bug 728866 - pt 3 - add support for Hebrew presentation forms, excluded from canonical composition. r=smontagu
authorJonathan Kew <jfkthame@gmail.com>
Mon, 20 Feb 2012 20:39:59 +0000
changeset 87250 969493463a87c7942ac30d75c9548e547f03b976
parent 87249 2c3f7f3ace8913fa91417c4dc9c462b536a031fb
child 87251 c714722a7aad88ec64b96c59adc22de4608bd6d5
push id22103
push userbmo@edmorley.co.uk
push dateTue, 21 Feb 2012 12:01:45 +0000
treeherdermozilla-central@4038ffaa5d82 [default view] [failures only]
perfherder[talos] [build metrics] [platform microbench] (compared to previous push)
reviewerssmontagu
bugs728866
milestone13.0a1
first release with
nightly linux32
nightly linux64
nightly mac
nightly win32
nightly win64
last release without
nightly linux32
nightly linux64
nightly mac
nightly win32
nightly win64
bug 728866 - pt 3 - add support for Hebrew presentation forms, excluded from canonical composition. r=smontagu
gfx/thebes/gfxHarfBuzzShaper.cpp
--- a/gfx/thebes/gfxHarfBuzzShaper.cpp
+++ b/gfx/thebes/gfxHarfBuzzShaper.cpp
@@ -695,24 +695,138 @@ HBGetCombiningClass(hb_unicode_funcs_t *
 }
 
 static unsigned int
 HBGetEastAsianWidth(hb_unicode_funcs_t *ufuncs, hb_codepoint_t aCh, void *user_data)
 {
     return gfxUnicodeProperties::GetEastAsianWidth(aCh);
 }
 
+// Hebrew presentation forms with dagesh, for characters 0x05D0..0x05EA;
+// note that some letters do not have a dagesh presForm encoded
+static const PRUnichar sDageshForms[0x05EA - 0x05D0 + 1] = {
+    0xFB30, // ALEF
+    0xFB31, // BET
+    0xFB32, // GIMEL
+    0xFB33, // DALET
+    0xFB34, // HE
+    0xFB35, // VAV
+    0xFB36, // ZAYIN
+    0, // HET
+    0xFB38, // TET
+    0xFB39, // YOD
+    0xFB3A, // FINAL KAF
+    0xFB3B, // KAF
+    0xFB3C, // LAMED
+    0, // FINAL MEM
+    0xFB3E, // MEM
+    0, // FINAL NUN
+    0xFB40, // NUN
+    0xFB41, // SAMEKH
+    0, // AYIN
+    0xFB43, // FINAL PE
+    0xFB44, // PE
+    0, // FINAL TSADI
+    0xFB46, // TSADI
+    0xFB47, // QOF
+    0xFB48, // RESH
+    0xFB49, // SHIN
+    0xFB4A // TAV
+};
+
 static hb_bool_t
 HBUnicodeCompose(hb_unicode_funcs_t *ufuncs,
                  hb_codepoint_t      a,
                  hb_codepoint_t      b,
                  hb_codepoint_t     *ab,
                  void               *user_data)
 {
-    return nsUnicodeNormalizer::Compose(a, b, ab);
+    hb_bool_t found = nsUnicodeNormalizer::Compose(a, b, ab);
+
+    if (!found && (b & 0x1fff80) == 0x0580) {
+        // special-case Hebrew presentation forms that are excluded from
+        // standard normalization, but wanted for old fonts
+        switch (b) {
+        case 0x05B4: // HIRIQ
+            if (a == 0x05D9) { // YOD
+                *ab = 0xFB1D;
+                found = true;
+            }
+            break;
+        case 0x05B7: // patah
+            if (a == 0x05F2) { // YIDDISH YOD YOD
+                *ab = 0xFB1F;
+                found = true;
+            } else if (a == 0x05D0) { // ALEF
+                *ab = 0xFB2E;
+                found = true;
+            }
+            break;
+        case 0x05B8: // QAMATS
+            if (a == 0x05D0) { // ALEF
+                *ab = 0xFB2F;
+                found = true;
+            }
+            break;
+        case 0x05B9: // HOLAM
+            if (a == 0x05D5) { // VAV
+                *ab = 0xFB4B;
+                found = true;
+            }
+            break;
+        case 0x05BC: // DAGESH
+            if (a >= 0x05D0 && a <= 0x05EA) {
+                *ab = sDageshForms[a - 0x05D0];
+                found = (*ab != 0);
+            } else if (a == 0xFB2A) { // SHIN WITH SHIN DOT
+                *ab = 0xFB2C;
+                found = true;
+            } else if (a == 0xFB2B) { // SHIN WITH SIN DOT
+                *ab = 0xFB2D;
+                found = true;
+            }
+            break;
+        case 0x05BF: // RAFE
+            switch (a) {
+            case 0x05D1: // BET
+                *ab = 0xFB4C;
+                found = true;
+                break;
+            case 0x05DB: // KAF
+                *ab = 0xFB4D;
+                found = true;
+                break;
+            case 0x05E4: // PE
+                *ab = 0xFB4E;
+                found = true;
+                break;
+            }
+            break;
+        case 0x05C1: // SHIN DOT
+            if (a == 0x05E9) { // SHIN
+                *ab = 0xFB2A;
+                found = true;
+            } else if (a == 0xFB49) { // SHIN WITH DAGESH
+                *ab = 0xFB2C;
+                found = true;
+            }
+            break;
+        case 0x05C2: // SIN DOT
+            if (a == 0x05E9) { // SHIN
+                *ab = 0xFB2B;
+                found = true;
+            } else if (a == 0xFB49) { // SHIN WITH DAGESH
+                *ab = 0xFB2D;
+                found = true;
+            }
+            break;
+        }
+    }
+
+    return found;
 }
 
 static hb_bool_t
 HBUnicodeDecompose(hb_unicode_funcs_t *ufuncs,
                    hb_codepoint_t      ab,
                    hb_codepoint_t     *a,
                    hb_codepoint_t     *b,
                    void               *user_data)