From 8ae019ffc34b98bcd8546598de6037addaa29a4f Mon Sep 17 00:00:00 2001
From: Silvan Jegen <s.jegen@gmail.com>
Date: Sat, 10 Apr 2021 13:40:51 +0200
Subject: vis-menu: try to preserve valid Unicode points

Before we were not taking non-ascii characters into account properly. With
this patch we still mix byte counts and "grapheme cluster" (i.e. complete
glyphs that are rendered in a terminal cell) counts but the code should
be less broken in the more common case now.
---
 vis-menu.c | 40 +++++++++++++++++++++++++++++++++++++---
 1 file changed, 37 insertions(+), 3 deletions(-)

(limited to 'vis-menu.c')

diff --git a/vis-menu.c b/vis-menu.c
index 444a26a..66e23e3 100644
--- a/vis-menu.c
+++ b/vis-menu.c
@@ -84,12 +84,46 @@ appenditem(Item *item, Item **list, Item **last) {
 
 static size_t
 textwn(const char *s, int l) {
-	int b, c; /* bytes and UTF-8 characters */
+	int c;
 
-	for(b=c=0; s && s[b] && (l<0 || b<l); b++) if((s[b] & 0xc0) != 0x80) c++;
+	for (c=0; s && s[c] && (l<0 || c<l); c++);
 	return c+4; /* Accomodate for the leading and trailing spaces */
 }
 
+/*
+ * textvalidn returns the highest amount of bytes <= l of string s that
+ * only contains valid Unicode points. This is used to make sure we don't
+ * cut off any valid UTF-8-encoded unicode point in case there is not
+ * enough space to render the whole text string.
+*/
+static ssize_t
+textvalidn(const char *s, int l) {
+  int c, utfcharbytes; /* byte count and UTF-8 codepoint length */
+
+  for (c=0; s && s[c] && (l<0 || c<l); ) {
+		utfcharbytes = 0;
+		if ((s[c] & 0x80) == 0) {
+			utfcharbytes = 1;
+		} else if ((s[c] & 0xf0) == 0xf0) {
+			utfcharbytes = 4;
+		} else if ((s[c] & 0xf0) == 0xe0) {
+			utfcharbytes = 3;
+		} else if ((s[c] & 0xe0) == 0xc0) {
+			utfcharbytes = 2;
+		} else {
+			return -1;
+		}
+
+		if ((l>0 && c + utfcharbytes >= l)) {
+			break;
+		}
+
+		c += utfcharbytes;
+  }
+
+	return c;
+}
+
 static size_t
 textw(const char *s) {
 	return textwn(s, -1);
@@ -149,7 +183,7 @@ drawtext(const char *t, size_t w, Color col) {
 	buf[tw] = '\0';
 	memcpy(buf, t, MIN(strlen(t), tw));
 	if (textw(t) > w) /* Remember textw returns the width WITH padding */
-		for (i = MAX((tw-4), 0); i < tw; i++) buf[i] = '.';
+		for (i = MAX(textvalidn(t, w-4), 0); i < tw; i++) buf[i] = '.';
 
 	fprintf(stderr, "%s  %s  %s", prestr, buf, poststr);
 	free(buf);
-- 
cgit v1.2.3