pango-script.c

GTK+-2.0源码之pango-1.15.6.tar.gz

      /*
       * Paired character handling:
       *
       * if it's an open character, push it onto the stack.
       * if it's a close character, find the matching open on the
       * stack, and use that script code. Any non-matching open
       * characters above it on the stack will be poped.
       */
      if (pair_index >= 0)
	{
	  if (IS_OPEN (pair_index))
	    {
	      /*
	       * If the paren stack is full, empty it. This
	       * means that deeply nested paired punctuation
	       * characters will be ignored, but that's an unusual
	       * case, and it's better to ignore them than to
	       * write off the end of the stack...
	       */
	      if (++iter->paren_sp >= PAREN_STACK_DEPTH)
		iter->paren_sp = 0;

	      iter->paren_stack[iter->paren_sp].pair_index = pair_index;
	      iter->paren_stack[iter->paren_sp].script_code = iter->script_code;
	    }
	  else if (iter->paren_sp >= 0)
	    {
	      int pi = pair_index & ~1;

	      while (iter->paren_sp >= 0 && iter->paren_stack[iter->paren_sp].pair_index != pi)
		iter->paren_sp--;

	      if (iter->paren_sp < start_sp)
		start_sp = iter->paren_sp;

	      if (iter->paren_sp >= 0)
		sc = iter->paren_stack[iter->paren_sp].script_code;
	    }
	}

      if (SAME_SCRIPT (iter->script_code, sc))
	{
	  if (!REAL_SCRIPT (iter->script_code) && REAL_SCRIPT (sc))
	    {
	      iter->script_code = sc;

	      /*
	       * now that we have a final script code, fix any open
	       * characters we pushed before we knew the script code.
	       */
	      while (start_sp < iter->paren_sp)
		iter->paren_stack[++start_sp].script_code = iter->script_code;
	    }

	  /*
	   * if this character is a close paired character,
	   * pop it from the stack
	   */
	  if (pair_index >= 0 && !IS_OPEN (pair_index) && iter->paren_sp >= 0)
	    {
	      iter->paren_sp--;

	      if (iter->paren_sp < start_sp)
		start_sp = iter->paren_sp;
	    }
	}
      else
	{
	  /* Different script, we're done */
	  break;
	}
    }

  return TRUE;
}

/**********************************************************
 * End of code from ICU
 **********************************************************/

#include "pango-script-lang-table.h"

/* The fact that this comparison function works is dependent
 * on a property of the pango_script_lang_table which accidental rather
 * than inherent.
 *
 * The property is if we take any element in the table and suffix it
 * <elem>-<suffix> then that must strcmp() between any elements
 * preceding the element in the table and any element following in the
 * table. So, if we had something like:
 *
 * 'zh'
 *' zh-cn'
 *
 * in the table we would have a problem since 'zh-tw' follows 'zh-cn'.
 * On the other hand:
 *
 * 'zh'
 * 'zha'
 *
 * Works because 'zh-tw' precedes 'zha'.
 */
static int
script_for_lang_compare (gconstpointer key,
			 gconstpointer member)
{
  PangoLanguage *lang = (PangoLanguage *)key;
  const PangoScriptForLang *script_for_lang = member;

  if (pango_language_matches (lang, script_for_lang->lang))
    return 0;
  else
    return strcmp (pango_language_to_string (lang),
		   script_for_lang->lang);
}

/**
 * pango_language_includes_script:
 * @language: a #PangoLanguage
 * @script: a #PangoScript
 *
 * Determines if @script is one of the scripts used to
 * write @language. The returned value is conservative;
 * if nothing is known about the language tag @language,
 * %TRUE will be returned, since, as far as Pango knows,
 * @script might be used to write @language.
 *
 * This routine is used in Pango's itemization process when
 * determining if a supplied language tag is relevant to
 * a particular section of text. It probably is not useful for
 * applications in most circumstances.
 *
 * Return value: %TRUE if @script is one of the scripts used
 * to write @language, or if nothing is known about @language.
 *
 * Since: 1.4
 **/
gboolean
pango_language_includes_script (PangoLanguage *language,
				PangoScript    script)
{
  PangoScriptForLang *script_for_lang;
  unsigned int j;

  g_return_val_if_fail (language != NULL, FALSE);

  if (!REAL_SCRIPT (script))
    return TRUE;

  /* This bsearch could be optimized to occur only once if
   * we store the pointer to the PangoScriptForLang in the
   * same block as the string value for the PangoLanguage.
   */
  script_for_lang = bsearch (pango_language_to_string (language),
			     pango_script_for_lang,
			     G_N_ELEMENTS (pango_script_for_lang),
			     sizeof (PangoScriptForLang),
			     script_for_lang_compare);
  if (!script_for_lang)
    return TRUE;

  for (j = 0; j < G_N_ELEMENTS (script_for_lang->scripts); j++)
    if (script_for_lang->scripts[j] == script)
      return TRUE;

  return FALSE;
}

/**
 * pango_script_get_sample_language:
 * @script: a #PangoScript
 *
 * Given a script, finds a language tag that is reasonably
 * representative of that script. This will usually be the
 * most widely spoken or used language written in that script:
 * for instance, the sample language for %PANGO_SCRIPT_CYRILLIC
 * is <literal>ru</literal> (Russian), the sample language
 * for %PANGO_SCRIPT_ARABIC is <literal>ar</literal>.
 *
 * For some
 * scripts, no sample language will be returned because there
 * is no language that is sufficiently representative. The best
 * example of this is %PANGO_SCRIPT_HAN, where various different
 * variants of written Chinese, Japanese, and Korean all use
 * significantly different sets of Han characters and forms
 * of shared characters. No sample language can be provided
 * for many historical scripts as well.
 *
 * Return value: a #PangoLanguage that is representative
 * of the script, or %NULL if no such language exists.
 *
 * Since: 1.4
 **/
PangoLanguage *
pango_script_get_sample_language (PangoScript script)
{
  /* Note that in the following, we want
   * pango_language_includes_script() for the sample language
   * to include the script, so alternate orthographies
   * (Shavian for English, Osmanya for Somali, etc), typically
   * have no sample language
   */
  const char sample_languages[][4] = {
    "",    /* PANGO_SCRIPT_COMMON */
    "",    /* PANGO_SCRIPT_INHERITED */
    "ar",  /* PANGO_SCRIPT_ARABIC */
    "hy",  /* PANGO_SCRIPT_ARMENIAN */
    "bn",  /* PANGO_SCRIPT_BENGALI */
    /* Used primarily in Taiwan, but not part of the standard
     * zh-tw orthography  */
    "",    /* PANGO_SCRIPT_BOPOMOFO */
    "chr", /* PANGO_SCRIPT_CHEROKEE */
    "cop", /* PANGO_SCRIPT_COPTIC */
    "ru",  /* PANGO_SCRIPT_CYRILLIC */
    /* Deseret was used to write English */
    "",    /* PANGO_SCRIPT_DESERET */
    "hi",  /* PANGO_SCRIPT_DEVANAGARI */
    "am",  /* PANGO_SCRIPT_ETHIOPIC */
    "ka",  /* PANGO_SCRIPT_GEORGIAN */
    "",    /* PANGO_SCRIPT_GOTHIC */
    "el",  /* PANGO_SCRIPT_GREEK */
    "gu",  /* PANGO_SCRIPT_GUJARATI */
    "pa",  /* PANGO_SCRIPT_GURMUKHI */
    "",    /* PANGO_SCRIPT_HAN */
    "ko",  /* PANGO_SCRIPT_HANGUL */
    "he",  /* PANGO_SCRIPT_HEBREW */
    "ja",  /* PANGO_SCRIPT_HIRAGANA */
    "kn",  /* PANGO_SCRIPT_KANNADA */
    "ja",  /* PANGO_SCRIPT_KATAKANA */
    "km",  /* PANGO_SCRIPT_KHMER */
    "lo",  /* PANGO_SCRIPT_LAO */
    "en",  /* PANGO_SCRIPT_LATIN */
    "ml",  /* PANGO_SCRIPT_MALAYALAM */
    "mn",  /* PANGO_SCRIPT_MONGOLIAN */
    "my",  /* PANGO_SCRIPT_MYANMAR */
    /* Ogham was used to write old Irish */
    "",    /* PANGO_SCRIPT_OGHAM */
    "",    /* PANGO_SCRIPT_OLD_ITALIC */
    "or",  /* PANGO_SCRIPT_ORIYA */
    "",    /* PANGO_SCRIPT_RUNIC */
    "si",  /* PANGO_SCRIPT_SINHALA */
    "syr", /* PANGO_SCRIPT_SYRIAC */
    "ta",  /* PANGO_SCRIPT_TAMIL */
    "te",  /* PANGO_SCRIPT_TELUGU */
    "dv",  /* PANGO_SCRIPT_THAANA */
    "th",  /* PANGO_SCRIPT_THAI */
    "bo",  /* PANGO_SCRIPT_TIBETAN */
    "iu",  /* PANGO_SCRIPT_CANADIAN_ABORIGINAL */
    "",    /* PANGO_SCRIPT_YI */
    "tl",  /* PANGO_SCRIPT_TAGALOG */
    /* There are no ISO-636 language codes for the following
     * Phillipino languages/scripts */
    "",    /* PANGO_SCRIPT_HANUNOO */
    "",    /* PANGO_SCRIPT_BUHID */
    "",    /* PANGO_SCRIPT_TAGBANWA */

    "",    /* PANGO_SCRIPT_BRAILLE */
    "",    /* PANGO_SCRIPT_CYPRIOT */
    "",    /* PANGO_SCRIPT_LIMBU */
    /* Used for Somali (so) in the past */
    "",    /* PANGO_SCRIPT_OSMANYA */
    /* The Shavian alphabet was designed for English */
    "",    /* PANGO_SCRIPT_SHAVIAN */
    "",    /* PANGO_SCRIPT_LINEAR_B */
    "",    /* PANGO_SCRIPT_TAI_LE */
    "uga", /* PANGO_SCRIPT_UGARITIC */

    "",    /* PANGO_SCRIPT_NEW_TAI_LUE */
    "bug", /* PANGO_SCRIPT_BUGINESE */
    /* The original script for Old Church Slavonic (chu), later
     * written with Cyrillic */
    "",    /* PANGO_SCRIPT_GLAGOLITIC */
    /* Used for for Berber (ber), but Arabic script is more common */
    "",    /* PANGO_SCRIPT_TIFINAGH */
    /* Syloti Nagri is used for Sylheti, no ISO 639 code */
    "",    /* PANGO_SCRIPT_SYLOTI_NAGRI */
    "peo", /* PANGO_SCRIPT_OLD_PERSIAN */
    "",    /* PANGO_SCRIPT_KHAROSHTHI */

    "",    /* PANGO_SCRIPT_UNKNOWN */
    "",    /* PANGO_SCRIPT_BALINESE */
    "",    /* PANGO_SCRIPT_CUNEIFORM */
    "",    /* PANGO_SCRIPT_PHOENICIAN */
    "",    /* PANGO_SCRIPT_PHAGS_PA */
    "nqo"  /* PANGO_SCRIPT_NKO */
  };
  const char *sample_language;

  g_return_val_if_fail (script >= 0, NULL);
  g_return_val_if_fail ((guint)script < G_N_ELEMENTS (sample_languages), NULL);

  sample_language = sample_languages[script];

  if (!sample_language[0])
    return NULL;
  else
    return pango_language_from_string (sample_language);
}