vim-patch:9.1.0227: Recording may still be wrong in Select mode (#28092)

Problem:  Recording may still be wrong in Select mode (after 8.2.3993).
Solution: Make sure a character isn't split between two buffer blocks.
          (zeertzjq)

closes: vim/vim#14326

https://github.com/vim/vim/commit/ea95f1a5ad2455c7fd1eee2413ac7a3460ef4f8a

3 files changed, 93 insertions, 60 deletions

diff --git a/src/nvim/getchar.c b/src/nvim/getchar.c
index f68bd7098b..2a5b8b00d4 100644
--- a/src/nvim/getchar.c
+++ b/src/nvim/getchar.c
@@ -1102,27 +1102,59 @@ void del_typebuf(int len, int offset)
 }
 
 /// Write typed characters to script file.
-/// If recording is on put the character in the recordbuffer.
+/// If recording is on put the character in the record buffer.
 static void gotchars(const uint8_t *chars, size_t len)
   FUNC_ATTR_NONNULL_ALL
 {
   const uint8_t *s = chars;
-  static uint8_t buf[4] = { 0 };
+  int c = NUL;
+  static int prev_c = NUL;
+  static uint8_t buf[MB_MAXBYTES * 3 + 4] = { 0 };
   static size_t buflen = 0;
+  static unsigned pending = 0;
+  static bool in_special = false;
+  static bool in_mbyte = false;
   size_t todo = len;
 
-  while (todo--) {
-    buf[buflen++] = *s++;
+  for (; todo--; prev_c = c) {
+    c = buf[buflen++] = *s++;
+    if (pending > 0) {
+      pending--;
+    }
 
-    // When receiving a special key sequence, store it until we have all
-    // the bytes and we can decide what to do with it.
-    if (buflen == 1 && buf[0] == K_SPECIAL) {
-      continue;
+    if ((pending == 0 || in_mbyte) && c == K_SPECIAL) {
+      pending += 2;
+      if (!in_mbyte) {
+        in_special = true;
+      }
     }
-    if (buflen == 2) {
+
+    if (pending > 0) {
       continue;
     }
 
+    if (!in_mbyte) {
+      if (in_special) {
+        in_special = false;
+        if (prev_c == KS_MODIFIER) {
+          // When receiving a modifier, wait for the modified key.
+          continue;
+        }
+        c = TO_SPECIAL(prev_c, c);
+      }
+      // When receiving a multibyte character, store it until we have all
+      // the bytes, so that it won't be split between two buffer blocks,
+      // and delete_buff_tail() will work properly.
+      pending = MB_BYTE2LEN_CHECK(c) - 1;
+      if (pending > 0) {
+        in_mbyte = true;
+        continue;
+      }
+    } else {
+      // Stored all bytes of a multibyte character.
+      in_mbyte = false;
+    }
+
     // Handle one byte at a time; no translation to be done.
     for (size_t i = 0; i < buflen; i++) {
       updatescript(buf[i]);
diff --git a/test/old/testdir/test_registers.vim b/test/old/testdir/test_registers.vim
index e113bd9e75..1a25844ad3 100644
--- a/test/old/testdir/test_registers.vim
+++ b/test/old/testdir/test_registers.vim
@@ -214,6 +214,58 @@ func Test_recording_with_select_mode()
   bwipe!
 endfunc
 
+func Run_test_recording_with_select_mode_utf8()
+  new
+
+  " Test with different text lengths: 5, 7, 9, 11, 13, 15, to check that
+  " a character isn't split between two buffer blocks.
+  for s in ['12345', '口=口', '口口口', '口=口=口', '口口=口口', '口口口口口']
+    " 0x80 is K_SPECIAL
+    " 0x9B is CSI
+    " 哦: 0xE5 0x93 0xA6
+    " 洛: 0xE6 0xB4 0x9B
+    " 固: 0xE5 0x9B 0xBA
+    " 四: 0xE5 0x9B 0x9B
+    " 最: 0xE6 0x9C 0x80
+    " 倒: 0xE5 0x80 0x92
+    " 倀: 0xE5 0x80 0x80
+    for c in ['哦', '洛', '固', '四', '最', '倒', '倀']
+      call setline(1, 'asdf')
+      call feedkeys($"qacc{s}\<Esc>gH{c}\<Esc>q", "tx")
+      call assert_equal(c, getline(1))
+      call assert_equal($"cc{s}\<Esc>gH{c}\<Esc>", @a)
+      call setline(1, 'asdf')
+      normal! @a
+      call assert_equal(c, getline(1))
+
+      " Test with Shift modifier.
+      let shift_c = eval($'"\<S-{c}>"')
+      call setline(1, 'asdf')
+      call feedkeys($"qacc{s}\<Esc>gH{shift_c}\<Esc>q", "tx")
+      call assert_equal(c, getline(1))
+      call assert_equal($"cc{s}\<Esc>gH{shift_c}\<Esc>", @a)
+      call setline(1, 'asdf')
+      normal! @a
+      call assert_equal(c, getline(1))
+    endfor
+  endfor
+
+  bwipe!
+endfunc
+
+func Test_recording_with_select_mode_utf8()
+  call Run_test_recording_with_select_mode_utf8()
+endfunc
+
+" This must be done as one of the last tests, because it starts the GUI, which
+" cannot be undone.
+func Test_zz_recording_with_select_mode_utf8_gui()
+  CheckCanRunGui
+
+  gui -f
+  call Run_test_recording_with_select_mode_utf8()
+endfunc
+
 " Test for executing the last used register (@)
 func Test_last_used_exec_reg()
   " Test for the @: command
diff --git a/test/old/testdir/test_utf8.vim b/test/old/testdir/test_utf8.vim
index a5a9624ec3..920000e2c7 100644
--- a/test/old/testdir/test_utf8.vim
+++ b/test/old/testdir/test_utf8.vim
@@ -292,55 +292,4 @@ func Test_print_overlong()
   bwipe!
 endfunc
 
-func Test_recording_with_select_mode_utf8()
-  call Run_test_recording_with_select_mode_utf8()
-endfunc
-
-func Run_test_recording_with_select_mode_utf8()
-  new
-
-  " No escaping
-  call feedkeys("qacc12345\<Esc>gH哦\<Esc>q", "tx")
-  call assert_equal("哦", getline(1))
-  call assert_equal("cc12345\<Esc>gH哦\<Esc>", @a)
-  call setline(1, 'asdf')
-  normal! @a
-  call assert_equal("哦", getline(1))
-
-  " 固 is 0xE5 0x9B 0xBA where 0x9B is CSI
-  call feedkeys("qacc12345\<Esc>gH固\<Esc>q", "tx")
-  call assert_equal("固", getline(1))
-  call assert_equal("cc12345\<Esc>gH固\<Esc>", @a)
-  call setline(1, 'asdf')
-  normal! @a
-  call assert_equal("固", getline(1))
-
-  " 四 is 0xE5 0x9B 0x9B where 0x9B is CSI
-  call feedkeys("qacc12345\<Esc>gH四\<Esc>q", "tx")
-  call assert_equal("四", getline(1))
-  call assert_equal("cc12345\<Esc>gH四\<Esc>", @a)
-  call setline(1, 'asdf')
-  normal! @a
-  call assert_equal("四", getline(1))
-
-  " 倒 is 0xE5 0x80 0x92 where 0x80 is K_SPECIAL
-  call feedkeys("qacc12345\<Esc>gH倒\<Esc>q", "tx")
-  call assert_equal("倒", getline(1))
-  call assert_equal("cc12345\<Esc>gH倒\<Esc>", @a)
-  call setline(1, 'asdf')
-  normal! @a
-  call assert_equal("倒", getline(1))
-
-  bwipe!
-endfunc
-
-" This must be done as one of the last tests, because it starts the GUI, which
-" cannot be undone.
-func Test_zz_recording_with_select_mode_utf8_gui()
-  CheckCanRunGui
-
-  gui -f
-  call Run_test_recording_with_select_mode_utf8()
-endfunc
-
 " vim: shiftwidth=2 sts=2 expandtab