Segmentation fault - weird debugging with gdb

Question

I'm working in C, Linux terminal. I need to find a pattern in a text and recolor it. GDB debugging can locate the function that is causing the problem via (gdb) backtrace, but it shows me a terrible message when I try to find the exact line:

Error

Program received signal SIGSEGV, Segmentation fault.
strstr_sse2_unaligned ()
at ../sysdeps/x86_64/multiarch/strstr-sse2-unaligned.S:40
40 ../sysdeps/x86_64/multiarch/strstr-sse2-unaligned.S: No such file or dir
ectory.
(gbd)

The broken function is find_and_recolor:

char* my_replace(char *text, char* replacement)
{
      int lgreplacement = strlen(replacement);
      int lgtext = strlen(text);
      char *aux = (char*)malloc((lgreplacement + lgtext + 10) * sizeof(char));
      strcpy(aux, replacement);
      strcat(aux, text);
      return(aux);
}

char* find_and_recolor(char* text, char* pattern)
{
      int lgpattern = strlen(pattern);
      int lgreplace = lgpattern + 10;//there are exactly 10 characters that must be inserted along the pattern word
      int dif = 0;
      char *p;
      char *replacement = (char*)malloc(lgreplace * sizeof(char));
      strcpy(replacement, "\e[0;31m");
      strcat(replacement, pattern);
      strcat(replacement, "\e[m");//to recolor a word in red, that word must be surrounded by those characters
      while(p = strstr(text + dif, pattern))
      {
            p = my_replace(p, replacement);
            p += lgreplace;
            dif = p - text;
      }
      free(replacement);
      return strdup(text);
}

Regarding strstr-sse2-unaligned.S: No such file or dir, I believe you can install glibc-dev (Debian) or glibc-devel (Fedora) to clear the missing source files. — jww
– jww, Commented Nov 24, 2019 at 9:30

Employed Russian · Accepted Answer · 2019-11-24 04:08:29Z

3

it shows me a terrible message when I try to find the exact line:

There is nothing terrible, weird or unusual about this message, you just need to learn proper debugging technique.

What's happening is that the segmentation fault doesn't happen in your code, it happens inside GLIBC code (inside strstr), because you called strstr with bad arguments.

To find which call to strstr that was, use GDB up command to step out of GLIBC code, and into your code. Once you are inside find_and_recolor, you would be able to see the exact line, and print values of text, dif and pattern which caused your crash (assuming you compiled your code for debugging, i.e. with the -g flag).

edited Nov 24, 2019 at 4:08

answered Nov 24, 2019 at 4:01

Employed Russian

217k36 gold badges324 silver badges396 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

jww · Accepted Answer · 2019-11-24 09:28:18Z

0

Updating diff to p-text in while loop where both pointer points to different array doesn't make sense. It is undefined behavior.

Also code has other issues.

Uninitialized variable.
Less optimized as number of call can be reduced.

edited Nov 24, 2019 at 9:28

jww

104k107 gold badges453 silver badges975 bronze badges

answered Nov 24, 2019 at 2:23

anuplohiya

463 bronze badges

3 Comments

Andrei Stan Over a year ago

p is pointer towards a location in text array. So it makes sense to calculate a difference. Example: strstr finds the pattern at position 5 in text array, p moves to text + 5, so the difference will be p - text = 5.

anuplohiya Over a year ago

No my_replace function return pointer to newly allocated array in that function. Try printing address of p and text.

Andrei Stan Over a year ago

You are right... Also, I know the number of function calls can be reduced, but I just need it to be done in any method. No need for efficiency.

Collectives™ on Stack Overflow

Segmentation fault - weird debugging with gdb

2 Answers 2

Comments

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related