Changeset 84027015

Timestamp:

Jan 17, 2008, 1:23:53 AM (16 years ago)

Author:

Alejandro R. Sedeño <asedeno@mit.edu>

Branches:

master, barnowl_perlaim, debian, release-1.10, release-1.4, release-1.5, release-1.6, release-1.7, release-1.8, release-1.9

Children:

b2c1bd4

Parents:

a8d5a39

Message:

editwin.c - lots of utf-8 cleanup that I had been putting off.
util.c - a can we break here'' function based on perl's Text::WrapI18N

Files:

• editwin.c (modified) (13 diffs)
• util.c (modified) (1 diff)

editwin.c

@@ -351 +351 @@
 int _owl_editwin_linewrap_word(owl_editwin *e)
 {
-  int i, z;
-
-  z = _owl_editwin_get_index_from_xy(e);
-  /* move back and line wrap the previous word */
-  for (i = z - 1; ; i--) {
-    /* move back until you find a space or hit the beginning of the line */
-    if (e->buff[i] == ' ') {
-      /* replace the space with a newline */
-      e->buff[i] = '\n';
-      e->buffy++;
-      e->buffx = z - i - 1;
-      /* were we on the last line */
-      return(0);
-    } else if (e->buff[i] == '\n' || i <= e->lock) {
-      /* we hit the beginning of the line or the buffer, we cannot
-       * wrap.
-       */
-      return(-1);
-    }
-  }
+  int x, y;
+  int i;
+  char *ptr1, *start;
+  gunichar c;
+
+  /* saving values */
+  x = e->buffx;
+  y = e->buffy;
+  start = e->buff + e->lock;
+
+  ptr1 = e->buff + _owl_editwin_get_index_from_xy(e);
+  ptr1 = g_utf8_find_prev_char(start, ptr1);
+
+  while (ptr1) {
+    c = g_utf8_get_char(ptr1);
+    if (owl_util_can_break_after(c)) {
+      if (c != ' ') {
+        _owl_editwin_set_xy_by_index(e, ptr1 - e->buff);
+        owl_editwin_key_right(e);
+        /* _owl_editwin_insert_bytes may move e->buff. */
+        i = ptr1 - e->buff;
+        _owl_editwin_insert_bytes(e,1);
+        ptr1 = e->buff + i;
+      }
+      *ptr1 = '\n';
+      return 0;
+    }
+    ptr1 = g_utf8_find_prev_char(start, ptr1);
+  }
+  return -1;
 }
 
@@ -642 +652 @@
 }
 
+/* We assume x,y are not set to point to a mid-char */
+gunichar _owl_editwin_get_char_at_xy(owl_editwin *e)
+{
+  return g_utf8_get_char(e->buff + _owl_editwin_get_index_from_xy(e));
+}
+
+
 void _owl_editwin_set_xy_by_index(owl_editwin *e, int index)
 {
@@ -687 +704 @@
   int x, i;
 
+  /* Find line */
   ptr1 = e->buff;
   ptr2 = strchr(ptr1, '\n');
@@ -694 +712 @@
   }
   ptr2 = ptr1;
+
+  /* Find char */
   x = 0;
   while (ptr2 != NULL && x < e->buffx) {
@@ -701 +721 @@
     ptr2 = g_utf8_next_char(ptr2);
   }
+  
+  /* calculate x offset */
   return x - e->buffx;
 }
@@ -805 +827 @@
 void owl_editwin_move_to_nextword(owl_editwin *e)
 {
-  /* asedeno: needs fixing for utf-8*/
   int i, x;
+  gunichar c = '\0';
 
   /* if we're starting on a space, find the first non-space */
@@ -819 +841 @@
   }
 
-  /* find the next space, newline or end of line and go there, if
-     already at the end of the line, continue on to the next */
-  i=owl_editwin_get_numchars_on_line(e, e->buffy);
+  /* find the next space, newline or end of line and go
+     there, if already at the end of the line, continue on to the next */
+  i=owl_editwin_get_numcells_on_line(e, e->buffy);
+  c = _owl_editwin_get_char_at_xy(e);
   if (e->buffx < i) {
     /* move right till end of line */
     while (e->buffx < i) {
-      e->buffx++;
-      if (e->buff[_owl_editwin_get_index_from_xy(e)]==' ') return;
+      owl_editwin_key_right(e);
+      c = _owl_editwin_get_char_at_xy(e);
+      if (c == ' ') return;
       if (e->buffx == i) return;
     }
@@ -832 +856 @@
     /* try to move down */
     if (e->style==OWL_EDITWIN_STYLE_MULTILINE) {
-      if (e->buffy+1 <  owl_editwin_get_numlines(e)) {
+      if (e->buffy+1 < owl_editwin_get_numlines(e)) {
         e->buffx=0;
         e->buffy++;
@@ -845 +869 @@
 void owl_editwin_move_to_previousword(owl_editwin *e)
 {
-  /* asedeno: needs fixing for utf-8*/
-  int i, x;
+  int i;
+  gunichar c;
+  char *ptr1, *ptr2;
 
   /* are we already at the beginning of the word? */
-  i=_owl_editwin_get_index_from_xy(e);
-  if ( (e->buff[i]!=' ' && e->buff[i]!='\n' && e->buff[i]!='\0') &&
-       (e->buff[i-1]==' ' || e->buff[i-1]=='\n') ) {
+  c = _owl_editwin_get_char_at_xy(e);
+  i = _owl_editwin_get_index_from_xy(e);
+  ptr1 = e->buff + i;
+  if (*ptr1 != ' ' && *ptr1 != '\n' && *ptr1 != '\0' ) {
+    ptr1 = g_utf8_find_prev_char(e->buff, ptr1);
+    c = g_utf8_get_char(ptr1);
+    if (c == ' ' || c == '\n') {
+      owl_editwin_key_left(e);      
+    }
+  }
+
+  /* are we starting on a space character? */
+  i = _owl_editwin_get_index_from_xy(e);
+  while (e->buff[i] == ' ' || e->buff[i] == '\n' || e->buff[i] == '\0') {
+    /* find the first non-space */
+    owl_editwin_key_left(e);      
+    i = _owl_editwin_get_index_from_xy(e);
+  }
+
+  /* find the last non-space */
+  owl_editwin_key_left(e);
+  ptr1 = e->buff + _owl_editwin_get_index_from_xy(e);
+  while (ptr1 >= e->buff + e->lock) {
+    ptr2 = g_utf8_find_prev_char(e->buff, ptr1);
+    if (!ptr2) break;
+    
+    c = g_utf8_get_char(ptr2);
+    if (c == ' ' || c == '\n'){
+      break;
+    }
     owl_editwin_key_left(e);
-  }
-    
-  /* are we starting on a space character? */
-  i=_owl_editwin_get_index_from_xy(e);
-  if (e->buff[i]==' ' || e->buff[i]=='\n' || e->buff[i]=='\0') {
-    /* find the first non-space */
-    for (x=i; x>=e->lock; x--) {
-      if (e->buff[x]!=' ' && e->buff[x]!='\n' && e->buff[x]!='\0') {
-        _owl_editwin_set_xy_by_index(e, x);
-        break;
-      }
-    }
-  }
-
-  /* find the last non-space */
-  i=_owl_editwin_get_index_from_xy(e);
-  for (x=i; x>=e->lock; x--) {
-    if (e->buff[x-1]==' ' || e->buff[x-1]=='\n') {
-      _owl_editwin_set_xy_by_index(e, x);
-      break;
-    }
-  }
-  _owl_editwin_set_xy_by_index(e, x);
+    ptr1 = e->buff + _owl_editwin_get_index_from_xy(e);
+  }
 }
 
@@ -881 +912 @@
 void owl_editwin_delete_nextword(owl_editwin *e)
 {
-  /* asedeno: needs fixing for utf-8*/
-  int z;
+  char *ptr1, *start;
+  gunichar c;
 
   if (e->bufflen==0) return;
 
-  /* if we start out on a space character then gobble all the spaces
-     up first */
-  while (1) {
-    z=_owl_editwin_get_index_from_xy(e);
-    if (e->buff[z]==' ' || e->buff[z]=='\n') {
-      owl_editwin_delete_char(e);
-    } else {
-      break;
-    }
-  }
-
-  /* then nuke the next word */
-  while (1) {
-    z=_owl_editwin_get_index_from_xy(e);
-    /* z == e->bufflen check added to prevent a hang I (nelhage) have
-       seen repeatedly while using owl. I'm not sure precisely what
-       conditions lead to it. */
-    if (z == e->bufflen
-        || e->buff[z+1]==' ' || e->buff[z+1]=='\n' || e->buff[z+1]=='\0') break;
-    owl_editwin_delete_char(e);
-  }
-  owl_editwin_delete_char(e);
+  start = ptr1 = e->buff + _owl_editwin_get_index_from_xy(e);
+  /* if we start out on a space character then jump past all the
+     spaces up first */
+  while (*ptr1 == ' ' || *ptr1 == '\n') {
+    ++ptr1;
+  }
+
+  /* then jump past the next word */
+  
+  while (ptr1 && ptr1 - e->buff < e->bufflen) {
+    c = g_utf8_get_char(ptr1);
+    if (c == ' ' || c == '\n' || c == '\0') break;
+    ptr1 = g_utf8_find_next_char(ptr1, NULL);
+  }
+
+  if (ptr1) { /* We broke on a space, */
+    ptr1 = g_utf8_find_next_char(ptr1, NULL);
+    if (ptr1) { /* and there's a character after it, */
+      /* nuke everything back to our starting point. */
+      _owl_editwin_remove_bytes(e, ptr1 - start);
+      return;
+    }
+  }
+  
+  /* If we get here, we ran out of string, drop what's left. */
+  *start = '\0';
+  e->bufflen = start - e->buff;
 }
 
 void owl_editwin_delete_previousword(owl_editwin *e)
 {
-  /* asedeno: needs fixing for utf-8*/
   /* go backwards to the last non-space character, then delete chars */
-  int i, startpos, endpos;
+  int startpos, endpos;
 
   startpos = _owl_editwin_get_index_from_xy(e);
   owl_editwin_move_to_previousword(e);
   endpos = _owl_editwin_get_index_from_xy(e);
-  for (i=0; i<startpos-endpos; i++) {
-    owl_editwin_delete_char(e);
-  }
+  _owl_editwin_remove_bytes(e, startpos-endpos);
 }
 
 void owl_editwin_delete_to_endofline(owl_editwin *e)
 {
-  /* asedeno: needs fixing for utf-8*/
   int i;
 
@@ -987 +1019 @@
 void owl_editwin_fill_paragraph(owl_editwin *e)
 {
-  /* asedeno: needs fixing for utf-8*/
   int i, save;
 
@@ -1018 +1049 @@
 
     /* did we hit the end of a line too soon? */
+    /* asedeno: Here we replace a newline with a space. We may want to
+       consider removing the space if the characters to either side
+       are CJK ideograms.*/
     i = _owl_editwin_get_index_from_xy(e);
     if (e->buff[i] == '\n' && e->buffx < e->fillcol - 1) {
@@ -1031 +1065 @@
       } else {
         owl_editwin_delete_char(e);
-        /* if we did this ahead of the save point, adjust it */
+        /* if we did this ahead of the save point, adjust it. Changing
+           by one is fine here because we're only removing an ASCII
+           space. */
         if (i < save) save--;
       }

util.c

@@ -832 +832 @@
 }
 
+/* This is based on _extract() and _isCJ() from perl's Text::WrapI18N */
+int owl_util_can_break_after(gunichar c)
+{
+  
+  if (c == ' ') return 1;
+  if (c >= 0x3000 && c <= 0x312f) {
+    /* CJK punctuations, Hiragana, Katakana, Bopomofo */
+    if (c == 0x300a || c == 0x300c || c == 0x300e ||
+        c == 0x3010 || c == 0x3014 || c == 0x3016 ||
+        c == 0x3018 || c == 0x301a)
+      return 0;
+    return 1;
+  }
+  if (c >= 0x31a0 && c <= 0x31bf) {return 1;}  /* Bopomofo */
+  if (c >= 0x31f0 && c <= 0x31ff) {return 1;}  /* Katakana extension */
+  if (c >= 0x3400 && c <= 0x9fff) {return 1;}  /* Han Ideogram */
+  if (c >= 0xf900 && c <= 0xfaff) {return 1;}  /* Han Ideogram */
+  if (c >= 0x20000 && c <= 0x2ffff) {return 1;}  /* Han Ideogram */
+  return 0;
+}
+
 /**************************************************************************/
 /************************* REGRESSION TESTS *******************************/