From 09f887758c03f2e434430ec0251750fa32d98a7e Mon Sep 17 00:00:00 2001
From: Thomas Koutcher <thomas.koutcher@online.fr>
Date: Tue, 23 Feb 2021 23:13:25 +0100
Subject: [PATCH] Fix loop when wrapping line with ISO8859-1 character

Fixes #1087
---
 include/tig/string.h |  3 +++
 src/pager.c          |  7 ++++++-
 src/string.c         | 38 ++++++++++++++++++++++++++++++++++----
 3 files changed, 43 insertions(+), 5 deletions(-)

diff --git a/include/tig/string.h b/include/tig/string.h
index 290d313db..2c01c7bbb 100644
--- a/include/tig/string.h
+++ b/include/tig/string.h
@@ -84,6 +84,7 @@ void string_copy_rev_from_commit_line(char *dst, const char *src);
 #define string_concat_path(dst, path1, path2) \
 	string_format(dst, !*path1 || path1[strlen(path1) - 1] == '/' ? "%s%s" : "%s/%s", path1, path2)
 
+size_t string_expanded_length(const char *src, size_t srclen, size_t tabsize, size_t max_size);
 size_t string_expand(char *dst, size_t dstlen, const char *src, int srclen, int tabsize);
 
 char *string_trim_end(char *name);
@@ -125,5 +126,7 @@ int utf8_width_of(const char *text, int max_bytes, int max_width);
 
 bool utf8_string_contains_uppercase(const char *text);
 
+bool utf8_is_valid(const char *string);
+
 #endif
 /* vim: set ts=8 sw=8 noexpandtab: */
diff --git a/src/pager.c b/src/pager.c
index b570de262..69d08d69f 100644
--- a/src/pager.c
+++ b/src/pager.c
@@ -77,9 +77,14 @@ pager_wrap_line(struct view *view, const char *data, enum line_type type)
 		int width;
 		int trimmed;
 		bool wrapped = !!first_line;
-		size_t linelen = utf8_length(&data, datalen, 0, &width, view->width, &trimmed, wrapped, opt_tab_size);
+		size_t linelen;
 		struct line *line;
 
+		if (utf8_is_valid(data))
+			linelen = utf8_length(&data, datalen, 0, &width, view->width, &trimmed, wrapped, opt_tab_size);
+		else
+			linelen = string_expanded_length(data, datalen, opt_tab_size, view->width - !!wrapped);
+
 		line = add_line_text_at_(view, view->lines, data, linelen, type, 1, wrapped);
 		if (!line)
 			break;
diff --git a/src/string.c b/src/string.c
index 2638d9254..27fce6220 100644
--- a/src/string.c
+++ b/src/string.c
@@ -88,6 +88,24 @@ string_copy_rev_from_commit_line(char *dst, const char *src)
 	string_copy_rev(dst, src);
 }
 
+size_t
+string_expanded_length(const char *src, size_t srclen, size_t tabsize, size_t max_size)
+{
+	size_t size, pos;
+
+	for (size = pos = 0; pos < srclen && size < max_size; pos++) {
+		if (src[pos] == '\t') {
+			size_t expanded = tabsize - (size % tabsize);
+
+			size += expanded;
+		} else {
+			size++;
+		}
+	}
+
+	return pos;
+}
+
 size_t
 string_expand(char *dst, size_t dstlen, const char *src, int srclen, int tabsize)
 {
@@ -311,10 +329,6 @@ utf8_length(const char **start, int max_chars, size_t skip, int *width, size_t m
 		 * it is a single- or double-width character. */
 
 		unicode = utf8_to_unicode(string, bytes);
-		/* FIXME: Graceful handling of invalid Unicode character. */
-		if (!unicode)
-			break;
-
 		ucwidth = unicode == '\t' ? tab_size - (*width % tab_size) :
 					    utf8proc_charwidth((utf8proc_int32_t) unicode);
 		if (skip > 0) {
@@ -385,4 +399,20 @@ utf8_string_contains_uppercase(const char *search)
 	return utf8_string_contains(search, UTF8PROC_CATEGORY_LU);
 }
 
+bool
+utf8_is_valid(const char *string)
+{
+	for (;;) {
+		int32_t unicode;
+		ssize_t slen = utf8proc_iterate((const unsigned char *) string, -1, &unicode);
+
+		if (slen <= 0 || !utf8proc_codepoint_valid(unicode))
+			break;
+		if (unicode == 0)
+			return true;
+		string += slen;
+	}
+	return false;
+}
+
 /* vim: set ts=8 sw=8 noexpandtab: */