will not match the pow from <math.h>. Of course, the simplest solution is to just #include <math.h> as you should have done in the first place.

EDIT:
Okay, I see that I did misread your question. You should have stated why you think zacs7's answers are not satisfactory.

Logically, the effects of default int do not seem to be in effect here, which is a little puzzling, but is confirmed by you being able to compile without including the header.

Consequently, I suspect that you may be observing something implementation defined: the answer is just don't do it.

__________________
C + C++ Compiler: MinGW port of GCC
Build + Version Control System: SCons + Bazaar

Look up a C/C++ Reference and learn How To Ask Questions The Smart Way

sean

You provided constants for both the inputs to pow, so I'm betting the compiler made an optimization and replaced the call with the literal value, or inlined pow for you, since pow(10,2) will always be the same value, and you didn't define pow in any way. That seems to be the case with this thread: gcc and pow in a C program - JustLinux Forums, where they actually look at the assembly outputted by gcc to see what was going on.

edit: as laserlight implies, this is NOT part of the C standard in anyway - it's specific to gcc, and even then, perhaps to only certain versions - so don't use it in your code or rely on it.