From 936486cb1ccaba88add3b7caa984cdf841e019f1 Mon Sep 17 00:00:00 2001
From: Manolo Gouy <Manolo>
Date: Sat, 17 Dec 2016 07:42:54 +0000
Subject: [PATCH] Handle non-ASCII characters when selecting a word or moving
 the cursor by one word.

git-svn-id: file:///fltk/svn/fltk/branches/branch-1.4@12149 ea41ed52-d2ee-0310-a9c1-e6b18d33e121
---
 FL/Fl_Text_Buffer.H     |  1 +
 src/Fl_Text_Buffer.cxx  | 28 +++++++++++++++++++---------
 src/Fl_Text_Display.cxx | 17 +++++------------
 3 files changed, 25 insertions(+), 21 deletions(-)

diff --git a/FL/Fl_Text_Buffer.H b/FL/Fl_Text_Buffer.H
index 4c6cf31a6..ee02322db 100644
--- a/FL/Fl_Text_Buffer.H
+++ b/FL/Fl_Text_Buffer.H
@@ -699,6 +699,7 @@ public:
    to be informed if file input required transcoding to UTF-8.
    */
   void (*transcoding_warning_action)(Fl_Text_Buffer*);
+  bool is_word_separator(int pos) const;
 
 protected:
 
diff --git a/src/Fl_Text_Buffer.cxx b/src/Fl_Text_Buffer.cxx
index 9efcd6609..1b42234f3 100644
--- a/src/Fl_Text_Buffer.cxx
+++ b/src/Fl_Text_Buffer.cxx
@@ -825,17 +825,29 @@ int Fl_Text_Buffer::line_end(int pos) const {
 } 
 
 
+/** Returns whether character at position \p pos is a word separator.
+ Pos must be at a character boundary.
+ */
+bool Fl_Text_Buffer::is_word_separator(int pos) const {
+  int c = char_at(pos);
+  if (c < 128) {
+    return !(isalnum(c) || c == '_');  // non alphanumeric ASCII
+  }
+  return (c == 0xA0 ||                 // NO-BREAK SPACE
+          (c >= 0x3000 && c <= 0x301F) // IDEOGRAPHIC punctuation
+         );
+}
+
+
 /*
  Find the beginning of a word.
- NOT UNICODE SAFE.
  */
 int Fl_Text_Buffer::word_start(int pos) const {
-  // FIXME: character is ucs-4
-  while (pos>0 && (isalnum(char_at(pos)) || char_at(pos) == '_')) {
+  while (pos > 0 && !is_word_separator(pos))
+  {
     pos = prev_char(pos);
-  } 
-  // FIXME: character is ucs-4
-  if (!(isalnum(char_at(pos)) || char_at(pos) == '_'))
+  }
+  if (is_word_separator(pos))
     pos = next_char(pos);
   return pos;
 }
@@ -843,11 +855,9 @@ int Fl_Text_Buffer::word_start(int pos) const {
 
 /*
  Find the end of a word.
- NOT UNICODE SAFE.
  */
 int Fl_Text_Buffer::word_end(int pos) const {
-  // FIXME: character is ucs-4
-  while (pos < length() && (isalnum(char_at(pos)) || char_at(pos) == '_'))
+  while (pos < length() && !is_word_separator(pos))
   {
     pos = next_char(pos);
   }
diff --git a/src/Fl_Text_Display.cxx b/src/Fl_Text_Display.cxx
index 52c506715..aacee88e6 100644
--- a/src/Fl_Text_Display.cxx
+++ b/src/Fl_Text_Display.cxx
@@ -1537,24 +1537,17 @@ int Fl_Text_Display::rewind_lines(int startPos, int nLines) {
 
 
 
-static inline int fl_isseparator(unsigned int c) {
-  // FIXME: this does not take UCS-4 encoding into account
-  return c != '$' && c != '_' && (isspace(c) || ispunct(c));
-}
-
-
-
 /**
  \brief Moves the current insert position right one word.
  */
 void Fl_Text_Display::next_word() {
   int pos = insert_position();
 
-  while (pos < buffer()->length() && !fl_isseparator(buffer()->char_at(pos))) {
+  while (pos < buffer()->length() && !buffer()->is_word_separator(pos)) {
     pos = buffer()->next_char(pos);
   }
 
-  while (pos < buffer()->length() && fl_isseparator(buffer()->char_at(pos))) {
+  while (pos < buffer()->length() && buffer()->is_word_separator(pos)) {
     pos = buffer()->next_char(pos);
   }
 
@@ -1571,15 +1564,15 @@ void Fl_Text_Display::previous_word() {
   if (pos==0) return;
   pos = buffer()->prev_char(pos);
 
-  while (pos && fl_isseparator(buffer()->char_at(pos))) {
+  while (pos && buffer()->is_word_separator(pos)) {
     pos = buffer()->prev_char(pos);
   }
 
-  while (pos && !fl_isseparator(buffer()->char_at(pos))) {
+  while (pos && !buffer()->is_word_separator(pos)) {
     pos = buffer()->prev_char(pos);
   }
 
-  if (fl_isseparator(buffer()->char_at(pos))) {
+  if (buffer()->is_word_separator(pos)) {
     pos = buffer()->next_char(pos);
   }