test/unit/collate_facets_test.cpp

#include "locale_test.h"

#if !defined (STLPORT) || !defined (_STLP_USE_NO_IOSTREAMS)
#  include <locale>
#  include <stdexcept>
#  include <algorithm>
#  include <vector>

#  if !defined (STLPORT) || defined(_STLP_USE_NAMESPACES)
using namespace std;
#  endif

//
// tests implementation
//
void LocaleTest::collate_facet()
{
  {
    CPPUNIT_ASSERT( has_facet<collate<char> >(locale::classic()) );
    collate<char> const& col = use_facet<collate<char> >(locale::classic());

    char const str1[] = "abcdef1";
    char const str2[] = "abcdef2";
    const size_t size1 = sizeof(str1) / sizeof(str1[0]) - 1;
    const size_t size2 = sizeof(str2) / sizeof(str2[0]) - 1;

    CPPUNIT_ASSERT( col.compare(str1, str1 + size1 - 1, str2, str2 + size2 - 1) == 0 );
    CPPUNIT_ASSERT( col.compare(str1, str1 + size1, str2, str2 + size2) == -1 );

    //Smallest string should be before largest one:
    CPPUNIT_ASSERT( col.compare(str1, str1 + size1 - 2, str2, str2 + size2 - 1) == -1 );
    CPPUNIT_ASSERT( col.compare(str1, str1 + size1 - 1, str2, str2 + size2 - 2) == 1 );
  }

#  if !defined (STLPORT) || defined (_STLP_USE_EXCEPTIONS)
  try {
    locale loc("fr_FR");
    {
      CPPUNIT_ASSERT( has_facet<collate<char> >(loc) );
      collate<char> const& col = use_facet<collate<char> >(loc);

      char const str1[] = "abcdef1";
      char const str2[] = "abcdef2";
      const size_t size1 = sizeof(str1) / sizeof(str1[0]) - 1;
      const size_t size2 = sizeof(str2) / sizeof(str2[0]) - 1;

      CPPUNIT_ASSERT( col.compare(str1, str1 + size1 - 1, str2, str2 + size2 - 1) == 0 );
      CPPUNIT_ASSERT( col.compare(str1, str1 + size1, str2, str2 + size2) == -1 );

      //Smallest string should be before largest one:
      CPPUNIT_ASSERT( col.compare(str1, str1 + size1 - 2, str2, str2 + size2 - 1) == -1 );
      CPPUNIT_ASSERT( col.compare(str1, str1 + size1 - 1, str2, str2 + size2 - 2) == 1 );
    }
    {
      CPPUNIT_ASSERT( has_facet<collate<char> >(loc) );
      collate<char> const& col = use_facet<collate<char> >(loc);

      string strs[] = {"abdd", "abçd", "abbd", "abcd"};

      string transformed[4];
      for (size_t i = 0; i < 4; ++i) {
        transformed[i] = col.transform(strs[i].data(), strs[i].data() + strs[i].size());
      }

      sort(strs, strs + 4, loc);
      CPPUNIT_ASSERT( strs[0] == "abbd" );
      CPPUNIT_ASSERT( strs[1] == "abcd" );
      CPPUNIT_ASSERT( strs[2] == "abçd" );
      CPPUNIT_ASSERT( strs[3] == "abdd" );

      sort(transformed, transformed + 4);

      CPPUNIT_ASSERT( col.transform(strs[0].data(), strs[0].data() + strs[0].size()) == transformed[0] );
      CPPUNIT_ASSERT( col.transform(strs[1].data(), strs[1].data() + strs[1].size()) == transformed[1] );
      CPPUNIT_ASSERT( col.transform(strs[2].data(), strs[2].data() + strs[2].size()) == transformed[2] );
      CPPUNIT_ASSERT( col.transform(strs[3].data(), strs[3].data() + strs[3].size()) == transformed[3] );

      // Check empty string result in empty key.
      CPPUNIT_ASSERT( col.transform(strs[0].data(), strs[0].data()).empty() );

      // Check that only characters that matter are taken into accout to build the key.
      CPPUNIT_ASSERT( col.transform(strs[0].data(), strs[0].data() + 2) == col.transform(strs[1].data(), strs[1].data() + 2) );
    }
#    if !defined (STLPORT) || !defined (_STLP_NO_WCHAR_T)
    {
      CPPUNIT_ASSERT( has_facet<collate<wchar_t> >(loc) );
      collate<wchar_t> const& col = use_facet<collate<wchar_t> >(loc);

      wchar_t const str1[] = L"abcdef1";
      wchar_t const str2[] = L"abcdef2";
      const size_t size1 = sizeof(str1) / sizeof(str1[0]) - 1;
      const size_t size2 = sizeof(str2) / sizeof(str2[0]) - 1;

      CPPUNIT_ASSERT( col.compare(str1, str1 + size1 - 1, str2, str2 + size2 - 1) == 0 );
      CPPUNIT_ASSERT( col.compare(str1, str1 + size1, str2, str2 + size2) == -1 );

      //Smallest string should be before largest one:
      CPPUNIT_ASSERT( col.compare(str1, str1 + size1 - 2, str2, str2 + size2 - 1) == -1 );
      CPPUNIT_ASSERT( col.compare(str1, str1 + size1 - 1, str2, str2 + size2 - 2) == 1 );
    }
    {
      size_t i;
      CPPUNIT_ASSERT( has_facet<collate<wchar_t> >(loc) );
      collate<wchar_t> const& col = use_facet<collate<wchar_t> >(loc);

      // Here we would like to use L"abçd" but it looks like all compilers
      // do not support storage of unicode characters in exe resulting in
      // compilation error. We avoid this test for the moment.
      wstring strs[] = {L"abdd", L"abcd", L"abbd", L"abcd"};

      wstring transformed[4];
      for (i = 0; i < 4; ++i) {
        transformed[i] = col.transform(strs[i].data(), strs[i].data() + strs[i].size());
      }

      sort(strs, strs + 4, loc);
      CPPUNIT_ASSERT( strs[0] == L"abbd" );
      CPPUNIT_ASSERT( strs[1] == L"abcd" );
      CPPUNIT_ASSERT( strs[2] == L"abcd" );
      CPPUNIT_ASSERT( strs[3] == L"abdd" );

      sort(transformed, transformed + 4);

      CPPUNIT_ASSERT( col.transform(strs[0].data(), strs[0].data() + strs[0].size()) == transformed[0] );
      CPPUNIT_ASSERT( col.transform(strs[1].data(), strs[1].data() + strs[1].size()) == transformed[1] );
      CPPUNIT_ASSERT( col.transform(strs[2].data(), strs[2].data() + strs[2].size()) == transformed[2] );
      CPPUNIT_ASSERT( col.transform(strs[3].data(), strs[3].data() + strs[3].size()) == transformed[3] );

      CPPUNIT_ASSERT( col.transform(strs[0].data(), strs[0].data()).empty() );

      CPPUNIT_ASSERT( col.transform(strs[0].data(), strs[0].data() + 2) == col.transform(strs[1].data(), strs[1].data() + 2) );
    }
#    endif
  }
  catch (runtime_error const&) {
    CPPUNIT_MESSAGE("No french locale to check collate facet");
  }
#  endif
}

void LocaleTest::collate_by_name()
{
  /*
   * Check of the 22.1.1.2.7 standard point. Construction of a locale
   * instance from a null pointer or an unknown name should result in
   * a runtime_error exception.
   */
#  if !defined (STLPORT) || defined (_STLP_USE_EXCEPTIONS)
#    if defined (STLPORT) || !defined (__GNUC__)
  try {
    locale loc(locale::classic(), new collate_byname<char>(static_cast<char const*>(0)));
    CPPUNIT_FAIL;
  }
  catch (runtime_error const& /* e */) {
    //CPPUNIT_MESSAGE( e.what() );
  }
  catch (...) {
    CPPUNIT_FAIL;
  }
#    endif

  try {
    locale loc(locale::classic(), new collate_byname<char>("yasli_language"));
    CPPUNIT_FAIL;
  }
  catch (runtime_error const& /* e */) {
    //CPPUNIT_MESSAGE( e.what() );
  }
  catch (...) {
    CPPUNIT_FAIL;
  }

  try {
    string veryLongFacetName("LC_COLLATE=");
    veryLongFacetName.append(512, '?');
    locale loc(locale::classic(), new collate_byname<char>(veryLongFacetName.c_str()));
    CPPUNIT_FAIL;
  }
  catch (runtime_error const& /* e */) {
    //CPPUNIT_MESSAGE( e.what() );
  }
  catch (...) {
    CPPUNIT_FAIL;
  }

  try {
    locale loc(locale::classic(), "C", locale::collate);
  }
  catch (runtime_error const& e) {
    CPPUNIT_MESSAGE( e.what() );
    CPPUNIT_FAIL;
  }
  catch (...) {
    CPPUNIT_FAIL;
  }

  try {
    // On platform without real localization support we should rely on the "C" facet.
    locale loc(locale::classic(), "", locale::collate);
  }
  catch (runtime_error const& e) {
    CPPUNIT_MESSAGE( e.what() );
    CPPUNIT_FAIL;
  }
  catch (...) {
    CPPUNIT_FAIL;
  }

  try {
    locale loc(locale::classic(), new collate_byname<char>("C"));

    //We check that the C locale gives a lexicographical comparison:
    collate<char> const& cfacet_byname = use_facet<collate<char> >(loc);
    collate<char> const& cfacet = use_facet<collate<char> >(locale::classic());

    char const str1[] = "abcdef1";
    char const str2[] = "abcdef2";
    const size_t size1 = sizeof(str1) / sizeof(str1[0]) - 1;
    const size_t size2 = sizeof(str2) / sizeof(str2[0]) - 1;

    CPPUNIT_ASSERT( cfacet_byname.compare(str1, str1 + size1 - 1, str2, str2 + size2 - 1) ==
                    cfacet.compare(str1, str1 + size1 - 1, str2, str2 + size2 - 1) );
    CPPUNIT_ASSERT( cfacet_byname.compare(str1, str1 + size1, str2, str2 + size2) ==
                    cfacet.compare(str1, str1 + size1, str2, str2 + size2) );

    //Smallest string should be before largest one:
    CPPUNIT_ASSERT( cfacet_byname.compare(str1, str1 + size1 - 2, str2, str2 + size2 - 1) ==
                    cfacet.compare(str1, str1 + size1 - 2, str2, str2 + size2 - 1) );
    CPPUNIT_ASSERT( cfacet_byname.compare(str1, str1 + size1 - 1, str2, str2 + size2 - 2) ==
                    cfacet.compare(str1, str1 + size1 - 1, str2, str2 + size2 - 2) );

    // We cannot play with 'ç' char here because doing so would make test result
    // dependant on char being consider as signed or not...
    string strs[] = {"abdd", /* "abçd",*/ "abbd", "abcd"};

    vector<string> v1(strs, strs + sizeof(strs) / sizeof(strs[0]));
    sort(v1.begin(), v1.end(), loc);
    vector<string> v2(strs, strs + sizeof(strs) / sizeof(strs[0]));
    sort(v2.begin(), v2.end(), locale::classic());
    CPPUNIT_ASSERT( v1 == v2 );

    CPPUNIT_ASSERT( (cfacet_byname.transform(v1[0].data(), v1[0].data() + v1[0].size()).compare(cfacet_byname.transform(v1[1].data(), v1[1].data() + v1[1].size())) ==
                    v1[0].compare(v1[1])) );
  }
  catch (runtime_error const& /* e */) {
    /* CPPUNIT_MESSAGE( e.what() ); */
    CPPUNIT_FAIL;
  }
  catch (...) {
    CPPUNIT_FAIL;
  }

#    if !defined (STLPORT) || !defined (_STLP_NO_WCHAR_T)
#      if defined (STLPORT) || !defined (__GNUC__)
  try {
    locale loc(locale::classic(), new collate_byname<wchar_t>(static_cast<char const*>(0)));
    CPPUNIT_FAIL;
  }
  catch (runtime_error const&) {
  }
  catch (...) {
    CPPUNIT_FAIL;
  }
#      endif

  try {
    locale loc(locale::classic(), new collate_byname<wchar_t>("yasli_language"));
    CPPUNIT_FAIL;
  }
  catch (runtime_error const&) {
  }
  catch (...) {
    CPPUNIT_FAIL;
  }
#    endif
#  endif
}

#endif