]> git.saurik.com Git - apple/javascriptcore.git/blobdiff - wtf/unicode/glib/UnicodeGLib.cpp
JavaScriptCore-554.1.tar.gz
[apple/javascriptcore.git] / wtf / unicode / glib / UnicodeGLib.cpp
diff --git a/wtf/unicode/glib/UnicodeGLib.cpp b/wtf/unicode/glib/UnicodeGLib.cpp
new file mode 100644 (file)
index 0000000..a779b36
--- /dev/null
@@ -0,0 +1,214 @@
+/*
+ *  Copyright (C) 2008 Jürg Billeter <j@bitron.ch>
+ *  Copyright (C) 2008 Dominik Röttsches <dominik.roettsches@access-company.com>
+ *
+ *  This library is free software; you can redistribute it and/or
+ *  modify it under the terms of the GNU Library General Public
+ *  License as published by the Free Software Foundation; either
+ *  version 2 of the License, or (at your option) any later version.
+ *
+ *  This library 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
+ *  Library General Public License for more details.
+ *
+ *  You should have received a copy of the GNU Library General Public License
+ *  along with this library; see the file COPYING.LIB.  If not, write to
+ *  the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
+ *  Boston, MA 02110-1301, USA.
+ *
+ */
+
+#include "UnicodeGLib.h"
+
+namespace WTF {
+namespace Unicode {
+
+UChar32 foldCase(UChar32 ch)
+{
+    GOwnPtr<GError> gerror;
+
+    GOwnPtr<char> utf8char;
+    utf8char.set(g_ucs4_to_utf8(reinterpret_cast<gunichar*>(&ch), 1, 0, 0, &gerror.outPtr()));
+    if (gerror)
+        return ch;
+
+    GOwnPtr<char> utf8caseFolded;
+    utf8caseFolded.set(g_utf8_casefold(utf8char.get(), -1));
+
+    GOwnPtr<gunichar> ucs4Result;
+    ucs4Result.set(g_utf8_to_ucs4_fast(utf8caseFolded.get(), -1, 0));
+
+    return *ucs4Result;
+}
+
+int foldCase(UChar* result, int resultLength, const UChar* src, int srcLength, bool* error)
+{
+    *error = false;
+    GOwnPtr<GError> gerror;
+
+    GOwnPtr<char> utf8src;
+    utf8src.set(g_utf16_to_utf8(src, srcLength, 0, 0, &gerror.outPtr()));
+    if (gerror) {
+        *error = true;
+        return -1;
+    }
+
+    GOwnPtr<char> utf8result;
+    utf8result.set(g_utf8_casefold(utf8src.get(), -1));
+
+    long utf16resultLength = -1;
+    GOwnPtr<UChar> utf16result;
+    utf16result.set(g_utf8_to_utf16(utf8result.get(), -1, 0, &utf16resultLength, &gerror.outPtr()));
+    if (gerror) {
+        *error = true;
+        return -1;
+    }
+
+    if (utf16resultLength > resultLength) {
+        *error = true;
+        return utf16resultLength;
+    }
+    memcpy(result, utf16result.get(), utf16resultLength * sizeof(UChar));
+
+    return utf16resultLength;
+}
+
+int toLower(UChar* result, int resultLength, const UChar* src, int srcLength, bool* error)
+{
+    *error = false;
+    GOwnPtr<GError> gerror;
+
+    GOwnPtr<char> utf8src;
+    utf8src.set(g_utf16_to_utf8(src, srcLength, 0, 0, &gerror.outPtr()));
+    if (gerror) {
+        *error = true;
+        return -1;
+    }
+
+    GOwnPtr<char> utf8result;
+    utf8result.set(g_utf8_strdown(utf8src.get(), -1));
+
+    long utf16resultLength = -1;
+    GOwnPtr<UChar> utf16result;
+    utf16result.set(g_utf8_to_utf16(utf8result.get(), -1, 0, &utf16resultLength, &gerror.outPtr()));
+    if (gerror) {
+        *error = true;
+        return -1;
+    }
+
+    if (utf16resultLength > resultLength) {
+        *error = true;
+        return utf16resultLength;
+    }
+    memcpy(result, utf16result.get(), utf16resultLength * sizeof(UChar));
+
+    return utf16resultLength;
+}
+
+int toUpper(UChar* result, int resultLength, const UChar* src, int srcLength, bool* error)
+{
+    *error = false;
+    GOwnPtr<GError> gerror;
+
+    GOwnPtr<char> utf8src;
+    utf8src.set(g_utf16_to_utf8(src, srcLength, 0, 0, &gerror.outPtr()));
+    if (gerror) {
+        *error = true;
+        return -1;
+    }
+
+    GOwnPtr<char> utf8result;
+    utf8result.set(g_utf8_strup(utf8src.get(), -1));
+
+    long utf16resultLength = -1;
+    GOwnPtr<UChar> utf16result;
+    utf16result.set(g_utf8_to_utf16(utf8result.get(), -1, 0, &utf16resultLength, &gerror.outPtr()));
+    if (gerror) {
+        *error = true;
+        return -1;
+    }
+
+    if (utf16resultLength > resultLength) {
+        *error = true;
+        return utf16resultLength;
+    }
+    memcpy(result, utf16result.get(), utf16resultLength * sizeof(UChar));
+
+    return utf16resultLength;
+}
+
+Direction direction(UChar32 c)
+{
+    PangoBidiType type = pango_bidi_type_for_unichar(c);
+    switch (type) {
+    case PANGO_BIDI_TYPE_L:
+        return LeftToRight;
+    case PANGO_BIDI_TYPE_R:
+        return RightToLeft;
+    case PANGO_BIDI_TYPE_AL:
+        return RightToLeftArabic;
+    case PANGO_BIDI_TYPE_LRE:
+        return LeftToRightEmbedding;
+    case PANGO_BIDI_TYPE_RLE:
+        return RightToLeftEmbedding;
+    case PANGO_BIDI_TYPE_LRO:
+        return LeftToRightOverride;
+    case PANGO_BIDI_TYPE_RLO:
+        return RightToLeftOverride;
+    case PANGO_BIDI_TYPE_PDF:
+        return PopDirectionalFormat;
+    case PANGO_BIDI_TYPE_EN:
+        return EuropeanNumber;
+    case PANGO_BIDI_TYPE_AN:
+        return ArabicNumber;
+    case PANGO_BIDI_TYPE_ES:
+        return EuropeanNumberSeparator;
+    case PANGO_BIDI_TYPE_ET:
+        return EuropeanNumberTerminator;
+    case PANGO_BIDI_TYPE_CS:
+        return CommonNumberSeparator;
+    case PANGO_BIDI_TYPE_NSM:
+        return NonSpacingMark;
+    case PANGO_BIDI_TYPE_BN:
+        return BoundaryNeutral;
+    case PANGO_BIDI_TYPE_B:
+        return BlockSeparator;
+    case PANGO_BIDI_TYPE_S:
+        return SegmentSeparator;
+    case PANGO_BIDI_TYPE_WS:
+        return WhiteSpaceNeutral;
+    default:
+        return OtherNeutral;
+    }
+}
+
+int umemcasecmp(const UChar* a, const UChar* b, int len)
+{
+    GOwnPtr<char> utf8a;
+    GOwnPtr<char> utf8b;
+
+    utf8a.set(g_utf16_to_utf8(a, len, 0, 0, 0));
+    utf8b.set(g_utf16_to_utf8(b, len, 0, 0, 0));
+
+    GOwnPtr<char> foldedA;
+    GOwnPtr<char> foldedB;
+
+    foldedA.set(g_utf8_casefold(utf8a.get(), -1));
+    foldedB.set(g_utf8_casefold(utf8b.get(), -1));
+
+    // FIXME: umemcasecmp needs to mimic u_memcasecmp of icu
+    // from the ICU docs:
+    // "Compare two strings case-insensitively using full case folding.
+    // his is equivalent to u_strcmp(u_strFoldCase(s1, n, options), u_strFoldCase(s2, n, options))."
+    //
+    // So it looks like we don't need the full g_utf8_collate here,
+    // but really a bitwise comparison of casefolded unicode chars (not utf-8 bytes).
+    // As there is no direct equivalent to this icu function in GLib, for now
+    // we'll use g_utf8_collate():
+
+    return g_utf8_collate(foldedA.get(), foldedB.get());
+}
+
+}
+}