patch 9.0.1629: having utf16idx() rounding up is inconvenient

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

diff --git a/runtime/doc/builtin.txt b/runtime/doc/builtin.txt
index b14851ca8918b59a18d4453fec93e7f9903359d9..1b7c6dcc6ef531918e05177cf6622c5aa09ac901 100644 (file)
--- a/runtime/doc/builtin.txt
+++ b/runtime/doc/builtin.txt
@@ -10127,8 +10127,8 @@ 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/strings.c b/src/strings.c
index 9e2c331bf4ed749788e36cc274a223285a89545a..e3be3354d39ee06f7b95009d09fac00896eba139 100644 (file)
--- a/src/strings.c
+++ b/src/strings.c
@@ -1743,8 +1743,10 @@ f_strtrans(typval_T *argvars, typval_T *rettv)
 
 
 /*
- *
  * "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)
@@ -1780,6 +1782,7 @@ f_utf16idx(typval_T *argvars, typval_T *rettv)
 
     char_u     *p;
     int                len;
+    int                utf16idx = 0;
     for (p = str, len = 0; charidx ? idx >= 0 : p <= str + idx; len++)
     {
        if (*p == NUL)
@@ -1791,6 +1794,7 @@ f_utf16idx(typval_T *argvars, typval_T *rettv)
                rettv->vval.v_number = len;
            return;
        }
+       utf16idx = len;
        int clen = ptr2len(p);
        int c = (clen > 1) ? utf_ptr2char(p) : *p;
        if (c > 0xFFFF)
@@ -1800,7 +1804,7 @@ f_utf16idx(typval_T *argvars, typval_T *rettv)
            idx--;
     }
 
-    rettv->vval.v_number = len > 0 ? len - 1 : 0;
+    rettv->vval.v_number = utf16idx;
 }
 
 /*

diff --git a/src/testdir/test_functions.vim b/src/testdir/test_functions.vim
index d7e7f923bab713f48bc2600517cdb7c34b5b5e80..a17c19e179b9516ffd0b57829ea4e6d76fc54caf 100644 (file)
--- a/src/testdir/test_functions.vim
+++ b/src/testdir/test_functions.vim
@@ -1518,14 +1518,14 @@ 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))
@@ -1621,8 +1621,8 @@ 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))

diff --git a/src/version.c b/src/version.c
index 1235ac94391d316467ed7f4d4b80d95be37067ae..c753acd3794212749e159fd61e22a79a55a8f1a4 100644 (file)
--- a/src/version.c
+++ b/src/version.c
@@ -695,6 +695,8 @@ static char *(features[]) =
 
 static int included_patches[] =
 {   /* Add new patch number below this line */
+/**/
+    1629,
 /**/
     1628,
 /**/