Linux C question using the gcc compiler.
gcc -Wall -W my_c_code.c -o ouput_program
How do you deal with a segmentation error?
I get this segmentation error at compile time
so is it a syntax error of some kind?
There is no clue as to what line of code is
causing it, so the entire program is suspect.
It drives me nuts! How do you get rid of a
segmentation error?

if

gcc -Wall -W my_c_code.c -o ouput_program

is producing a segmentation fault, then it the problem is not in your program. Since its not being executed yet and segmentation faults occur at runtime.

This means that your gcc is getting a segmentation fault. This could possibly be because of corrupted installation of gcc or libraries.

I would not like to comment until I could see the code. As its only one source file could you please post it, or dump it at a location in my signature.
thanks

Hi -

Like sibtay said: you've got a corrupted and/or incompatible version of gcc:

1. Find a version of gcc that matches your platform, OS and OS version
(for example, go to www.rpmfind.org)

2. De-install your current version. Reboot your PC.

3. Install the new, "known good" version of gcc.

Hmmm interesting that two people think it's gcc. As i always use an ide and never the command line, could please explain why it's gcc and not the code?

ty

What exactly is getting the SIGSEGV? Is it the compiler, during compilation, which may leave a large "gcc.core" file behind? Or is it your program, during execution, which probably leaves an "output_program.core" file? If it is gcc, which seems very very unlikely to me (unless you did some bleeding edge CVS installation), you may need to file a bug report. Otherwise try testing your output_program with a debugger: `gdb ./output_program` and type "run".

To dmail: a compiler / interpreter should never crash during compilation / execution. It'd be like your IDE crashing in the middle of a session. It shouldn't happen no matter how bad the code you pass into it is. The compiler exists to catch those sort of errors! And if it just dies then it's not doing its job.

Because TS indicates that it's at compile time and not at runtime He/She might have informed us incorrectly, but till that time I agree that it's not the compiled program as it's not running and so it must be the compiler.

Thanks I see it now.
I'm sure when I use an ide i can get seg faults, but this maybe due to the build in debugger.

I just ran the following from the command line and got no warnings so I will stay quite on this matter lol.

I got it fixed. There is nothing wrong with
gcc or even my ancient computer. It was the code.
I can't remember exactly what I did since I have
a habbit of panicing and trying everything at
once, but I think the problem was this:
I won't write all the code, only the stuff that
matters:
Defined a structure globally
struct my_structure {
float * big_array;
other stuff...
}
typedef struct my_structure STX;

void initialize_structure( STX * stxp )
{
float * temp_array;
temp_array = (float *)malloc( sizeof(float) * 200 );

This next line caused the segmentation error
stxp->big_array = temp_array;

I fixed this in main()
int main(void)
{
STX * my_structure_ptr;
The next two lines fixed the problem
STX the_structure;
my_structure_ptr = &the_structure;

Real dumb mistake! I should have caught it
right alway, but I suspected the everything else,
including the computer hardware. It's kind of
hard to admit to a mistake like this. It's so
stupid!
While changing code all over the program and
recompiling, I got all kinds of wierd errors.
One was even from a library I never heard of
before. ( Id.so or something like that )
I should have written it down.
I use the command line and the nano editor.
For debugging I use printf and /* */ all over.
Sometimes gdb or DDD. The IDEs I've tried
all seem to take a year to "configure" and
my little programs aren't worth the trouble.
Thanks to you all for your help.