Lower a std::string, but not completely

[denkfix]

Hi,

I use the stl features to deal with string conversions, especially facets. Now, I am faced with the task to lower a string except its first letter. Please note that I am dealing with multi-byte characters and so I cannot simply replace the first character of the string with its original after lowering it. Here is my code for lowering a string entirely:

std::string lower_string ( const std::string& a_string, const std::locale& a_locale )
{
	std::string result;
	char *p = (char*) alloca ( a_string.size() + 1 );

	memcpy ( p, a_string.c_str (), a_string.size() + 1);
	std::use_facet< std::ctype<char> > ( a_locale ).tolower ( p, p + a_string.length () );
	result.assign(p);
	return result;
}

Has anyone an idea how to avoid the conversion of the first (probably) multi-byte character in a c++ way? Or a way to subsequently upper the first character?

Many thanks in advance,
Kay

[gerard4143]

So what you want is a function that creates a new string which is a copy of the string being past... Why not change the 'const std::string& a_string' parameter to 'std::string& a_string' and change original string or is that outside the requirements?

Your question, get the iterators for the string's begin and end and then increment the begin value by one.

[gerard4143]

I didn't take the time and read your post completely..Ooops

Try something like below.

#include <iostream>
#include <string>
#include <algorithm>
#include <locale>


std::string lower_string (const std::string& a_string, const std::locale& a_locale )
{
    char *str = new char(a_string.size());
    std::string::const_iterator begin = a_string.begin();
    std::string::const_iterator end = a_string.end();

    copy(begin, end, str);

    std::string ans;

    std::use_facet< std::ctype<char> > ( a_locale ).tolower ( str + 1, str + a_string.size() );

    ans = str;
    
    return ans;
}

int main(int argc, char *argv[])
{
    std::locale loc("");
    std::string name("A nAME tO PASS aLONG");

    std::string ans = lower_string(name, loc);

    std::cout << ans << std::endl;
    return 0;
}

[gerard4143]

Here's a tighter version.

#include <iostream>
#include <string>
#include <algorithm>
#include <locale>

std::string lower_string (const std::string& a_string, const std::locale& a_locale )
{
    char *str = new char[a_string.size() + 1];
    std::string::const_iterator begin = a_string.begin();
    std::string::const_iterator end = a_string.end();

    copy(begin, end, str);

    std::use_facet< std::ctype<char> > ( a_locale ).tolower ( str + 1, str + a_string.size() );

    std::string ans(str);
    delete [] str;

    return ans;
}

int main(int argc, char *argv[])
{
    std::locale loc("");
    std::string name("A nAME tO PASS aLONG");

    std::string ans = lower_string(name, loc);

    std::cout << ans << std::endl;
    return 0;
}

Its been a while since I look at C++ code.

[denkfix]

Many thanks! I had to add the following line to make it work (right after the conversion):

str[a_string.length ()] = '\0';

The bad thing: it does not work with multi-byte characters as in

std::string name("Ä nAME tO PÄSS aLONG");

, even if I try to convert the whole string which is quit frustrating. What is the point in using locale then? Using std::wstring makes it work:

std::wstring lower_string2 ( const std::wstring& a_string, const std::locale& a_locale )
{
	std::wstring b_string; 
	b_string += a_string;
	std::use_facet< std::ctype<wchar_t> > ( a_locale ).tolower ( const_cast<wchar_t*> ( b_string.c_str () ), b_string.c_str () + b_string.length () );
	return b_string;
}

[denkfix]

In my point of view this behavior is a clear bug: Real multibyte characters in multibyte strings are not converted by the conversion functions offered in <locale> regardless of the chosen locale. Meanwhile I tried boost::algorithm::string::to_lower, but this function also reverts to std::ctype yielding the same results. For all who are faced with the same problem, try boost::locale (http://cppcms.sourceforge.net/boost_locale/html/index.html). The conversion methods in this library work as expected.