commit d5e0ee089932fa61ce56746c09c77019353d5614
from: Omar Polo <op@omarpolo.com>
date: Sun Oct 09 16:01:43 2022 UTC

use libgrapheme for text reflowing

unlike the previous hack, libgrapheme is UNICODE 15.0 compliant (at the
time of writing) and provides some nice API.  While here, drop the hard
wrap method for preformatted blocks: always use the normal wrapping.

Since libgrapheme is not yet widely available, future commits will
probably add a bundled version to ease packaging.  I'll also start to
replace various home-grown UNICODE/utf8 handling routines with functions
from libgrapheme.

commit - db6fcbbe651bdf690df980ff1a4f73310b33dd87
commit + d5e0ee089932fa61ce56746c09c77019353d5614
blob - f037f5f80862c31a7ab21a42c9eb786c04e10776
blob + 5273190b7424f14ff5247c661d582ab711dfa994
--- configure.ac
+++ configure.ac
@@ -82,6 +82,10 @@ AC_CHECK_LIB(tls, tls_init, [], [
 	AC_MSG_ERROR([requires libtls])
 ])
 
+AC_CHECK_LIB(grapheme, grapheme_next_line_break_utf8, [], [
+	AC_MSG_ERROR([requires libgrapheme])
+])
+
 case "$host_os" in
 	*openbsd*)
 		AC_CHECK_LIB([event], [event_init], [],
blob - 1c7a101b1280b8e8775b23e98b543b0df4b4cb88
blob + fcffadb7e89a7c3f8800eee0aacb3e232a4c4be0
--- include/telescope.h
+++ include/telescope.h
@@ -347,7 +347,6 @@ void		 empty_linelist(struct buffer*);
 void		 empty_vlist(struct buffer*);
 int		 wrap_one(struct buffer *, const char *, struct line *, size_t);
 int		 wrap_text(struct buffer*, const char*, struct line*, size_t);
-int		 hardwrap_text(struct buffer*, struct line*, size_t);
 int		 wrap_page(struct buffer *, int width);
 
 #endif /* TELESCOPE_H */
blob - 338068315003a7f604337fc0f6c91a4b98a7739e
blob + 7c359d32f7211243d1f277da709a91db6df22fde
--- u/wrap.c
+++ u/wrap.c
@@ -17,29 +17,16 @@
 #include "compat.h"
 
 #include <ctype.h>
+#include <stdint.h>
 #include <stdlib.h>
 #include <string.h>
 
+#include <grapheme.h>
+
 #include "defaults.h"
 #include "telescope.h"
 #include "utf8.h"
 
-/*
- * Text wrapping
- * =============
- *
- * There's a simple text wrapping algorithm.
- *
- * 1. if it's a line in a pre-formatted block:
- *    a. hard wrap.
- *    b. repeat
- * 2. otherwise advance the line char by char.
- * 3. when ending the space, split the line at the last occurrence of
- *    a "word separator" (i.e. " \t-") or at point if none.
- * 4. repeat
- *
- */
-
 void
 erase_buffer(struct buffer *buffer)
 {
@@ -150,20 +137,16 @@ wrap_one(struct buffer *buffer, const char *prfx, stru
 int
 wrap_text(struct buffer *buffer, const char *prfx, struct line *l, size_t width)
 {
-	const char	*separators = " \t-";
-	const char	*start, *end, *line, *lastsep, *lastchar, *space;
-	uint32_t	 cp = 0, state = 0;
-	size_t		 cur, prfxwidth, w;
+	const char	*line, *space;
+	size_t		 ret, off, start, cur, prfxwidth;
 	int		 flags;
 
-	if ((line = l->line) == NULL)
+	if ((line = l->line) == NULL || *line == '\0')
 		return push_line(buffer, l, NULL, 0, 0);
 
 	prfxwidth = utf8_swidth(prfx);
 	cur = prfxwidth;
-	start = line;
-	lastsep = NULL;
-	lastchar = line;
+	start = 0;
 	flags = 0;
 
 	if (l->type == LINE_LINK && emojify_link &&
@@ -173,64 +156,28 @@ wrap_text(struct buffer *buffer, const char *prfx, str
 		line = space + 1;
 	}
 
-	for (; *line; line++) {
-		if (utf8_decode(&state, &cp, *line))
-			continue;
-		w = utf8_chwidth(cp);
-		if (cur + w > width) {
-			end = lastsep == NULL
-				? utf8_next_cp((char*)lastchar)
-				: utf8_next_cp((char*)lastsep);
-			if (!push_line(buffer, l, start, end - start, flags))
-				return 0;
-			flags = L_CONTINUATION;
-			start = end;
-			cur = prfxwidth + utf8_swidth_between(start, lastchar);
-		} else if (strchr(separators, *line) != NULL) {
-			lastsep = line;
-		}
-
-		lastchar = utf8_prev_cp(line, l->line);
-		cur += w;
-	}
+	for (off = 0; line[off] != '\0'; off += ret) {
+		size_t t;
 
-	return push_line(buffer, l, start, line - start, flags);
-}
+		ret = grapheme_next_line_break_utf8(&line[off], SIZE_MAX);
+		t = utf8_swidth_between(&line[off], &line[off + ret]);
 
-int
-hardwrap_text(struct buffer *buffer, struct line *l, size_t width)
-{
-	const char	*line, *start, *lastchar;
-	int		 cont;
-	uint32_t	 state = 0, cp = 0;
-	size_t		 cur, w;
-
-	if ((line = l->line) == NULL)
-		return push_line(buffer, l, NULL, 0, 0);
-
-	start = line;
-	lastchar = line;
-	cont = 0;
-	cur = 0;
-	for (; *line; line++) {
-		if (utf8_decode(&state, &cp, *line))
+		if (cur + t <= width) {
+			cur += t;
 			continue;
-		w = utf8_chwidth(cp);
-		if (cur + w > width) {
-			if (!push_line(buffer, l, start, lastchar-start, cont))
-				return 0;
-			cont = L_CONTINUATION;
-			if (dont_wrap_pre)
-				return 1;
-			cur = 0;
-			start = lastchar;
 		}
 
-		lastchar = utf8_prev_cp(line, l->line);
-		cur += w;
+		if (!push_line(buffer, l, &line[start], off - start, flags))
+			return 0;
+
+		flags = L_CONTINUATION;
+		start = off;
+		cur = t;
 	}
 
-	return push_line(buffer, l, start, line - start, cont);
+	if (off != start)
+		return push_line(buffer, l, &line[start], off - start, flags);
+	return 0;
 }
 
 int
@@ -265,15 +212,13 @@ wrap_page(struct buffer *buffer, int width)
 		case LINE_QUOTE:
 		case LINE_PRE_START:
 		case LINE_PRE_END:
-			wrap_text(buffer, prfx, l, MIN(fill_column, width));
-			break;
 		case LINE_PRE_CONTENT:
 		case LINE_PATCH:
 		case LINE_PATCH_HDR:
 		case LINE_PATCH_HUNK_HDR:
 		case LINE_PATCH_ADD:
 		case LINE_PATCH_DEL:
-			hardwrap_text(buffer, l, MIN(fill_column, width));
+			wrap_text(buffer, prfx, l, MIN(fill_column, width));
 			break;
 		case LINE_COMPL:
 		case LINE_COMPL_CURRENT: