Fix text-terminal output with UTF-8 encoding on MS-Windows

* src/w32console.c (w32con_write_glyphs)
(w32con_write_glyphs_with_face): Support UTF-8 encoded text
better, by counting characters and using display columns, not
bytes, to move the cursor after writing the text.  (Bug#79298)

1 files changed, 30 insertions, 8 deletions

diff --git a/src/w32console.c b/src/w32console.c
index 1bca0cadff9..30f511d93f9 100644
--- a/src/w32console.c
+++ b/src/w32console.c
@@ -351,10 +351,23 @@ w32con_write_glyphs (struct frame *f, register struct glyph *string,
       conversion_buffer = (LPCSTR) encode_terminal_code (string, n, coding);
       if (coding->produced > 0)
         {
+          /* By default, assume single-byte encoding and single-column
+             characters...  */
+          ptrdiff_t nchars = coding->produced;
+          ptrdiff_t ncols = nchars;
+          /* ...but if we are using UTF-8, correct that by computing
+             characters.  Note: multibyte_chars_in_text and strwidth
+             handle the internal encoding of characters, which is a
+             superset of UTF-8.
+             FIXME: this doesn't handle character compositions.  */
+          if (coding->encoder == encode_coding_utf_8)
+            {
+              ncols = strwidth (conversion_buffer, nchars);
+              nchars = multibyte_chars_in_text (conversion_buffer, nchars);
+            }
           /* Set the attribute for these characters.  */
-          if (!FillConsoleOutputAttribute (cur_screen, char_attr,
-                                           coding->produced, cursor_coords,
-                                           &r))
+          if (!FillConsoleOutputAttribute (cur_screen, char_attr, nchars,
+                                           cursor_coords, &r))
             {
               printf ("Failed writing console attributes: %lu\n",
                       GetLastError ());
@@ -371,7 +384,7 @@ w32con_write_glyphs (struct frame *f, register struct glyph *string,
               fflush (stdout);
             }
 
-          cursor_coords.X += coding->produced;
+          cursor_coords.X += ncols;
           w32con_move_cursor (f, cursor_coords.Y, cursor_coords.X);
         }
       len -= n;
@@ -407,19 +420,28 @@ w32con_write_glyphs_with_face (struct frame *f, register int x, register int y,
       /* Compute the character attributes corresponding to the face.  */
       DWORD char_attr = w32_face_attributes (f, face_id);
       COORD start_coords;
+      /* By default, assume single-byte encoding...  */
+      ptrdiff_t nchars = coding->produced;
+      /* ...but if we are using UTF-8, correct that by counting
+         characters.  Note: multibyte_chars_in_text handles the
+         internal encoding of characters, which is a superset of
+         UTF-8.
+         FIXME: this doesn't handle character compositions.  */
+      if (coding->encoder == encode_coding_utf_8)
+        nchars = multibyte_chars_in_text (conversion_buffer, nchars);
 
       start_coords.X = x;
       start_coords.Y = y;
       /* Set the attribute for these characters.  */
-      if (!FillConsoleOutputAttribute (cur_screen, char_attr,
-                                       coding->produced, start_coords,
-                                       &filled))
+      if (!FillConsoleOutputAttribute (cur_screen, char_attr, nchars,
+                                       start_coords, &filled))
         DebPrint (("Failed writing console attributes: %d\n", GetLastError ()));
       else
         {
           /* Write the characters.  */
           if (!WriteConsoleOutputCharacter (cur_screen, conversion_buffer,
-                                            filled, start_coords, &written))
+                                            coding->produced, start_coords,
+                                            &written))
             DebPrint (("Failed writing console characters: %d\n",
                        GetLastError ()));
         }