vim-patch:9.0.1598: screenchar() and others are wrong with DBCS 'encoding' (#23872)

Problem:    screenchar(), screenchars() and screenstring() do not work
            properly when 'encoding' is set to a double-byte encoding.
Solution:   Fix the way the bytes of the characters are obtained.
            (issue vim/vim#12469)

47eec6716b

src/nvim/eval/funcs.c

@@ -62,7 +62,7 @@
 #include "nvim/getchar.h"
 #include "nvim/gettext.h"
 #include "nvim/globals.h"
-#include "nvim/grid_defs.h"
+#include "nvim/grid.h"
 #include "nvim/hashtab.h"
 #include "nvim/highlight_defs.h"
 #include "nvim/highlight_group.h"
@@ -6745,7 +6745,9 @@ static void f_screenchar(typval_T *argvars, typval_T *rettv, EvalFuncData fptr)
   if (row < 0 || row >= grid->rows || col < 0 || col >= grid->cols) {
     c = -1;
   } else {
-    c = utf_ptr2char((char *)grid->chars[grid->line_offset[row] + (size_t)col]);
+    char buf[MB_MAXBYTES + 1];
+    grid_getbytes(grid, row, col, buf, NULL);
+    c = utf_ptr2char(buf);
   }
   rettv->vval.v_number = c;
 }
@@ -6763,10 +6765,13 @@ static void f_screenchars(typval_T *argvars, typval_T *rettv, EvalFuncData fptr)
     tv_list_alloc_ret(rettv, 0);
     return;
   }
+
+  char buf[MB_MAXBYTES + 1];
+  grid_getbytes(grid, row, col, buf, NULL);
   int pcc[MAX_MCO];
-  int c = utfc_ptr2char((char *)grid->chars[grid->line_offset[row] + (size_t)col], pcc);
+  int c = utfc_ptr2char(buf, pcc);
   int composing_len = 0;
-  while (pcc[composing_len] != 0) {
+  while (composing_len < MAX_MCO && pcc[composing_len] != 0) {
     composing_len++;
   }
   tv_list_alloc_ret(rettv, composing_len + 1);
@@ -6806,7 +6811,9 @@ static void f_screenstring(typval_T *argvars, typval_T *rettv, EvalFuncData fptr
     return;
   }
 
-  rettv->vval.v_string = xstrdup((char *)grid->chars[grid->line_offset[row] + (size_t)col]);
+  char buf[MB_MAXBYTES + 1];
+  grid_getbytes(grid, row, col, buf, NULL);
+  rettv->vval.v_string = xstrdup(buf);
 }
 
 /// "search()" function
@@ -7201,7 +7208,7 @@ static void f_serverstop(typval_T *argvars, typval_T *rettv, EvalFuncData fptr)
 }
 
 /// Set the cursor or mark position.
-/// If 'charpos' is true, then use the column number as a character offset.
+/// If "charpos" is true, then use the column number as a character offset.
 /// Otherwise use the column number as a byte offset.
 static void set_position(typval_T *argvars, typval_T *rettv, bool charpos)
 {

src/nvim/grid.c

@@ -138,8 +138,9 @@ void grid_putchar(ScreenGrid *grid, int c, int row, int col, int attr)
   grid_puts(grid, buf, row, col, attr);
 }
 
-/// get a single character directly from grid.chars into "bytes[]".
+/// Get a single character directly from grid.chars into "bytes", which must
-/// Also return its attribute in *attrp;
+/// have a size of "MB_MAXBYTES + 1".
+/// If "attrp" is not NULL, return the character's attribute in "*attrp".
 void grid_getbytes(ScreenGrid *grid, int row, int col, char *bytes, int *attrp)
 {
   grid_adjust(&grid, &row, &col);
@@ -150,7 +151,9 @@ void grid_getbytes(ScreenGrid *grid, int row, int col, char *bytes, int *attrp)
   }
 
   size_t off = grid->line_offset[row] + (size_t)col;
-  *attrp = grid->attrs[off];
+  if (attrp != NULL) {
+    *attrp = grid->attrs[off];
+  }
   schar_copy(bytes, grid->chars[off]);
 }
 

test/old/testdir/test_functions.vim

@@ -2834,6 +2834,34 @@ func Test_screen_functions()
   call assert_equal(-1, screenattr(-1, -1))
   call assert_equal(-1, screenchar(-1, -1))
   call assert_equal([], screenchars(-1, -1))
+
+  " Run this in a separate Vim instance to avoid messing up.
+  let after =<< trim [CODE]
+    scriptencoding utf-8
+    call setline(1, '口')
+    redraw
+    call assert_equal(0, screenattr(1, 1))
+    call assert_equal(char2nr('口'), screenchar(1, 1))
+    call assert_equal([char2nr('口')], screenchars(1, 1))
+    call assert_equal('口', screenstring(1, 1))
+    call writefile(v:errors, 'Xresult')
+    qall!
+  [CODE]
+
+  let encodings = ['utf-8', 'cp932', 'cp936', 'cp949', 'cp950']
+  if !has('win32')
+    let encodings += ['euc-jp']
+  endif
+  if has('nvim')
+    let encodings = ['utf-8']
+  endif
+  for enc in encodings
+    let msg = 'enc=' .. enc
+    if RunVim([], after, $'--clean --cmd "set encoding={enc}"')
+      call assert_equal([], readfile('Xresult'), msg)
+    endif
+    call delete('Xresult')
+  endfor
 endfunc
 
 " Test for getcurpos() and setpos()

test/old/testdir/test_utf8.vim

@@ -88,7 +88,20 @@ func Test_screenchar_utf8()
   call assert_equal("B", screenstring(1, 2))
   call assert_equal("C\u0308", screenstring(1, 3))
 
-  " 2-cells, with composing characters 
+  " 1-cell, with 6 composing characters
+  set maxcombine=6
+  call setline(1, ["ABC" .. repeat("\u0308", 6)])
+  redraw
+  call assert_equal([0x0041], screenchars(1, 1))
+  call assert_equal([0x0042], 1->screenchars(2))
+  " This should not use uninitialized memory
+  call assert_equal([0x0043] + repeat([0x0308], 6), screenchars(1, 3))
+  call assert_equal("A", screenstring(1, 1))
+  call assert_equal("B", screenstring(1, 2))
+  call assert_equal("C" .. repeat("\u0308", 6), screenstring(1, 3))
+  set maxcombine&
+
+  " 2-cells, with composing characters
   let text = "\u3042\u3044\u3046\u3099"
   call setline(1, text)
   redraw