diff options
| author | Caolán McNamara <caolanm@redhat.com> | 2020-11-27 12:14:44 +0000 |
|---|---|---|
| committer | Caolán McNamara <caolanm@redhat.com> | 2020-11-28 20:31:29 +0100 |
| commit | fc011ac027da7574baa4b431ac800020170ba8b0 (patch) | |
| tree | 21066a3bda35695cd55c4cba4913f34da8e2eacf /i18npool/qa | |
| parent | 5ec2cf4e1ff448a3d9b3b30a825a40f962e0a53d (diff) | |
drop custom Indic grapheme rules and rely on contemporary icu defaults
similar to...
commit 8578a1c9d167c19f1d8038fac5946b4b3cae305e
Date: Thu Nov 26 15:47:26 2020 +0200
tdf#138481: Trust the built-in break iterator character data in ICU
Don't use our own char.txt.
the char_in.txt hasn't really changed since 2008 and is woefully out of
date at this point.
we have cppunit tests for the only documented bug that touched
char_in.txt, #i111152# and tdf#40292, for tdf#40292 change the test
to test what was actually reported as a bug
Change-Id: I8e35b102b0a46d2c63e47e055e472892f65022ac
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106763
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Diffstat (limited to 'i18npool/qa')
| -rw-r--r-- | i18npool/qa/cppunit/test_breakiterator.cxx | 24 |
1 files changed, 18 insertions, 6 deletions
diff --git a/i18npool/qa/cppunit/test_breakiterator.cxx b/i18npool/qa/cppunit/test_breakiterator.cxx index 7d900fcfdbac..384041ccd5db 100644 --- a/i18npool/qa/cppunit/test_breakiterator.cxx +++ b/i18npool/qa/cppunit/test_breakiterator.cxx @@ -662,17 +662,29 @@ void TestBreakIterator::testGraphemeIteration() aLocale.Country = "IN"; { - const sal_Unicode KA_VIRAMA_SSA[] = { 0x0B95, 0x0BCD, 0x0BB7 }; - OUString aTest(KA_VIRAMA_SSA, SAL_N_ELEMENTS(KA_VIRAMA_SSA)); + const sal_Unicode CA_VOWELSIGNI_TA_VIRAMA_TA_VOWELSIGNI_RA_VOWELSIGNAI [] = { 0x0B9A, 0x0BBF, 0x0BA4, 0x0BCD, 0x0BA4, 0x0BBF, 0x0BB0, 0x0BC8 }; + + OUString aTest(CA_VOWELSIGNI_TA_VIRAMA_TA_VOWELSIGNI_RA_VOWELSIGNAI, SAL_N_ELEMENTS(CA_VOWELSIGNI_TA_VIRAMA_TA_VOWELSIGNI_RA_VOWELSIGNAI)); sal_Int32 nDone=0; sal_Int32 nPos = 0; - nPos = m_xBreak->nextCharacters(aTest, 0, aLocale, - i18n::CharacterIteratorMode::SKIPCELL, 1, nDone); - CPPUNIT_ASSERT_EQUAL_MESSAGE("Should skip full grapheme", static_cast<sal_Int32>(SAL_N_ELEMENTS(KA_VIRAMA_SSA)), nPos); - nPos = m_xBreak->previousCharacters(aTest, SAL_N_ELEMENTS(KA_VIRAMA_SSA), aLocale, + nPos = m_xBreak->nextCharacters(aTest, 0, aLocale, i18n::CharacterIteratorMode::SKIPCELL, 1, nDone); + CPPUNIT_ASSERT_EQUAL_MESSAGE("Should skip full grapheme", static_cast<sal_Int32>(2), nPos); + nPos = m_xBreak->nextCharacters(aTest, nPos, aLocale, i18n::CharacterIteratorMode::SKIPCELL, 1, nDone); + CPPUNIT_ASSERT_EQUAL_MESSAGE("Should skip full grapheme", static_cast<sal_Int32>(4), nPos); + nPos = m_xBreak->nextCharacters(aTest, nPos, aLocale, i18n::CharacterIteratorMode::SKIPCELL, 1, nDone); + CPPUNIT_ASSERT_EQUAL_MESSAGE("Should skip full grapheme", static_cast<sal_Int32>(6), nPos); + nPos = m_xBreak->nextCharacters(aTest, nPos, aLocale, i18n::CharacterIteratorMode::SKIPCELL, 1, nDone); + CPPUNIT_ASSERT_EQUAL_MESSAGE("Should skip full grapheme", static_cast<sal_Int32>(SAL_N_ELEMENTS(CA_VOWELSIGNI_TA_VIRAMA_TA_VOWELSIGNI_RA_VOWELSIGNAI)), nPos); + nPos = m_xBreak->previousCharacters(aTest, SAL_N_ELEMENTS(CA_VOWELSIGNI_TA_VIRAMA_TA_VOWELSIGNI_RA_VOWELSIGNAI), aLocale, i18n::CharacterIteratorMode::SKIPCELL, 1, nDone); + CPPUNIT_ASSERT_EQUAL_MESSAGE("Should skip full grapheme", static_cast<sal_Int32>(6), nPos); + nPos = m_xBreak->previousCharacters(aTest, nPos, aLocale, i18n::CharacterIteratorMode::SKIPCELL, 1, nDone); + CPPUNIT_ASSERT_EQUAL_MESSAGE("Should skip full grapheme", static_cast<sal_Int32>(4), nPos); + nPos = m_xBreak->previousCharacters(aTest, nPos, aLocale, i18n::CharacterIteratorMode::SKIPCELL, 1, nDone); + CPPUNIT_ASSERT_EQUAL_MESSAGE("Should skip full grapheme", static_cast<sal_Int32>(2), nPos); + nPos = m_xBreak->previousCharacters(aTest, nPos, aLocale, i18n::CharacterIteratorMode::SKIPCELL, 1, nDone); CPPUNIT_ASSERT_EQUAL_MESSAGE("Should skip full grapheme", static_cast<sal_Int32>(0), nPos); } |