(mfont_list): If FONT is null, use a temporary font.

[m17n/m17n-lib.git] / src / mtext.c

diff --git a/src/mtext.c b/src/mtext.c
index 22d6cee..b13ef88 100644 (file)
--- a/src/mtext.c
+++ b/src/mtext.c
@@ -1,5 +1,5 @@
 /* mtext.c -- M-text module.
-   Copyright (C) 2003, 2004
+   Copyright (C) 2003, 2004, 2005
      National Institute of Advanced Industrial Science and Technology (AIST)
      Registration Number H15PRO112
 
@@ -93,6 +93,9 @@
 #include "character.h"
 #include "mtext.h"
 #include "plist.h"
+#ifdef HAVE_THAI_WORDSEG
+#include "word-thai.h"
+#endif
 
 static M17NObjectArray mtext_table;
 
@@ -682,11 +685,18 @@ case_compare (MText *mt1, int from1, int to1, MText *mt2, int from2, int to2)
 \f
 /* Internal API */
 
+MCharTable *wordseg_func_table;
+
 int
 mtext__init ()
 {
+  M17N_OBJECT_ADD_ARRAY (mtext_table, "M-text");
   M_charbag = msymbol_as_managing_key ("  charbag");
   mtext_table.count = 0;
+  wordseg_func_table = mchartable (Mnil, NULL);
+#ifdef HAVE_THAI_WORDSEG
+  mtext__word_thai_init ();
+#endif
   return 0;
 }
 
@@ -694,7 +704,11 @@ mtext__init ()
 void
 mtext__fini (void)
 {
-  mdebug__report_object ("M-text", &mtext_table);
+#ifdef HAVE_THAI_WORDSEG
+  mtext__word_thai_fini ();
+#endif
+  M17N_OBJECT_UNREF (wordseg_func_table);
+  wordseg_func_table = NULL;
 }
 
 
@@ -1103,6 +1117,22 @@ mtext__eol (MText *mt, int pos)
     }
 }
 
+typedef int (*MTextWordsegFunc) (MText *mt, int pos, int *from, int *to);
+
+int
+mtext__word_segment (MText *mt, int pos, int *from, int *to)
+{
+  int c = mtext_ref_char (mt, pos);
+  MTextWordsegFunc func
+    = (MTextWordsegFunc) mchartable_lookup (wordseg_func_table, c);
+
+  if (func)
+    return (func) (mt, pos, from, to);
+  *from = *to = pos;
+  return -1;
+}
+
+
 /*** @} */
 #endif /* !FOR_DOXYGEN || DOXYGEN_INTERNAL_MODULE */
 
@@ -1110,9 +1140,9 @@ mtext__eol (MText *mt, int pos)
 /* External API */
 
 #ifdef WORDS_BIGENDIAN
-const int MTEXT_FORMAT_UTF_16 = MTEXT_FORMAT_UTF_16BE;
+const enum MTextFormat MTEXT_FORMAT_UTF_16 = MTEXT_FORMAT_UTF_16BE;
 #else
-const int MTEXT_FORMAT_UTF_16 = MTEXT_FORMAT_UTF_16LE;
+const enum MTextFormat MTEXT_FORMAT_UTF_16 = MTEXT_FORMAT_UTF_16LE;
 #endif
 
 #ifdef WORDS_BIGENDIAN
@@ -1232,6 +1262,86 @@ mtext_from_data (const void *data, int nitems, enum MTextFormat format)
 /*=*/
 
 /***en
+    @brief Get information about the text data in M-text.
+
+    The mtext_data () function returns a pointer to the text data of
+    M-text $MT.  If $FMT is not NULL, the format of the text data is
+    stored in it.  If $NUNITS is not NULL, the number of units of the
+    text data is stored in it.
+
+    If $POS_IDX is not NULL and it points to a non-negative number,
+    what it points to is a character position.  In this case, the
+    return value is a pointer to the text data of a character at that
+    position.
+
+    Otherwise, if $UNIT_IDX is not NULL, it points to a unit position.
+    In this case, the return value is a pointer to the text data of a
+    character containing that unit.
+
+    The character position and unit position of the return value are
+    stored in $POS_IDX and $UNIT_DIX respectively if they are not
+    NULL.
+
+    <ul>
+
+    <li> If the format of the text data is MTEXT_FORMAT_US_ASCII or
+    MTEXT_FORMAT_UTF_8, one unit is unsigned char.
+
+    <li> If the format is MTEXT_FORMAT_UTF_16LE or
+    MTEXT_FORMAT_UTF_16BE, one unit is unsigned short.
+
+    <li> If the format is MTEXT_FORMAT_UTF_32LE or
+    MTEXT_FORMAT_UTF_32BE, one unit is unsigned int.
+
+    </ul> */
+
+void *
+mtext_data (MText *mt, enum MTextFormat *fmt, int *nunits,
+           int *pos_idx, int *unit_idx)
+{
+  void *data;
+  int pos = 0, unit_pos = 0;
+
+  if (fmt)
+    *fmt = mt->format;
+  data = MTEXT_DATA (mt);
+  if (pos_idx && *pos_idx >= 0)
+    {
+      pos = *pos_idx;
+      if (pos > mtext_nchars (mt))
+       MERROR (MERROR_MTEXT, NULL);
+      unit_pos = POS_CHAR_TO_BYTE (mt, pos);
+    }
+  else if (unit_idx)
+    {
+      unit_pos = *unit_idx;
+
+      if (unit_pos < 0 || unit_pos > mtext_nbytes (mt))
+       MERROR (MERROR_MTEXT, NULL);
+      pos = POS_BYTE_TO_CHAR (mt, unit_pos);
+      unit_pos = POS_CHAR_TO_BYTE (mt, pos);
+    }
+  if (nunits)
+    *nunits = mtext_nbytes (mt) - unit_pos;
+  if (pos_idx)
+    *pos_idx = pos;
+  if (unit_idx)
+    *unit_idx = unit_pos;
+  if (unit_pos > 0)
+    {
+      if (mt->format <= MTEXT_FORMAT_UTF_8)
+       data = (unsigned char *) data + unit_pos;
+      else if (mt->format <= MTEXT_FORMAT_UTF_16BE)
+       data = (unsigned short *) data + unit_pos;
+      else
+       data = (unsigned int *) data + unit_pos;
+    }
+  return data;
+}
+
+/*=*/
+
+/***en
     @brief Number of characters in M-text.
 
     The mtext_len () function returns the number of characters in
@@ -2477,9 +2587,7 @@ int
 mtext_text (MText *mt1, int pos, MText *mt2)
 {
   int from = pos;
-  int pos_byte = POS_CHAR_TO_BYTE (mt1, pos);
   int c = mtext_ref_char (mt2, 0);
-  int nbytes1 = mtext_nbytes (mt1);
   int nbytes2 = mtext_nbytes (mt2);
   int limit;
   int use_memcmp = (mt1->format == mt2->format
@@ -2487,13 +2595,14 @@ mtext_text (MText *mt1, int pos, MText *mt2)
                        && mt2->format == MTEXT_FORMAT_UTF_8));
   int unit_bytes = UNIT_BYTES (mt1->format);
 
-  if (nbytes2 > pos_byte + nbytes1)
+  if (from + mtext_nchars (mt2) > mtext_nchars (mt1))
     return -1;
-  pos_byte = nbytes1 - nbytes2;
-  limit = POS_BYTE_TO_CHAR (mt1, pos_byte);
+  limit = mtext_nchars (mt1) - mtext_nchars (mt2) + 1;
 
   while (1)
     {
+      int pos_byte;
+
       if ((pos = mtext_character (mt1, from, limit, c)) < 0)
        return -1;
       pos_byte = POS_CHAR_TO_BYTE (mt1, pos);
@@ -2568,7 +2677,7 @@ mtext_search (MText *mt1, int from, int to, MText *mt2)
        return -1;
       while (1)
        {
-         if ((from = find_char_backward (mt1, from, to, c)) < 0)
+         if ((from = find_char_backward (mt1, to, from + 1, c)) < 0)
            return -1;
          from_byte = POS_CHAR_TO_BYTE (mt1, from);
          if (! memcmp (mt1->data + from_byte, mt2->data, nbytes2))