/* Test of canonical decomposition of Unicode characters. Copyright (C) 2009-2023 Free Software Foundation, Inc. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program. If not, see . */ /* Written by Bruno Haible , 2009. */ #include #include "uninorm.h" #include "macros.h" int main () { ucs4_t decomposed[UC_DECOMPOSITION_MAX_LENGTH]; int ret; /* SPACE */ ret = uc_canonical_decomposition (0x0020, decomposed); ASSERT (ret == -1); /* LATIN CAPITAL LETTER A WITH DIAERESIS */ ret = uc_canonical_decomposition (0x00C4, decomposed); ASSERT (ret == 2); ASSERT (decomposed[0] == 0x0041); ASSERT (decomposed[1] == 0x0308); /* LATIN CAPITAL LETTER A WITH DIAERESIS AND MACRON */ ret = uc_canonical_decomposition (0x01DE, decomposed); ASSERT (ret == 2); ASSERT (decomposed[0] == 0x00C4); ASSERT (decomposed[1] == 0x0304); /* GREEK DIALYTIKA AND PERISPOMENI */ ret = uc_canonical_decomposition (0x1FC1, decomposed); ASSERT (ret == 2); ASSERT (decomposed[0] == 0x00A8); ASSERT (decomposed[1] == 0x0342); /* SCRIPT SMALL L */ ret = uc_canonical_decomposition (0x2113, decomposed); ASSERT (ret == -1); /* NO-BREAK SPACE */ ret = uc_canonical_decomposition (0x00A0, decomposed); ASSERT (ret == -1); /* ARABIC LETTER VEH INITIAL FORM */ ret = uc_canonical_decomposition (0xFB6C, decomposed); ASSERT (ret == -1); /* ARABIC LETTER VEH MEDIAL FORM */ ret = uc_canonical_decomposition (0xFB6D, decomposed); ASSERT (ret == -1); /* ARABIC LETTER VEH FINAL FORM */ ret = uc_canonical_decomposition (0xFB6B, decomposed); ASSERT (ret == -1); /* ARABIC LETTER VEH ISOLATED FORM */ ret = uc_canonical_decomposition (0xFB6A, decomposed); ASSERT (ret == -1); /* CIRCLED NUMBER FIFTEEN */ ret = uc_canonical_decomposition (0x246E, decomposed); ASSERT (ret == -1); /* TRADE MARK SIGN */ ret = uc_canonical_decomposition (0x2122, decomposed); ASSERT (ret == -1); /* LATIN SUBSCRIPT SMALL LETTER I */ ret = uc_canonical_decomposition (0x1D62, decomposed); ASSERT (ret == -1); /* PRESENTATION FORM FOR VERTICAL LEFT PARENTHESIS */ ret = uc_canonical_decomposition (0xFE35, decomposed); ASSERT (ret == -1); /* FULLWIDTH LATIN CAPITAL LETTER A */ ret = uc_canonical_decomposition (0xFF21, decomposed); ASSERT (ret == -1); /* HALFWIDTH IDEOGRAPHIC COMMA */ ret = uc_canonical_decomposition (0xFF64, decomposed); ASSERT (ret == -1); /* SMALL IDEOGRAPHIC COMMA */ ret = uc_canonical_decomposition (0xFE51, decomposed); ASSERT (ret == -1); /* SQUARE MHZ */ ret = uc_canonical_decomposition (0x3392, decomposed); ASSERT (ret == -1); /* VULGAR FRACTION THREE EIGHTHS */ ret = uc_canonical_decomposition (0x215C, decomposed); ASSERT (ret == -1); /* MICRO SIGN */ ret = uc_canonical_decomposition (0x00B5, decomposed); ASSERT (ret == -1); /* ARABIC LIGATURE SALLALLAHOU ALAYHE WASALLAM */ ret = uc_canonical_decomposition (0xFDFA, decomposed); ASSERT (ret == -1); /* HANGUL SYLLABLE GEUL */ ret = uc_canonical_decomposition (0xAE00, decomposed); /* See the clarification at , section "Clarification of Hangul Jamo Handling". */ #if 1 ASSERT (ret == 2); ASSERT (decomposed[0] == 0xADF8); ASSERT (decomposed[1] == 0x11AF); #else ASSERT (ret == 3); ASSERT (decomposed[0] == 0x1100); ASSERT (decomposed[1] == 0x1173); ASSERT (decomposed[2] == 0x11AF); #endif /* HANGUL SYLLABLE GEU */ ret = uc_canonical_decomposition (0xADF8, decomposed); ASSERT (ret == 2); ASSERT (decomposed[0] == 0x1100); ASSERT (decomposed[1] == 0x1173); return 0; }