diff options
| -rw-r--r-- | runtime/doc/options.txt | 5 | ||||
| -rw-r--r-- | runtime/lua/vim/_meta/options.lua | 5 | ||||
| -rw-r--r-- | src/nvim/diff.c | 9 | ||||
| -rw-r--r-- | src/nvim/mbyte.c | 10 | ||||
| -rw-r--r-- | src/nvim/options.lua | 5 | ||||
| -rw-r--r-- | test/functional/ui/diff_spec.lua | 16 | ||||
| -rw-r--r-- | test/old/testdir/test_diffmode.vim | 5 |
7 files changed, 45 insertions, 10 deletions
diff --git a/runtime/doc/options.txt b/runtime/doc/options.txt index 972d2cce51..e94b72a260 100644 --- a/runtime/doc/options.txt +++ b/runtime/doc/options.txt @@ -2132,7 +2132,10 @@ A jump table for the options with a short description can be found at |Q_op|. difference. word Use internal diff to perform a |word|-wise diff and highlight the - difference. + difference. Non-alphanumeric + multi-byte characters such as emoji + and CJK characters are considered + individual words. internal Use the internal diff library. This is ignored when 'diffexpr' is set. *E960* diff --git a/runtime/lua/vim/_meta/options.lua b/runtime/lua/vim/_meta/options.lua index 0fc8518aa4..c032edf43c 100644 --- a/runtime/lua/vim/_meta/options.lua +++ b/runtime/lua/vim/_meta/options.lua @@ -1729,7 +1729,10 @@ vim.go.dex = vim.go.diffexpr --- difference. --- word Use internal diff to perform a --- `word`-wise diff and highlight the ---- difference. +--- difference. Non-alphanumeric +--- multi-byte characters such as emoji +--- and CJK characters are considered +--- individual words. --- --- internal Use the internal diff library. This is --- ignored when 'diffexpr' is set. *E960* diff --git a/src/nvim/diff.c b/src/nvim/diff.c index 6309aa6c5e..ad3e093dde 100644 --- a/src/nvim/diff.c +++ b/src/nvim/diff.c @@ -2990,10 +2990,15 @@ static void diff_find_change_inline_diff(diff_T *dp) char *s = curline; while (*s != NUL) { - // Always use the first buffer's 'iskeyword' to have a consistent diff bool new_in_keyword = false; if (diff_flags & DIFF_INLINE_WORD) { - new_in_keyword = vim_iswordp_buf(s, curtab->tp_diffbuf[file1_idx]); + // Always use the first buffer's 'iskeyword' to have a + // consistent diff. + // For multibyte chars, only treat alphanumeric chars + // (class 2) as "word", as other classes such as emojis and + // CJK ideographs do not usually benefit from word diff as + // Vim doesn't have a good way to segment them. + new_in_keyword = (mb_get_class_tab(s, curtab->tp_diffbuf[file1_idx]->b_chartab) == 2); } if (in_keyword && !new_in_keyword) { ga_append(curstr, NL); diff --git a/src/nvim/mbyte.c b/src/nvim/mbyte.c index add650e7a9..18df6b7e76 100644 --- a/src/nvim/mbyte.c +++ b/src/nvim/mbyte.c @@ -417,11 +417,11 @@ void remove_bom(char *s) } } -// Get class of pointer: -// 0 for blank or NUL -// 1 for punctuation -// 2 for an (ASCII) word character -// >2 for other word characters +/// Get class of pointer: +/// 0 for blank or NUL +/// 1 for punctuation +/// 2 for an alphanumeric word character +/// >2 for other word characters, including CJK and emoji int mb_get_class(const char *p) FUNC_ATTR_PURE { diff --git a/src/nvim/options.lua b/src/nvim/options.lua index 0a807979ac..a5171ed58b 100644 --- a/src/nvim/options.lua +++ b/src/nvim/options.lua @@ -2286,7 +2286,10 @@ local options = { difference. word Use internal diff to perform a |word|-wise diff and highlight the - difference. + difference. Non-alphanumeric + multi-byte characters such as emoji + and CJK characters are considered + individual words. internal Use the internal diff library. This is ignored when 'diffexpr' is set. *E960* diff --git a/test/functional/ui/diff_spec.lua b/test/functional/ui/diff_spec.lua index 223622eb6b..21174eaa45 100644 --- a/test/functional/ui/diff_spec.lua +++ b/test/functional/ui/diff_spec.lua @@ -2485,6 +2485,22 @@ it('diff mode inline highlighting', function() command('windo set iskeyword& | 1wincmd w') + screen:try_resize(75, 20) + command('wincmd =') + -- word diff: test handling of multi-byte characters. Only alphanumeric chars + -- (e.g. Greek alphabet, but not CJK/emoji) count as words. + WriteDiffFiles( + '🚀⛵️一二三ひらがなΔέλτα Δelta foobar', + '🚀🛸一二四ひらなδέλτα δelta foobar' + ) + command('set diffopt=internal,filler diffopt+=inline:word') + screen:expect([[ + {7: }{4:^🚀}{27:⛵️}{4:一二}{27:三}{4:ひら}{100:が}{4:な}{27:Δέλτα}{4: }{27:Δelta}{4: fooba}│{7: }{4:🚀}{27:🛸}{4:一二}{27:四}{4:ひらな}{27:δέλτα}{4: }{27:δelta}{4: foobar }| + {1:~ }│{1:~ }|*17 + {3:Xdifile1 }{2:Xdifile2 }| + | + ]]) + screen:try_resize(69, 20) command('wincmd =') -- char diff: should slide highlight to whitespace boundary if possible for diff --git a/test/old/testdir/test_diffmode.vim b/test/old/testdir/test_diffmode.vim index 6a5054b37b..2d3bcb8234 100644 --- a/test/old/testdir/test_diffmode.vim +++ b/test/old/testdir/test_diffmode.vim @@ -2193,6 +2193,11 @@ func Test_diff_inline() call term_sendkeys(buf, ":windo set iskeyword&\<CR>:1wincmd w\<CR>") + " word diff: test handling of multi-byte characters. Only alphanumeric chars + " (e.g. Greek alphabet, but not CJK/emoji) count as words. + call WriteDiffFiles(buf, ["🚀⛵️一二三ひらがなΔέλτα Δelta foobar"], ["🚀🛸一二四ひらなδέλτα δelta foobar"]) + call VerifyInternal(buf, "Test_diff_inline_word_03", " diffopt+=inline:word") + " char diff: should slide highlight to whitespace boundary if possible for " better readability (by using forced indent-heuristics). A wrong result " would be if the highlight is "Bar, prefix". It should be "prefixBar, " |