Update harfbuzz-ng to 1a5a91dc0d8bf4b72a2f22dc6300b06ad7000b79.

third_party/harfbuzz-ng/src/hb-ot-tag.cc

 /*
- * Copyright (C) 2009  Red Hat, Inc.
+ * Copyright © 2009  Red Hat, Inc.
+ * Copyright © 2011  Google, Inc.
  *
  *  This is part of HarfBuzz, a text shaping library.
  *
  * Permission is hereby granted, without written agreement and without
  * license or royalty fees, to use, copy, modify, and distribute this
  * software and its documentation for any purpose, provided that the
  * above copyright notice and the following two paragraphs appear in
  * all copies of this software.
  *
  * IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
  * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
  * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
  * IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
  * DAMAGE.
  *
  * THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
  * BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
  * FITNESS FOR A PARTICULAR PURPOSE.  THE SOFTWARE PROVIDED HEREUNDER IS
  * ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
  * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
  *
  * Red Hat Author(s): Behdad Esfahbod
+ * Google Author(s): Behdad Esfahbod
  */
 
-#include "hb-private.h"
+#include "hb-private.hh"
 #include "hb-ot.h"
 
 #include <string.h>
 
-HB_BEGIN_DECLS
 
 
+/* hb_script_t */
+
+static hb_tag_t
+hb_ot_old_tag_from_script (hb_script_t script)
+{
+  switch ((hb_tag_t) script) {
+    case HB_SCRIPT_INVALID:             return HB_OT_TAG_DEFAULT_SCRIPT;
+
+    /* KATAKANA and HIRAGANA both map to 'kana' */
+    case HB_SCRIPT_HIRAGANA:            return HB_TAG('k','a','n','a');
+
+    /* Spaces at the end are preserved, unlike ISO 15924 */
+    case HB_SCRIPT_LAO:                 return HB_TAG('l','a','o',' ');
+    case HB_SCRIPT_YI:                  return HB_TAG('y','i',' ',' ');
+    /* Unicode-5.0 additions */
+    case HB_SCRIPT_NKO:                 return HB_TAG('n','k','o',' ');
+    /* Unicode-5.1 additions */
+    case HB_SCRIPT_VAI:                 return HB_TAG('v','a','i',' ');
+    /* Unicode-5.2 additions */
+    /* Unicode-6.0 additions */
+  }
+
+  /* Else, just change first char to lowercase and return */
+  return ((hb_tag_t) script) | 0x20000000;
+}
+
+static hb_script_t
+hb_ot_old_tag_to_script (hb_tag_t tag)
+{
+  if (unlikely (tag == HB_OT_TAG_DEFAULT_SCRIPT))
+    return HB_SCRIPT_INVALID;
+
+  /* This side of the conversion is fully algorithmic. */
+
+  /* Any spaces at the end of the tag are replaced by repeating the last
+   * letter.  Eg 'nko ' -> 'Nkoo' */
+  if (unlikely ((tag & 0x0000FF00) == 0x00002000))
+    tag |= (tag >> 8) & 0x0000FF00; /* Copy second letter to third */
+  if (unlikely ((tag & 0x000000FF) == 0x00000020))
+    tag |= (tag >> 8) & 0x000000FF; /* Copy third letter to fourth */
+
+  /* Change first char to uppercase and return */
+  return (hb_script_t) (tag & ~0x20000000);
+}
+
+static hb_tag_t
+hb_ot_new_tag_from_script (hb_script_t script)
+{
+  switch ((hb_tag_t) script) {
+    case HB_SCRIPT_BENGALI:             return HB_TAG('b','n','g','2');
+    case HB_SCRIPT_DEVANAGARI:          return HB_TAG('d','e','v','2');
+    case HB_SCRIPT_GUJARATI:            return HB_TAG('g','j','r','2');
+    case HB_SCRIPT_GURMUKHI:            return HB_TAG('g','u','r','2');
+    case HB_SCRIPT_KANNADA:             return HB_TAG('k','n','d','2');
+    case HB_SCRIPT_MALAYALAM:           return HB_TAG('m','l','m','2');
+    case HB_SCRIPT_ORIYA:               return HB_TAG('o','r','y','2');
+    case HB_SCRIPT_TAMIL:               return HB_TAG('t','m','l','2');
+    case HB_SCRIPT_TELUGU:              return HB_TAG('t','e','l','2');
+  }
+
+  return HB_OT_TAG_DEFAULT_SCRIPT;
+}
+
+static hb_script_t
+hb_ot_new_tag_to_script (hb_tag_t tag)
+{
+  switch (tag) {
+    case HB_TAG('b','n','g','2'):       return HB_SCRIPT_BENGALI;
+    case HB_TAG('d','e','v','2'):       return HB_SCRIPT_DEVANAGARI;
+    case HB_TAG('g','j','r','2'):       return HB_SCRIPT_GUJARATI;
+    case HB_TAG('g','u','r','2'):       return HB_SCRIPT_GURMUKHI;
+    case HB_TAG('k','n','d','2'):       return HB_SCRIPT_KANNADA;
+    case HB_TAG('m','l','m','2'):       return HB_SCRIPT_MALAYALAM;
+    case HB_TAG('o','r','y','2'):       return HB_SCRIPT_ORIYA;
+    case HB_TAG('t','m','l','2'):       return HB_SCRIPT_TAMIL;
+    case HB_TAG('t','e','l','2'):       return HB_SCRIPT_TELUGU;
+  }
+
+  return HB_SCRIPT_UNKNOWN;
+}
+
 /*
  * Complete list at:
- * http://www.microsoft.com/typography/otspec/scripttags.htm
+ * https://www.microsoft.com/typography/otspec/scripttags.htm
+ * https://www.microsoft.com/typography/otspec160/scripttagsProposed.htm
+ *
+ * Most of the script tags are the same as the ISO 15924 tag but lowercased.
+ * So we just do that, and handle the exceptional cases in a switch.
  */
-static const hb_tag_t ot_scripts[][3] = {
-  {HB_TAG('D','F','L','T')},    /* HB_SCRIPT_COMMON */
-  {HB_TAG('D','F','L','T')},    /* HB_SCRIPT_INHERITED */
-  {HB_TAG('a','r','a','b')},    /* HB_SCRIPT_ARABIC */
-  {HB_TAG('a','r','m','n')},    /* HB_SCRIPT_ARMENIAN */
-  {HB_TAG('b','n','g','2'), HB_TAG('b','e','n','g')},   /* HB_SCRIPT_BENGALI */
-  {HB_TAG('b','o','p','o')},    /* HB_SCRIPT_BOPOMOFO */
-  {HB_TAG('c','h','e','r')},    /* HB_SCRIPT_CHEROKEE */
-  {HB_TAG('c','o','p','t')},    /* HB_SCRIPT_COPTIC */
-  {HB_TAG('c','y','r','l')},    /* HB_SCRIPT_CYRILLIC */
-  {HB_TAG('d','s','r','t')},    /* HB_SCRIPT_DESERET */
-  {HB_TAG('d','e','v','2'), HB_TAG('d','e','v','a')},   /* HB_SCRIPT_DEVANAGARI */
-  {HB_TAG('e','t','h','i')},    /* HB_SCRIPT_ETHIOPIC */
-  {HB_TAG('g','e','o','r')},    /* HB_SCRIPT_GEORGIAN */
-  {HB_TAG('g','o','t','h')},    /* HB_SCRIPT_GOTHIC */
-  {HB_TAG('g','r','e','k')},    /* HB_SCRIPT_GREEK */
-  {HB_TAG('g','j','r','2'), HB_TAG('g','u','j','r')},   /* HB_SCRIPT_GUJARATI */
-  {HB_TAG('g','u','r','2'), HB_TAG('g','u','r','u')},   /* HB_SCRIPT_GURMUKHI */
-  {HB_TAG('h','a','n','i')},    /* HB_SCRIPT_HAN */
-  {HB_TAG('h','a','n','g')},    /* HB_SCRIPT_HANGUL */
-  {HB_TAG('h','e','b','r')},    /* HB_SCRIPT_HEBREW */
-  {HB_TAG('k','a','n','a')},    /* HB_SCRIPT_HIRAGANA */
-  {HB_TAG('k','n','d','2'), HB_TAG('k','n','d','a')},   /* HB_SCRIPT_KANNADA */
-  {HB_TAG('k','a','n','a')},    /* HB_SCRIPT_KATAKANA */
-  {HB_TAG('k','h','m','r')},    /* HB_SCRIPT_KHMER */
-  {HB_TAG('l','a','o',' ')},    /* HB_SCRIPT_LAO */
-  {HB_TAG('l','a','t','n')},    /* HB_SCRIPT_LATIN */
-  {HB_TAG('m','l','m','2'), HB_TAG('m','l','y','m')},   /* HB_SCRIPT_MALAYALAM */
-  {HB_TAG('m','o','n','g')},    /* HB_SCRIPT_MONGOLIAN */
-  {HB_TAG('m','y','m','r')},    /* HB_SCRIPT_MYANMAR */
-  {HB_TAG('o','g','a','m')},    /* HB_SCRIPT_OGHAM */
-  {HB_TAG('i','t','a','l')},    /* HB_SCRIPT_OLD_ITALIC */
-  {HB_TAG('o','r','y','2'), HB_TAG('o','r','y','a')},   /* HB_SCRIPT_ORIYA */
-  {HB_TAG('r','u','n','r')},    /* HB_SCRIPT_RUNIC */
-  {HB_TAG('s','i','n','h')},    /* HB_SCRIPT_SINHALA */
-  {HB_TAG('s','y','r','c')},    /* HB_SCRIPT_SYRIAC */
-  {HB_TAG('t','m','l','2'), HB_TAG('t','a','m','l')},   /* HB_SCRIPT_TAMIL */
-  {HB_TAG('t','e','l','2'), HB_TAG('t','e','l','u')},   /* HB_SCRIPT_TELUGU */
-  {HB_TAG('t','h','a','a')},    /* HB_SCRIPT_THAANA */
-  {HB_TAG('t','h','a','i')},    /* HB_SCRIPT_THAI */
-  {HB_TAG('t','i','b','t')},    /* HB_SCRIPT_TIBETAN */
-  {HB_TAG('c','a','n','s')},    /* HB_SCRIPT_CANADIAN_ABORIGINAL */
-  {HB_TAG('y','i',' ',' ')},    /* HB_SCRIPT_YI */
-  {HB_TAG('t','g','l','g')},    /* HB_SCRIPT_TAGALOG */
-  {HB_TAG('h','a','n','o')},    /* HB_SCRIPT_HANUNOO */
-  {HB_TAG('b','u','h','d')},    /* HB_SCRIPT_BUHID */
-  {HB_TAG('t','a','g','b')},    /* HB_SCRIPT_TAGBANWA */
 
-  /* Unicode-4.0 additions */
-  {HB_TAG('b','r','a','i')},    /* HB_SCRIPT_BRAILLE */
-  {HB_TAG('c','p','r','t')},    /* HB_SCRIPT_CYPRIOT */
-  {HB_TAG('l','i','m','b')},    /* HB_SCRIPT_LIMBU */
-  {HB_TAG('o','s','m','a')},    /* HB_SCRIPT_OSMANYA */
-  {HB_TAG('s','h','a','w')},    /* HB_SCRIPT_SHAVIAN */
-  {HB_TAG('l','i','n','b')},    /* HB_SCRIPT_LINEAR_B */
-  {HB_TAG('t','a','l','e')},    /* HB_SCRIPT_TAI_LE */
-  {HB_TAG('u','g','a','r')},    /* HB_SCRIPT_UGARITIC */
+void
+hb_ot_tags_from_script (hb_script_t  script,
+                        hb_tag_t    *script_tag_1,
+                        hb_tag_t    *script_tag_2)
+{
+  hb_tag_t new_tag;
 
-  /* Unicode-4.1 additions */
-  {HB_TAG('t','a','l','u')},    /* HB_SCRIPT_NEW_TAI_LUE */
-  {HB_TAG('b','u','g','i')},    /* HB_SCRIPT_BUGINESE */
-  {HB_TAG('g','l','a','g')},    /* HB_SCRIPT_GLAGOLITIC */
-  {HB_TAG('t','f','n','g')},    /* HB_SCRIPT_TIFINAGH */
-  {HB_TAG('s','y','l','o')},    /* HB_SCRIPT_SYLOTI_NAGRI */
-  {HB_TAG('x','p','e','o')},    /* HB_SCRIPT_OLD_PERSIAN */
-  {HB_TAG('k','h','a','r')},    /* HB_SCRIPT_KHAROSHTHI */
+  *script_tag_2 = HB_OT_TAG_DEFAULT_SCRIPT;
+  *script_tag_1 = hb_ot_old_tag_from_script (script);
 
-  /* Unicode-5.0 additions */
-  {HB_TAG('D','F','L','T')},    /* HB_SCRIPT_UNKNOWN */
-  {HB_TAG('b','a','l','i')},    /* HB_SCRIPT_BALINESE */
-  {HB_TAG('x','s','u','x')},    /* HB_SCRIPT_CUNEIFORM */
-  {HB_TAG('p','h','n','x')},    /* HB_SCRIPT_PHOENICIAN */
-  {HB_TAG('p','h','a','g')},    /* HB_SCRIPT_PHAGS_PA */
-  {HB_TAG('n','k','o',' ')},    /* HB_SCRIPT_NKO */
-
-  /* Unicode-5.1 additions */
-  {HB_TAG('k','a','l','i')},    /* HB_SCRIPT_KAYAH_LI */
-  {HB_TAG('l','e','p','c')},    /* HB_SCRIPT_LEPCHA */
-  {HB_TAG('r','j','n','g')},    /* HB_SCRIPT_REJANG */
-  {HB_TAG('s','u','n','d')},    /* HB_SCRIPT_SUNDANESE */
-  {HB_TAG('s','a','u','r')},    /* HB_SCRIPT_SAURASHTRA */
-  {HB_TAG('c','h','a','m')},    /* HB_SCRIPT_CHAM */
-  {HB_TAG('o','l','c','k')},    /* HB_SCRIPT_OL_CHIKI */
-  {HB_TAG('v','a','i',' ')},    /* HB_SCRIPT_VAI */
-  {HB_TAG('c','a','r','i')},    /* HB_SCRIPT_CARIAN */
-  {HB_TAG('l','y','c','i')},    /* HB_SCRIPT_LYCIAN */
-  {HB_TAG('l','y','d','i')},    /* HB_SCRIPT_LYDIAN */
-
-  /* Unicode-5.2 additions */
-  {HB_TAG('a','v','s','t')},    /* HB_SCRIPT_AVESTAN */
-  {HB_TAG('b','a','m','u')},    /* HB_SCRIPT_BAMUM */
-  {HB_TAG('e','g','y','p')},    /* HB_SCRIPT_EGYPTIAN_HIEROGLYPHS */
-  {HB_TAG('a','r','m','i')},    /* HB_SCRIPT_IMPERIAL_ARAMAIC */
-  {HB_TAG('p','h','l','i')},    /* HB_SCRIPT_INSCRIPTIONAL_PAHLAVI */
-  {HB_TAG('p','r','t','i')},    /* HB_SCRIPT_INSCRIPTIONAL_PARTHIAN */
-  {HB_TAG('j','a','v','a')},    /* HB_SCRIPT_JAVANESE */
-  {HB_TAG('k','t','h','i')},    /* HB_SCRIPT_KAITHI */
-  {HB_TAG('l','i','s','u')},    /* HB_SCRIPT_LISU */
-  {HB_TAG('m','y','e','i')},    /* HB_SCRIPT_MEETEI_MAYEK */
-  {HB_TAG('s','a','r','b')},    /* HB_SCRIPT_OLD_SOUTH_ARABIAN */
-  {HB_TAG('o','r','k','h')},    /* HB_SCRIPT_OLD_TURKIC */
-  {HB_TAG('s','a','m','r')},    /* HB_SCRIPT_SAMARITAN */
-  {HB_TAG('l','a','n','a')},    /* HB_SCRIPT_TAI_THAM */
-  {HB_TAG('t','a','v','t')},    /* HB_SCRIPT_TAI_VIET */
-
-  /* Unicode-6.0 additions */
-  {HB_TAG('b','a','t','k')},    /* HB_SCRIPT_BATAK */
-  {HB_TAG('b','r','a','h')},    /* HB_SCRIPT_BRAHMI */
-  {HB_TAG('m','a','n','d')}     /* HB_SCRIPT_MANDAIC */
-};
-
-const hb_tag_t *
-hb_ot_tags_from_script (hb_script_t script)
-{
-  static const hb_tag_t def_tag[] = {HB_OT_TAG_DEFAULT_SCRIPT, HB_TAG_NONE};
-
-  if (unlikely ((unsigned int) script >= ARRAY_LENGTH (ot_scripts)))
-    return def_tag;
-
-  return ot_scripts[script];
+  new_tag = hb_ot_new_tag_from_script (script);
+  if (unlikely (new_tag != HB_OT_TAG_DEFAULT_SCRIPT)) {
+    *script_tag_2 = *script_tag_1;
+    *script_tag_1 = new_tag;
+  }
 }
 
 hb_script_t
 hb_ot_tag_to_script (hb_tag_t tag)
 {
-  int i;
+  if (unlikely ((tag & 0x000000FF) == '2'))
+    return hb_ot_new_tag_to_script (tag);
 
-  for (i = 0; i < ARRAY_LENGTH (ot_scripts); i++) {
-    const hb_tag_t *p;
-    for (p = ot_scripts[i]; *p; p++)
-      if (tag == *p)
-        return i;
-  }
+  return hb_ot_old_tag_to_script (tag);
+}
 
-  return HB_SCRIPT_UNKNOWN;
-}
+
+/* hb_language_t */
 
 typedef struct {
   char language[6];
   hb_tag_t tag;
 } LangTag;
 
 /*
  * Complete list at:
  * http://www.microsoft.com/typography/otspec/languagetags.htm
  *
  * Generated by intersecting the OpenType language tag list from
  * Draft OpenType 1.5 spec, with with the ISO 639-3 codes from
  * 2008/08/04, matching on name, and finally adjusted manually.
  *
  * Many items still missing.  Those are commented out at the end.
  * Keep sorted for bsearch.
  */
+
 static const LangTag ot_languages[] = {
   {"aa",        HB_TAG('A','F','R',' ')},       /* Afar */
   {"ab",        HB_TAG('A','B','K',' ')},       /* Abkhazian */
   {"abq",       HB_TAG('A','B','A',' ')},       /* Abaza */
   {"ady",       HB_TAG('A','D','Y',' ')},       /* Adyghe */
   {"af",        HB_TAG('A','F','K',' ')},       /* Afrikaans */
   {"aiw",       HB_TAG('A','R','I',' ')},       /* Aari */
   {"am",        HB_TAG('A','M','H',' ')},       /* Amharic */
   {"ar",        HB_TAG('A','R','A',' ')},       /* Arabic */
   {"arn",       HB_TAG('M','A','P',' ')},       /* Mapudungun */
@@ skipping 266 matching lines @@
   {"wbm",       HB_TAG('W','A',' ',' ')},       /* Wa */
   {"wbr",       HB_TAG('W','A','G',' ')},       /* Wagdi */
   {"wo",        HB_TAG('W','L','F',' ')},       /* Wolof */
   {"xal",       HB_TAG('K','L','M',' ')},       /* Kalmyk */
   {"xh",        HB_TAG('X','H','S',' ')},       /* Xhosa */
   {"xom",       HB_TAG('K','M','O',' ')},       /* Komo (Sudan) */
   {"xsl",       HB_TAG('S','S','L',' ')},       /* South Slavey */
   {"yi",        HB_TAG('J','I','I',' ')},       /* Yiddish */
   {"yo",        HB_TAG('Y','B','A',' ')},       /* Yoruba */
   {"yso",       HB_TAG('N','I','S',' ')},       /* Nisi (China) */
-  {"zh-cn",     HB_TAG('Z','H','S',' ')},       /* Chinese (China) */
-  {"zh-hk",     HB_TAG('Z','H','H',' ')},       /* Chinese (Hong Kong) */
-  {"zh-mo",     HB_TAG('Z','H','T',' ')},       /* Chinese (Macao) */
-  {"zh-sg",     HB_TAG('Z','H','S',' ')},       /* Chinese (Singapore) */
-  {"zh-tw",     HB_TAG('Z','H','T',' ')},       /* Chinese (Taiwan) */
   {"zne",       HB_TAG('Z','N','D',' ')},       /* Zande */
   {"zu",        HB_TAG('Z','U','L',' ')}        /* Zulu */
 
   /* I couldn't find the language id for these */
 
 /*{"??",        HB_TAG('A','G','W',' ')},*/     /* Agaw */
 /*{"??",        HB_TAG('A','L','S',' ')},*/     /* Alsatian */
 /*{"??",        HB_TAG('A','L','T',' ')},*/     /* Altai */
 /*{"??",        HB_TAG('A','R','K',' ')},*/     /* Arakanese */
 /*{"??",        HB_TAG('A','T','H',' ')},*/     /* Athapaskan */
@@ skipping 95 matching lines @@
 /*{"??",        HB_TAG('T','U','A',' ')},*/     /* Turoyo Aramaic */
 /*{"??",        HB_TAG('T','U','V',' ')},*/     /* Tuvin */
 /*{"??",        HB_TAG('W','C','R',' ')},*/     /* West-Cree */
 /*{"??",        HB_TAG('X','B','D',' ')},*/     /* Tai Lue */
 /*{"??",        HB_TAG('Y','C','R',' ')},*/     /* Y-Cree */
 /*{"??",        HB_TAG('Y','I','C',' ')},*/     /* Yi Classic */
 /*{"??",        HB_TAG('Y','I','M',' ')},*/     /* Yi Modern */
 /*{"??",        HB_TAG('Z','H','P',' ')},*/     /* Chinese Phonetic */
 };
 
+static const LangTag ot_languages_zh[] = {
+  {"zh-cn",     HB_TAG('Z','H','S',' ')},       /* Chinese (China) */
+  {"zh-hk",     HB_TAG('Z','H','H',' ')},       /* Chinese (Hong Kong) */
+  {"zh-mo",     HB_TAG('Z','H','T',' ')},       /* Chinese (Macao) */
+  {"zh-sg",     HB_TAG('Z','H','S',' ')},       /* Chinese (Singapore) */
+  {"zh-tw",     HB_TAG('Z','H','T',' ')}        /* Chinese (Taiwan) */
+};
+
 static int
 lang_compare_first_component (const char *a,
                               const char *b)
 {
   unsigned int da, db;
   const char *p;
 
-  p = strstr (a, "-");
+  p = strchr (a, '-');
   da = p ? (unsigned int) (p - a) : strlen (a);
 
-  p = strstr (b, "-");
+  p = strchr (b, '-');
   db = p ? (unsigned int) (p - b) : strlen (b);
 
   return strncmp (a, b, MAX (da, db));
 }
 
 static hb_bool_t
 lang_matches (const char *lang_str, const char *spec)
 {
   unsigned int len = strlen (spec);
 
-  return lang_str && strncmp (lang_str, spec, len) == 0 &&
+  return strncmp (lang_str, spec, len) == 0 &&
          (lang_str[len] == '\0' || lang_str[len] == '-');
 }
 
 hb_tag_t
 hb_ot_tag_from_language (hb_language_t language)
 {
-  const char *lang_str;
-  LangTag *lang_tag;
+  const char *lang_str, *s;
+  const LangTag *lang_tag;
 
-  if (language == NULL)
+  if (language == HB_LANGUAGE_INVALID)
     return HB_OT_TAG_DEFAULT_LANGUAGE;
 
   lang_str = hb_language_to_string (language);
 
-  if (0 == strcmp (lang_str, "x-hbot")) {
+  s = strstr (lang_str, "x-hbot");
+  if (s) {
     char tag[4];
     int i;
-    lang_str += 6;
-#define IS_LETTER(c) (((c) >= 'a' && (c) <= 'z') || ((c) >= 'A' && (c) <= 'Z'))
-#define TO_UPPER(c) (((c) >= 'a' && (c) <= 'z') ? (c) + 'A' - 'a' : (c))
-    for (i = 0; i < 4 && IS_LETTER (lang_str[i]); i++)
-      tag[i] = TO_UPPER (lang_str[i]);
-    for (; i < 4; i++)
-      tag[i] = ' ';
-    return HB_TAG_STR (tag);
+    s += 6;
+    for (i = 0; i < 4 && ISALPHA (s[i]); i++)
+      tag[i] = TOUPPER (s[i]);
+    if (i) {
+      for (; i < 4; i++)
+        tag[i] = ' ';
+      return HB_TAG_CHAR4 (tag);
+    }
   }
 
-  /* find a language matching in the first component */
-  lang_tag = bsearch (lang_str, ot_languages,
-                      ARRAY_LENGTH (ot_languages), sizeof (LangTag),
-                      (hb_compare_func_t) lang_compare_first_component);
-
-  /* we now need to find the best language matching */
-  if (lang_tag)
-  {
-    hb_bool_t found = FALSE;
-
-    /* go to the final one matching in the first component */
-    while (lang_tag + 1 < ot_languages + ARRAY_LENGTH (ot_languages) &&
-           lang_compare_first_component (lang_str, (lang_tag + 1)->language) == 0)
-      lang_tag++;
-
-    /* go back, find which one matches completely */
-    while (lang_tag >= ot_languages &&
-           lang_compare_first_component (lang_str, lang_tag->language) == 0)
-    {
-      if (lang_matches (lang_str, lang_tag->language)) {
-        found = TRUE;
-        break;
-      }
-
-      lang_tag--;
-    }
-
-    if (!found)
-      lang_tag = NULL;
-  }
-
+  /* Find a language matching in the first component */
+  lang_tag = (LangTag *) bsearch (lang_str, ot_languages,
+                                  ARRAY_LENGTH (ot_languages), sizeof (LangTag),
+                                  (hb_compare_func_t) lang_compare_first_component);
   if (lang_tag)
     return lang_tag->tag;
 
+  /* Otherwise, check the Chinese ones */
+  if (0 == lang_compare_first_component (lang_str, "zh"))
+  {
+    unsigned int i;
+
+    for (i = 0; i < ARRAY_LENGTH (ot_languages_zh); i++)
+    {
+      lang_tag = &ot_languages_zh[i];
+      if (lang_matches (lang_tag->language, lang_str))
+        return lang_tag->tag;
+    }
+
+    /* Otherwise just return 'ZHS ' */
+    return HB_TAG('Z','H','S',' ');
+  }
+
+  s = strchr (lang_str, '-');
+  if (!s)
+    s = lang_str + strlen (lang_str);
+  if (s - lang_str == 3) {
+    /* Assume it's ISO-639-3 and upper-case and use it. */
+    return hb_tag_from_string (lang_str, s - lang_str) & ~0x20202000;
+  }
+
   return HB_OT_TAG_DEFAULT_LANGUAGE;
 }
 
 hb_language_t
 hb_ot_tag_to_language (hb_tag_t tag)
 {
   unsigned int i;
-  unsigned char buf[11] = "x-hbot";
+
+  if (tag == HB_OT_TAG_DEFAULT_LANGUAGE)
+    return NULL;
 
   for (i = 0; i < ARRAY_LENGTH (ot_languages); i++)
     if (ot_languages[i].tag == tag)
-      return hb_language_from_string (ot_languages[i].language);
+      return hb_language_from_string (ot_languages[i].language, -1);
 
-  buf[6] = tag >> 24;
-  buf[7] = (tag >> 16) & 0xFF;
-  buf[8] = (tag >> 8) & 0xFF;
-  buf[9] = tag & 0xFF;
-  buf[10] = '\0';
-  return hb_language_from_string ((char *) buf);
+  /* If tag starts with ZH, it's Chinese */
+  if ((tag & 0xFFFF0000)  == 0x5A480000) {
+    switch (tag) {
+      case HB_TAG('Z','H','H',' '): return hb_language_from_string ("zh-hk", -1); /* Hong Kong */
+      default: {
+        /* Encode the tag... */
+        unsigned char buf[14] = "zh-x-hbot";
+        buf[9] = tag >> 24;
+        buf[10] = (tag >> 16) & 0xFF;
+        buf[11] = (tag >> 8) & 0xFF;
+        buf[12] = tag & 0xFF;
+        if (buf[12] == 0x20)
+          buf[12] = '\0';
+        buf[13] = '\0';
+        return hb_language_from_string ((char *) buf, -1);
+      }
+    }
+  }
+
+  /* Else return a custom language in the form of "x-hbotABCD" */
+  {
+    unsigned char buf[11] = "x-hbot";
+    buf[6] = tag >> 24;
+    buf[7] = (tag >> 16) & 0xFF;
+    buf[8] = (tag >> 8) & 0xFF;
+    buf[9] = tag & 0xFF;
+    if (buf[9] == 0x20)
+      buf[9] = '\0';
+    buf[10] = '\0';
+    return hb_language_from_string ((char *) buf, -1);
+  }
 }
 
 
-HB_END_DECLS