patch 8.2.2605: Vim9: string index and slice does not include composing chars
Problem: Vim9: string index and slice does not include composing chars.
Solution: Include composing characters. (issue #6563)
diff --git a/src/testdir/test_vim9_expr.vim b/src/testdir/test_vim9_expr.vim
index a78cf19..9829d88 100644
--- a/src/testdir/test_vim9_expr.vim
+++ b/src/testdir/test_vim9_expr.vim
@@ -2367,6 +2367,35 @@
assert_equal('abcd', g:teststring[: -3])
assert_equal('', g:teststring[: -9])
+ # composing characters are included
+ g:teststring = 'àéû'
+ assert_equal('à', g:teststring[0])
+ assert_equal('é', g:teststring[1])
+ assert_equal('û', g:teststring[2])
+ assert_equal('', g:teststring[3])
+ assert_equal('', g:teststring[4])
+
+ assert_equal('û', g:teststring[-1])
+ assert_equal('é', g:teststring[-2])
+ assert_equal('à', g:teststring[-3])
+ assert_equal('', g:teststring[-4])
+ assert_equal('', g:teststring[-5])
+
+ assert_equal('à', g:teststring[0 : 0])
+ assert_equal('é', g:teststring[1 : 1])
+ assert_equal('àé', g:teststring[0 : 1])
+ assert_equal('àéû', g:teststring[0 : -1])
+ assert_equal('àé', g:teststring[0 : -2])
+ assert_equal('à', g:teststring[0 : -3])
+ assert_equal('', g:teststring[0 : -4])
+ assert_equal('', g:teststring[0 : -5])
+ assert_equal('àéû', g:teststring[ : ])
+ assert_equal('àéû', g:teststring[0 : ])
+ assert_equal('éû', g:teststring[1 : ])
+ assert_equal('û', g:teststring[2 : ])
+ assert_equal('', g:teststring[3 : ])
+ assert_equal('', g:teststring[4 : ])
+
# blob index cannot be out of range
g:testblob = 0z01ab
assert_equal(0x01, g:testblob[0])
diff --git a/src/version.c b/src/version.c
index b4307fb..c5797c3 100644
--- a/src/version.c
+++ b/src/version.c
@@ -751,6 +751,8 @@
static int included_patches[] =
{ /* Add new patch number below this line */
/**/
+ 2605,
+/**/
2604,
/**/
2603,
diff --git a/src/vim9execute.c b/src/vim9execute.c
index b2c28a3..c0a4874 100644
--- a/src/vim9execute.c
+++ b/src/vim9execute.c
@@ -985,8 +985,9 @@
}
/*
- * Return the character "str[index]" where "index" is the character index. If
- * "index" is out of range NULL is returned.
+ * Return the character "str[index]" where "index" is the character index,
+ * including composing characters.
+ * If "index" is out of range NULL is returned.
*/
char_u *
char_from_string(char_u *str, varnumber_T index)
@@ -1005,7 +1006,7 @@
int clen = 0;
for (nbyte = 0; nbyte < slen; ++clen)
- nbyte += MB_CPTR2LEN(str + nbyte);
+ nbyte += mb_ptr2len(str + nbyte);
nchar = clen + index;
if (nchar < 0)
// unlike list: index out of range results in empty string
@@ -1013,15 +1014,15 @@
}
for (nbyte = 0; nchar > 0 && nbyte < slen; --nchar)
- nbyte += MB_CPTR2LEN(str + nbyte);
+ nbyte += mb_ptr2len(str + nbyte);
if (nbyte >= slen)
return NULL;
- return vim_strnsave(str + nbyte, MB_CPTR2LEN(str + nbyte));
+ return vim_strnsave(str + nbyte, mb_ptr2len(str + nbyte));
}
/*
* Get the byte index for character index "idx" in string "str" with length
- * "str_len".
+ * "str_len". Composing characters are included.
* If going over the end return "str_len".
* If "idx" is negative count from the end, -1 is the last character.
* When going over the start return -1.
@@ -1036,7 +1037,7 @@
{
while (nchar > 0 && nbyte < str_len)
{
- nbyte += MB_CPTR2LEN(str + nbyte);
+ nbyte += mb_ptr2len(str + nbyte);
--nchar;
}
}
@@ -1056,7 +1057,8 @@
}
/*
- * Return the slice "str[first:last]" using character indexes.
+ * Return the slice "str[first : last]" using character indexes. Composing
+ * characters are included.
* "exclusive" is TRUE for slice().
* Return NULL when the result is empty.
*/
@@ -1079,7 +1081,7 @@
end_byte = char_idx2byte(str, slen, last);
if (!exclusive && end_byte >= 0 && end_byte < (long)slen)
// end index is inclusive
- end_byte += MB_CPTR2LEN(str + end_byte);
+ end_byte += mb_ptr2len(str + end_byte);
}
if (start_byte >= (long)slen || end_byte <= start_byte)
@@ -3249,8 +3251,9 @@
res = string_slice(tv->vval.v_string, n1, n2, FALSE);
else
// Index: The resulting variable is a string of a
- // single character. If the index is too big or
- // negative the result is empty.
+ // single character (including composing characters).
+ // If the index is too big or negative the result is
+ // empty.
res = char_from_string(tv->vval.v_string, n2);
vim_free(tv->vval.v_string);
tv->vval.v_string = res;