Mercurial > vim
changeset 32595:a699363f01cd v9.0.1629
patch 9.0.1629: having utf16idx() rounding up is inconvenient
Commit: https://github.com/vim/vim/commit/95707037afa1aeae4f3494dc623a721ceed7fc4e
Author: Yegappan Lakshmanan <yegappan@yahoo.com>
Date: Wed Jun 14 13:10:15 2023 +0100
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 https://github.com/vim/vim/issues/12523)
| author | Bram Moolenaar <Bram@vim.org> |
|---|---|
| date | Wed, 14 Jun 2023 14:15:04 +0200 |
| parents | b3dc1ac1ad16 |
| children | 1f51b1180d87 |
| files | runtime/doc/builtin.txt src/strings.c src/testdir/test_functions.vim src/version.c |
| diffstat | 4 files changed, 20 insertions(+), 14 deletions(-) [+] |
line wrap: on
line diff
--- a/runtime/doc/builtin.txt +++ b/runtime/doc/builtin.txt @@ -10127,8 +10127,8 @@ utf16idx({string}, {idx} [, {countcc} [, 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
--- a/src/strings.c +++ b/src/strings.c @@ -1743,8 +1743,10 @@ f_strtrans(typval_T *argvars, typval_T * /* + * "utf16idx()" function * - * "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 * 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->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 * idx--; } - rettv->vval.v_number = len > 0 ? len - 1 : 0; + rettv->vval.v_number = utf16idx; } /*
--- 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))