Thread: ctype segmentation fault

Hi all,

The program I wrote compiles, but it always causes a segmentation fault in runtime and I have no idea why.

Code:

#include <iostream>
#include <locale>

using namespace std;

const unsigned int MAXBUF = 2048;

int main(int argc, char* argv[])
{
  const ctype<char>& ct_de = use_facet<ctype<char> >(locale("de_DE.utf8"));

  char buf[MAXBUF];
  char* c;

  cin.imbue(locale::classic());
  cout.imbue(locale::classic());
  while(cin.getline(buf, MAXBUF))
  {
    c = buf;
    while(*c)
      {
	cout << ct_de.tolower(*c++);
      }
    cout << endl;
  }
  return 0;
}

I'm running this on an Ubuntu whose default locale is "en_EN.UTF-8".

A libation and hecatomb for anyone who can help.

Yes, I tried that. I'm still trying to get the hang of gdb, but anyway, I'm still not proficient enough to understand what I'm looking at.

I was trying to ween myself off MSVC, but I had to go back, just to see what was going on, and even there, the error message is inscrutable.

Code:

First-chance exception at 0x004358b1 in de_tolower.exe: 0xC0000005: Access violation reading location 0xfeeefef6.
Unhandled exception at 0x004358b1 in de_tolower.exe: 0xC0000005: Access violation reading location 0xfeeefef6.

The only thing I did to the code was change the ctype definition so that it would conform to Microsoft locale settings:

Code:

const ctype<char>& ct_de = use_facet<ctype<char> >(locale("German_Germany"));

I tried to step into the code as much as possible in debug mode but it would not go beyond "xlocale":

Code:

	_Elem tolower(_Elem _Ch) const
		{	// convert element to lower case
		return (do_tolower(_Ch));
		}

I'm stumped.

To be honest, it would be much easier for me just to set up my own conversion table for the upper ASCII, but since c++ had localized libraries (which seem to be platform dependent to a certain degree), I thought it might be a good idea to learn how to use them. If no solutions turn up, I'll just have to manually encode the case conversion.

Code:

const ctype<char>& ct_de = use_facet<ctype<char> >(locale("de_DE.utf8"));

this line is giving it most probably

THANK YOU!!!

But would you mind telling me why? I mean, what would be the difference between creating a locale object, then passing it to use_facet, and doing "locale("de_DE.utf8")" inside use_facet?

Also, if you don't mind answering another question as well, I've discovered that to REALLY use C/C++, you need to understand machine level code (or assembly or whatever). I've run into so many of these snafus ever since I learned C/C++ that I am now sorely tempted to dive into assembly. Do you guys think this is a good idea? Or could you suggest a better alternative that would enhance my skills at debugging errors, most especially segmentation faults?

I first learned c++ programming through two books "on to c++" by Patrick Winston and "c++ in plain English" by Brian Overland. I actually managed to read them almost cover to cover.
After that, I just used a bunch of reference books which I look at whenever the need arises. "The C++ Standard Library" by Josuttis is my main resource, as it probably is for many other people.

Thanks, JaWib, I kind of understand what you're saying, and I'll look into my references for a better understanding.

And Salem, thanks for setting me aright on the debugging issue.