From cfdf68a7acde16597fbd896674af68c42361102c Mon Sep 17 00:00:00 2001
From: bfredl <bjorn.linse@gmail.com>
Date: Thu, 8 Aug 2024 10:42:08 +0200
Subject: feat(mbyte): support extended grapheme clusters including more emoji

Use the grapheme break algorithm from utf8proc to support grapheme
clusters from recent unicode versions.

Handle variant selector VS16 turning some codepoints into double-width
emoji. This means we need to use ptr2cells rather than char2cells when
possible.
---
 src/nvim/drawline.c | 22 +++++++++++-----------
 1 file changed, 11 insertions(+), 11 deletions(-)

(limited to 'src/nvim/drawline.c')

diff --git a/src/nvim/drawline.c b/src/nvim/drawline.c
index 8a948716e5..b5273a54ca 100644
--- a/src/nvim/drawline.c
+++ b/src/nvim/drawline.c
@@ -1826,7 +1826,7 @@ int win_line(win_T *wp, linenr_T lnum, int startrow, int endrow, int col_rows, s
 
         // If a double-width char doesn't fit display a '>' in the last column.
         // Don't advance the pointer but put the character at the start of the next line.
-        if (wlv.col >= grid->cols - 1 && utf_char2cells(mb_c) == 2) {
+        if (wlv.col >= grid->cols - 1 && schar_cells(mb_schar) == 2) {
           mb_c = '>';
           mb_l = 1;
           (void)mb_l;
@@ -1922,7 +1922,7 @@ int win_line(win_T *wp, linenr_T lnum, int startrow, int endrow, int col_rows, s
       // If a double-width char doesn't fit display a '>' in the
       // last column; the character is displayed at the start of the
       // next line.
-      if (wlv.col >= grid->cols - 1 && utf_char2cells(mb_c) == 2) {
+      if (wlv.col >= grid->cols - 1 && schar_cells(mb_schar) == 2) {
         mb_schar = schar_from_ascii('>');
         mb_c = '>';
         mb_l = 1;
@@ -2393,6 +2393,12 @@ int win_line(win_T *wp, linenr_T lnum, int startrow, int endrow, int col_rows, s
                 || (decor_conceal && decor_state.conceal_char)
                 || wp->w_p_cole == 1)
             && wp->w_p_cole != 3) {
+          if (schar_cells(mb_schar) > 1) {
+            // When the first char to be concealed is double-width,
+            // need to advance one more virtual column.
+            wlv.n_extra++;
+          }
+
           // First time at this concealed item: display one
           // character.
           if (has_match_conc && match_conc) {
@@ -2410,12 +2416,6 @@ int win_line(win_T *wp, linenr_T lnum, int startrow, int endrow, int col_rows, s
             mb_schar = schar_from_ascii(' ');
           }
 
-          if (utf_char2cells(mb_c) > 1) {
-            // When the first char to be concealed is double-width,
-            // need to advance one more virtual column.
-            wlv.n_extra++;
-          }
-
           mb_c = schar_get_first_codepoint(mb_schar);
 
           prev_syntax_id = syntax_seqnr;
@@ -2484,7 +2484,7 @@ int win_line(win_T *wp, linenr_T lnum, int startrow, int endrow, int col_rows, s
         && mb_schar != NUL) {
       mb_schar = wp->w_p_lcs_chars.prec;
       lcs_prec_todo = NUL;
-      if (utf_char2cells(mb_c) > 1) {
+      if (schar_cells(mb_schar) > 1) {
         // Double-width character being overwritten by the "precedes"
         // character, need to fill up half the character.
         wlv.sc_extra = schar_from_ascii(MB_FILLER_CHAR);
@@ -2725,7 +2725,7 @@ int win_line(win_T *wp, linenr_T lnum, int startrow, int endrow, int col_rows, s
 
       linebuf_vcol[wlv.off] = wlv.vcol;
 
-      if (utf_char2cells(mb_c) > 1) {
+      if (schar_cells(mb_schar) > 1) {
         // Need to fill two screen columns.
         wlv.off++;
         wlv.col++;
@@ -2744,7 +2744,7 @@ int win_line(win_T *wp, linenr_T lnum, int startrow, int endrow, int col_rows, s
       wlv.off++;
       wlv.col++;
     } else if (wp->w_p_cole > 0 && is_concealing) {
-      bool concealed_wide = utf_char2cells(mb_c) > 1;
+      bool concealed_wide = schar_cells(mb_schar) > 1;
 
       wlv.skip_cells--;
       wlv.vcol_off_co++;
-- 
cgit