Issue #12567: The curses module uses Unicode functions for Unicode arguments

when it is linked to the ncurses library. It encodes also Unicode strings to
the locale encoding instead of UTF-8.

Doc/library/curses.rst

@@ -653,7 +653,7 @@ Window Objects
 --------------
 
 Window objects, as returned by :func:`initscr` and :func:`newwin` above, have
-the following methods:
+the following methods and attributes:
 
 
 .. method:: window.addch([y, x,] ch[, attr])
@@ -834,6 +834,16 @@ the following methods:
    event.
 
 
+.. attribute:: window.encoding
+
+   Encoding used to encode method arguments (Unicode strings and characters).
+   The encoding attribute is inherited from by parent window when a subwindow
+   is created, for example with :meth:`window.subwin`. By default, the locale
+   encoding is used (see :func:`locale.getpreferredencoding`).
+
+   .. versionadded:: 3.3
+
+
 .. method:: window.erase()
 
    Clear the window.

Doc/whatsnew/3.3.rst

@@ -333,6 +333,11 @@ function to the :mod:`crypt` module.
 curses
 ------
 
+ * If the :mod:`curses` module is linked to the ncursesw library, use Unicode
+   functions when Unicode strings or characters are passed (e.g.
+   :c:func:`waddwstr`), and bytes functions otherwise (e.g. :c:func:`waddstr`).
+ * Use the locale encoding instead of ``utf-8`` to encode Unicode strings.
+ * :class:`curses.window` has a new :attr:`curses.window.encoding` attribute.
  * The :class:`curses.window` class has a new :meth:`~curses.window.get_wch`
    method to get a wide character
  * The :mod:`curses` module has a new :meth:`~curses.unget_wch` function to

Include/py_curses.h

@@ -76,6 +76,7 @@ extern "C" {
 typedef struct {
 	PyObject_HEAD
 	WINDOW *win;
+	char *encoding;
 } PyCursesWindowObject;
 
 #define PyCursesWindow_Check(v)	 (Py_TYPE(v) == &PyCursesWindow_Type)

Lib/test/test_curses.py

@@ -267,24 +267,42 @@ def test_issue6243(stdscr):
 def test_unget_wch(stdscr):
     if not hasattr(curses, 'unget_wch'):
         return
-    ch = 'a'
-    curses.unget_wch(ch)
-    read = stdscr.get_wch()
-    read = chr(read)
-    if read != ch:
-        raise AssertionError("%r != %r" % (read, ch))
-
-    ch = ord('a')
-    curses.unget_wch(ch)
-    read = stdscr.get_wch()
-    if read != ch:
-        raise AssertionError("%r != %r" % (read, ch))
+    for ch in ('a', '\xe9', '\u20ac', '\U0010FFFF'):
+        curses.unget_wch(ch)
+        read = stdscr.get_wch()
+        read = chr(read)
+        if read != ch:
+            raise AssertionError("%r != %r" % (read, ch))
+
+        code = ord(ch)
+        curses.unget_wch(code)
+        read = stdscr.get_wch()
+        if read != code:
+            raise AssertionError("%r != %r" % (read, code))
 
 def test_issue10570():
     b = curses.tparm(curses.tigetstr("cup"), 5, 3)
     assert type(b) is bytes
     curses.putp(b)
 
+def test_encoding(stdscr):
+    import codecs
+    encoding = stdscr.encoding
+    codecs.lookup(encoding)
+    try:
+        stdscr.encoding = 10
+    except TypeError:
+        pass
+    else:
+        raise AssertionError("TypeError not raised")
+    stdscr.encoding = encoding
+    try:
+        del stdscr.encoding
+    except TypeError:
+        pass
+    else:
+        raise AssertionError("TypeError not raised")
+
 def main(stdscr):
     curses.savetty()
     try:
@@ -295,6 +313,7 @@ def main(stdscr):
         test_issue6243(stdscr)
         test_unget_wch(stdscr)
         test_issue10570()
+        test_encoding(stdscr)
     finally:
         curses.resetty()
 

Misc/NEWS

@@ -395,6 +395,10 @@ Core and Builtins
 Library
 -------
 
+- Issue #12567: The curses module uses Unicode functions for Unicode arguments
+  when it is linked to the ncurses library. It encodes also Unicode strings to
+  the locale encoding instead of UTF-8.
+
 - Issue #12856: Ensure child processes do not inherit the parent's random
   seed for filename generation in the tempfile module.  Patch by Brian
   Harring.