vim-patch:9.0.1629: having utf16idx() rounding up is inconvenient (#24019)

Problem:    Having utf16idx() rounding up is inconvenient.
Solution:   Make utf16idx() round down. (Yegappan Lakshmanan, closes vim/vim#12523)

https://github.com/vim/vim/commit/95707037afa1aeae4f3494dc623a721ceed7fc4e

Co-authored-by: Yegappan Lakshmanan <yegappan@yahoo.com>

3 files changed, 18 insertions, 13 deletions

diff --git a/runtime/doc/builtin.txt b/runtime/doc/builtin.txt
index 1fea6ad715..938523b6ee 100644
--- a/runtime/doc/builtin.txt
+++ b/runtime/doc/builtin.txt
@@ -9146,8 +9146,8 @@ utf16idx({string}, {idx} [, {countcc} [, {charidx}]])
 		When {charidx} is present and TRUE, {idx} is used as the
 		character index in the String {string} instead of as the byte
 		index.
-		An {idx} in the middle of a UTF-8 sequence is rounded upwards
-		to the end of that sequence.
+		An {idx} in the middle of a UTF-8 sequence is rounded
+		downwards to the beginning of that sequence.
 
 		Returns -1 if the arguments are invalid or if there are less
 		than {idx} bytes in {string}. If there are exactly {idx} bytes
diff --git a/src/nvim/strings.c b/src/nvim/strings.c
index a0d62f5df5..ada3a45d60 100644
--- a/src/nvim/strings.c
+++ b/src/nvim/strings.c
@@ -2014,6 +2014,9 @@ void f_strtrans(typval_T *argvars, typval_T *rettv, EvalFuncData fptr)
 }
 
 /// "utf16idx()" function
+///
+/// Converts a byte or character offset in a string to the corresponding UTF-16
+/// code unit offset.
 void f_utf16idx(typval_T *argvars, typval_T *rettv, EvalFuncData fptr)
 {
   rettv->vval.v_number = -1;
@@ -2050,6 +2053,7 @@ void f_utf16idx(typval_T *argvars, typval_T *rettv, EvalFuncData fptr)
 
   const char *p;
   int len;
+  int utf16idx = 0;
   for (p = str, len = 0; charidx ? idx >= 0 : p <= str + idx; len++) {
     if (*p == NUL) {
       // If the index is exactly the number of bytes or characters in the
@@ -2059,6 +2063,7 @@ void f_utf16idx(typval_T *argvars, typval_T *rettv, EvalFuncData fptr)
       }
       return;
     }
+    utf16idx = len;
     const int clen = ptr2len(p);
     const int c = (clen > 1) ? utf_ptr2char(p) : *p;
     if (c > 0xFFFF) {
@@ -2070,7 +2075,7 @@ void f_utf16idx(typval_T *argvars, typval_T *rettv, EvalFuncData fptr)
     }
   }
 
-  rettv->vval.v_number = len > 0 ? len - 1 : 0;
+  rettv->vval.v_number = utf16idx;
 }
 
 /// "tolower(string)" function
diff --git a/test/old/testdir/test_functions.vim b/test/old/testdir/test_functions.vim
index d80004d70f..0fca0e322e 100644
--- a/test/old/testdir/test_functions.vim
+++ b/test/old/testdir/test_functions.vim
@@ -1395,14 +1395,14 @@ func Test_utf16idx_from_byteidx()
   " UTF-16 index of a string with four byte characters
   let str = 'a😊😊b'
   call assert_equal(0, utf16idx(str, 0))
-  call assert_equal(2, utf16idx(str, 1))
-  call assert_equal(2, utf16idx(str, 2))
-  call assert_equal(2, utf16idx(str, 3))
-  call assert_equal(2, utf16idx(str, 4))
-  call assert_equal(4, utf16idx(str, 5))
-  call assert_equal(4, utf16idx(str, 6))
-  call assert_equal(4, utf16idx(str, 7))
-  call assert_equal(4, utf16idx(str, 8))
+  call assert_equal(1, utf16idx(str, 1))
+  call assert_equal(1, utf16idx(str, 2))
+  call assert_equal(1, utf16idx(str, 3))
+  call assert_equal(1, utf16idx(str, 4))
+  call assert_equal(3, utf16idx(str, 5))
+  call assert_equal(3, utf16idx(str, 6))
+  call assert_equal(3, utf16idx(str, 7))
+  call assert_equal(3, utf16idx(str, 8))
   call assert_equal(5, utf16idx(str, 9))
   call assert_equal(6, utf16idx(str, 10))
   call assert_equal(-1, utf16idx(str, 11))
@@ -1498,8 +1498,8 @@ func Test_utf16idx_from_charidx()
   " UTF-16 index of a string with four byte characters
   let str = "a😊😊b"
   call assert_equal(0, utf16idx(str, 0, v:false, v:true))
-  call assert_equal(2, utf16idx(str, 1, v:false, v:true))
-  call assert_equal(4, utf16idx(str, 2, v:false, v:true))
+  call assert_equal(1, utf16idx(str, 1, v:false, v:true))
+  call assert_equal(3, utf16idx(str, 2, v:false, v:true))
   call assert_equal(5, utf16idx(str, 3, v:false, v:true))
   call assert_equal(6, utf16idx(str, 4, v:false, v:true))
   call assert_equal(-1, utf16idx(str, 5, v:false, v:true))