Shorthand format controls should not be ignored

Fixes https://github.com/behdad/harfbuzz/issues/503

diff --git a/src/hb-unicode-private.hh b/src/hb-unicode-private.hh
index 34513e1324c72786d881722f4fba0ae4fbd2dd6a..cd6c67c872f00b1d0e4e4025161d94c6ad9aa858 100644 (file)
--- a/src/hb-unicode-private.hh
+++ b/src/hb-unicode-private.hh
@@ -141,6 +141,7 @@ HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS_SIMPLE
    * we do NOT want to hide them, as the way Uniscribe has implemented them
    * is with regular spacing glyphs, and that's the way fonts are made to work.
    * As such, we make exceptions for those four.
+   * Also ignoring U+1BCA0..1BCA3. https://github.com/behdad/harfbuzz/issues/503
    *
    * Unicode 7.0:
    * $ grep '; Default_Ignorable_Code_Point ' DerivedCoreProperties.txt | sed 's/;.*#/#/'
@@ -197,8 +198,7 @@ HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS_SIMPLE
     {
       /* Other planes */
       switch (plane) {
-       case 0x01: return hb_in_ranges<hb_codepoint_t> (ch, 0x1BCA0u, 0x1BCA3u,
-                                           0x1D173u, 0x1D17Au);
+       case 0x01: return hb_in_range<hb_codepoint_t> (ch, 0x1D173u, 0x1D17Au);
        case 0x0E: return hb_in_range<hb_codepoint_t> (ch, 0xE0000u, 0xE0FFFu);
        default: return false;
       }