Follow UAX-29 and present user-perceived character counts

Not a count of code-units, nor a count of code-points, but try and follow
http://unicode.org/reports/tr29/ and present the grapheme count.

Add a few representative tests to try and avoid gotchas.
This commit is contained in:
Caolán McNamara
2011-11-03 11:00:38 +00:00
parent 56d5a48099
commit 12db5315fc
5 changed files with 76 additions and 14 deletions

View File

@@ -151,6 +151,24 @@ void TestBreakIterator::testGraphemeIteration()
i18n::CharacterIteratorMode::SKIPCELL, 1, nDone); i18n::CharacterIteratorMode::SKIPCELL, 1, nDone);
CPPUNIT_ASSERT_MESSAGE("Should skip full grapheme", nPos == 0); CPPUNIT_ASSERT_MESSAGE("Should skip full grapheme", nPos == 0);
} }
{
const sal_Unicode ALEF_QAMATS [] = { 0x05D0, 0x05B8 };
::rtl::OUString aText(ALEF_QAMATS, SAL_N_ELEMENTS(ALEF_QAMATS));
sal_Int32 nGraphemeCount = 0;
sal_Int32 nCurPos = 0;
while (nCurPos < aText.getLength())
{
sal_Int32 nCount2 = 1;
nCurPos = m_xBreak->nextCharacters(aText, nCurPos, lang::Locale(),
i18n::CharacterIteratorMode::SKIPCELL, nCount2, nCount2);
++nGraphemeCount;
}
CPPUNIT_ASSERT_MESSAGE("Should be considered 1 grapheme", nGraphemeCount == 1);
}
} }
//A test to ensure that certain ranges and codepoints that are categorized as //A test to ensure that certain ranges and codepoints that are categorized as

View File

@@ -111,6 +111,8 @@ public:
sal_uInt16 GetRealScriptOfText( const String& rTxt, xub_StrLen nPos ) const; sal_uInt16 GetRealScriptOfText( const String& rTxt, xub_StrLen nPos ) const;
sal_uInt16 GetAllScriptsOfText( const String& rTxt ) const; sal_uInt16 GetAllScriptsOfText( const String& rTxt ) const;
sal_Int32 getGraphemeCount(const rtl::OUString& rStr) const;
}; };
#define SW_BREAKITER() SwBreakIt::Get() #define SW_BREAKITER() SwBreakIt::Get()

View File

@@ -41,19 +41,20 @@
#include <sfx2/docfile.hxx> #include <sfx2/docfile.hxx>
#include <sfx2/sfxmodelfactory.hxx> #include <sfx2/sfxmodelfactory.hxx>
#include "init.hxx" #include "breakit.hxx"
#include "swtypes.hxx"
#include "docstat.hxx"
#include "doc.hxx" #include "doc.hxx"
#include "ndtxt.hxx"
#include "docsh.hxx" #include "docsh.hxx"
#include "shellres.hxx" #include "docstat.hxx"
#include "docufld.hxx" #include "docufld.hxx"
#include "fmtanchr.hxx" #include "fmtanchr.hxx"
#include "swscanner.hxx" #include "init.hxx"
#include "swcrsr.hxx" #include "ndtxt.hxx"
#include "swmodule.hxx"
#include "shellio.hxx" #include "shellio.hxx"
#include "shellres.hxx"
#include "swcrsr.hxx"
#include "swscanner.hxx"
#include "swmodule.hxx"
#include "swtypes.hxx"
SO2_DECL_REF(SwDocShell) SO2_DECL_REF(SwDocShell)
SO2_IMPL_REF(SwDocShell) SO2_IMPL_REF(SwDocShell)
@@ -73,14 +74,15 @@ public:
void testFileNameFields(); void testFileNameFields();
void testDocStat(); void testDocStat();
void testSwScanner(); void testSwScanner();
void testUserPerceivedCharCount();
void testGraphicAnchorDeletion(); void testGraphicAnchorDeletion();
CPPUNIT_TEST_SUITE(SwDocTest); CPPUNIT_TEST_SUITE(SwDocTest);
CPPUNIT_TEST(randomTest); CPPUNIT_TEST(randomTest);
CPPUNIT_TEST(testPageDescName); CPPUNIT_TEST(testPageDescName);
CPPUNIT_TEST(testFileNameFields); CPPUNIT_TEST(testFileNameFields);
CPPUNIT_TEST(testDocStat);
CPPUNIT_TEST(testSwScanner); CPPUNIT_TEST(testSwScanner);
CPPUNIT_TEST(testUserPerceivedCharCount);
CPPUNIT_TEST(testGraphicAnchorDeletion); CPPUNIT_TEST(testGraphicAnchorDeletion);
CPPUNIT_TEST_SUITE_END(); CPPUNIT_TEST_SUITE_END();
@@ -189,6 +191,28 @@ void SwDocTest::testDocStat()
CPPUNIT_ASSERT_MESSAGE("And cache is updated too", m_pDoc->GetDocStat().nChar == nLen); CPPUNIT_ASSERT_MESSAGE("And cache is updated too", m_pDoc->GetDocStat().nChar == nLen);
} }
//For UI character counts we should follow UAX#29 and display the user
//perceived characters, not the number of codepoints, nor the number of code
//units http://unicode.org/reports/tr29/
void SwDocTest::testUserPerceivedCharCount()
{
SwBreakIt *pBreakIter = SwBreakIt::Get();
//Grapheme example, two different unicode code-points perceived by the user as a single
//glyph
const sal_Unicode ALEF_QAMATS [] = { 0x05D0, 0x05B8 };
::rtl::OUString sALEF_QAMATS(ALEF_QAMATS, SAL_N_ELEMENTS(ALEF_QAMATS));
sal_Int32 nGraphemeCount = pBreakIter->getGraphemeCount(sALEF_QAMATS);
CPPUNIT_ASSERT_MESSAGE("Grapheme Count should be 1", nGraphemeCount == 1);
//Surrogate pair example, one single unicode code-point (U+1D11E)
//represented as two code units in UTF-8
const sal_Unicode GCLEF[] = { 0xD834, 0xDD1E };
::rtl::OUString sGCLEF(GCLEF, SAL_N_ELEMENTS(GCLEF));
sal_Int32 nCount = pBreakIter->getGraphemeCount(sGCLEF);
CPPUNIT_ASSERT_MESSAGE("Surrogate Pair should be counted as single character", nCount == 1);
}
//See https://bugs.freedesktop.org/show_bug.cgi?id=40449 for motivation //See https://bugs.freedesktop.org/show_bug.cgi?id=40449 for motivation
void SwDocTest::testSwScanner() void SwDocTest::testSwScanner()
{ {

View File

@@ -33,6 +33,7 @@
#include <unicode/uchar.h> #include <unicode/uchar.h>
#include <com/sun/star/lang/XMultiServiceFactory.hpp> #include <com/sun/star/lang/XMultiServiceFactory.hpp>
#include <com/sun/star/i18n/ScriptType.hdl> #include <com/sun/star/i18n/ScriptType.hdl>
#include <com/sun/star/i18n/CharacterIteratorMode.hpp>
#include <unotools/localedatawrapper.hxx> #include <unotools/localedatawrapper.hxx>
#include <editeng/unolingu.hxx> #include <editeng/unolingu.hxx>
@@ -169,4 +170,20 @@ sal_uInt16 SwBreakIt::GetAllScriptsOfText( const String& rTxt ) const
return nRet; return nRet;
} }
sal_Int32 SwBreakIt::getGraphemeCount(const rtl::OUString& rText) const
{
sal_Int32 nGraphemeCount = 0;
sal_Int32 nCurPos = 0;
while (nCurPos < rText.getLength())
{
sal_Int32 nCount2 = 1;
nCurPos = xBreak->nextCharacters(rText, nCurPos, lang::Locale(),
i18n::CharacterIteratorMode::SKIPCELL, nCount2, nCount2);
++nGraphemeCount;
}
return nGraphemeCount;
}
/* vim:set shiftwidth=4 softtabstop=4 expandtab: */ /* vim:set shiftwidth=4 softtabstop=4 expandtab: */

View File

@@ -1862,12 +1862,13 @@ void SwTxtNode::CountWords( SwDocStat& rStat,
if( 1 != aExpandText.match(aBreakWord, aScanner.GetBegin() )) if( 1 != aExpandText.match(aBreakWord, aScanner.GetBegin() ))
{ {
++nTmpWords; ++nTmpWords;
nTmpCharsExcludingSpaces += aScanner.GetLen(); nTmpCharsExcludingSpaces += pBreakIt->getGraphemeCount(aScanner.GetWord());
} }
} }
} }
nTmpChars = nExpandEnd - nExpandBegin - nNumOfMaskedChars; nTmpChars = pBreakIt->getGraphemeCount(aExpandText.copy(nExpandBegin, nExpandEnd - nExpandBegin));
nTmpChars -= nNumOfMaskedChars;
// no nTmpCharsExcludingSpaces adjust needed neither for blanked out MaskedChars // no nTmpCharsExcludingSpaces adjust needed neither for blanked out MaskedChars
// nor for mid-word selection - set scanner bClip = true at creation // nor for mid-word selection - set scanner bClip = true at creation
@@ -1889,10 +1890,10 @@ void SwTxtNode::CountWords( SwDocStat& rStat,
while ( aScanner.NextWord() ) while ( aScanner.NextWord() )
{ {
++nTmpWords; ++nTmpWords;
nTmpCharsExcludingSpaces += aScanner.GetLen(); nTmpCharsExcludingSpaces += pBreakIt->getGraphemeCount(aScanner.GetWord());
} }
nTmpChars += nNumStringLen; nTmpChars = pBreakIt->getGraphemeCount(aNumString);
} }
else if ( HasBullet() ) else if ( HasBullet() )
{ {