Cut text between characters, never inside one

The index description was cut at a byte count, which could split a
multibyte character and leave an invalid sequence on the page. Commit
subjects already backed up to the start of the character, so html_ntxt
now does that for every caller.

5 files changed, 34 insertions, 9 deletions

diff --git a/source/html.c b/source/html.cindex 1a6b7c5..b6e62a6 100644--- a/source/html.c+++ b/source/html.c@@ -247,6 +247,12 @@ ssize_t html_ntxt(const char *txt, size_t len) 		} 		p++; 	}+	// A cut inside a multibyte character would leave an invalid sequence+	// on the page, so it backs up to where that character starts. The+	// escaped bytes are all ASCII, so no character spans one.+	if (left < 0)+		while (p > txt && (*p & 0xC0) == 0x80)+			p--; 	if (p != txt) 		html_raw(txt, p - txt); 	return left;
diff --git a/source/html.c b/source/html.c
index 1a6b7c5..b6e62a6 100644
--- a/source/html.c
+++ b/source/html.c
@@ -247,6 +247,12 @@ ssize_t html_ntxt(const char *txt, size_t len)
 		}
 		p++;
 	}
+	// A cut inside a multibyte character would leave an invalid sequence
+	// on the page, so it backs up to where that character starts. The
+	// escaped bytes are all ASCII, so no character spans one.
+	if (left < 0)
+		while (p > txt && (*p & 0xC0) == 0x80)
+			p--;
 	if (p != txt)
 		html_raw(txt, p - txt);
 	return left;
diff --git a/source/html.h b/source/html.hindex 095804c..c6c0e19 100644--- a/source/html.h+++ b/source/html.h@@ -55,8 +55,9 @@ extern void html_attrf(const char *format, ...); extern void html_txt(const char *txt);  /*- * Escape at most len bytes of txt. The return is what is left of that budget,- * and a caller prints an ellipsis when it comes back negative.+ * Escape at most len bytes of txt, cutting short of a character the budget+ * would split. The return is what is left of that budget, and a caller prints+ * an ellipsis when it comes back negative.  */ extern ssize_t html_ntxt(const char *txt, size_t len); 
diff --git a/source/html.h b/source/html.h
index 095804c..c6c0e19 100644
--- a/source/html.h
+++ b/source/html.h
@@ -55,8 +55,9 @@ extern void html_attrf(const char *format, ...);
 extern void html_txt(const char *txt);
 
 /*
- * Escape at most len bytes of txt. The return is what is left of that budget,
- * and a caller prints an ellipsis when it comes back negative.
+ * Escape at most len bytes of txt, cutting short of a character the budget
+ * would split. The return is what is left of that budget, and a caller prints
+ * an ellipsis when it comes back negative.
  */
 extern ssize_t html_ntxt(const char *txt, size_t len);
 
diff --git a/source/ui-shared.c b/source/ui-shared.cindex 7d28c39..f3d815f 100644--- a/source/ui-shared.c+++ b/source/ui-shared.c@@ -1039,13 +1039,7 @@ void cgit_commit_link(const char *name, const char *title, const char *class, 	html("'>"); 	if (name && name[0] != '\0') { 		if (ctx.cfg.max_msg_len >= MIN_TRUNCATE_LEN && strlen(name) > (size_t)ctx.cfg.max_msg_len) {-			size_t len = ctx.cfg.max_msg_len - ELLIPSIS_LEN;--			// A cut inside a multibyte character would leave an-			// invalid sequence on the page.-			while (len > 0 && (name[len] & 0xC0) == 0x80)-				len--;-			html_ntxt(name, len);+			html_ntxt(name, ctx.cfg.max_msg_len - ELLIPSIS_LEN); 			html("..."); 		} else { 			html_txt(name);
diff --git a/source/ui-shared.c b/source/ui-shared.c
index 7d28c39..f3d815f 100644
--- a/source/ui-shared.c
+++ b/source/ui-shared.c
@@ -1039,13 +1039,7 @@ void cgit_commit_link(const char *name, const char *title, const char *class,
 	html("'>");
 	if (name && name[0] != '\0') {
 		if (ctx.cfg.max_msg_len >= MIN_TRUNCATE_LEN && strlen(name) > (size_t)ctx.cfg.max_msg_len) {
-			size_t len = ctx.cfg.max_msg_len - ELLIPSIS_LEN;
-
-			// A cut inside a multibyte character would leave an
-			// invalid sequence on the page.
-			while (len > 0 && (name[len] & 0xC0) == 0x80)
-				len--;
-			html_ntxt(name, len);
+			html_ntxt(name, ctx.cfg.max_msg_len - ELLIPSIS_LEN);
 			html("...");
 		} else {
 			html_txt(name);
diff --git a/tests/t0101-index.sh b/tests/t0101-index.shindex adad6be..e433beb 100755--- a/tests/t0101-index.sh+++ b/tests/t0101-index.sh@@ -53,4 +53,18 @@ test_expect_success 'the feed link encodes the repository and the branch' ' 	grep "title=.Atom feed. href=.http://example.org/foo%2Bbar/atom/?h=1%2B2." tmp ' +# The description limit counts bytes, so a cut has to land between+# characters rather than leave half of one on the page.+test_expect_success 'a cut description keeps whole characters' '+	{+		echo "max-repodesc-length=10" &&+		sed "s/^cache-size=.*/cache-size=0/" cgitrc &&+		echo "repo.url=mb" &&+		echo "repo.path=$PWD/repos/foo/.git" &&+		echo "repo.desc=日本語の説明文です"+	} >descrc &&+	CGIT_CONFIG="$PWD/descrc" QUERY_STRING="url=" cgit >tmp &&+	grep ">日本語\.\.\.</a>" tmp+'+ test_done
diff --git a/tests/t0101-index.sh b/tests/t0101-index.sh
index adad6be..e433beb 100755
--- a/tests/t0101-index.sh
+++ b/tests/t0101-index.sh
@@ -53,4 +53,18 @@ test_expect_success 'the feed link encodes the repository and the branch' '
 	grep "title=.Atom feed. href=.http://example.org/foo%2Bbar/atom/?h=1%2B2." tmp
 '
 
+# The description limit counts bytes, so a cut has to land between
+# characters rather than leave half of one on the page.
+test_expect_success 'a cut description keeps whole characters' '
+	{
+		echo "max-repodesc-length=10" &&
+		sed "s/^cache-size=.*/cache-size=0/" cgitrc &&
+		echo "repo.url=mb" &&
+		echo "repo.path=$PWD/repos/foo/.git" &&
+		echo "repo.desc=日本語の説明文です"
+	} >descrc &&
+	CGIT_CONFIG="$PWD/descrc" QUERY_STRING="url=" cgit >tmp &&
+	grep ">日本語\.\.\.</a>" tmp
+'
+
 test_done
diff --git a/tests/t0103-log.sh b/tests/t0103-log.shindex ed92a76..f9612e6 100755--- a/tests/t0103-log.sh+++ b/tests/t0103-log.sh@@ -91,4 +91,14 @@ test_expect_success 'the pager of a followed log keeps the name followed' ' 	grep "href=./fol/log/b.txt?ofs=6&amp;follow=1.>\[next\]" tmp ' +test_expect_success 'a cut subject keeps whole characters' '+	git -C repos/foo commit --allow-empty -m "日本語の説明文です日本語の説明文です" &&+	{+		echo "max-message-length=20" &&+		sed "s/^cache-size=.*/cache-size=0/" cgitrc+	} >subjectrc &&+	CGIT_CONFIG="$PWD/subjectrc" QUERY_STRING="url=foo/log/" cgit >tmp &&+	grep ">日本語の説\.\.\.</a>" tmp+'+ test_done
diff --git a/tests/t0103-log.sh b/tests/t0103-log.sh
index ed92a76..f9612e6 100755
--- a/tests/t0103-log.sh
+++ b/tests/t0103-log.sh
@@ -91,4 +91,14 @@ test_expect_success 'the pager of a followed log keeps the name followed' '
 	grep "href=./fol/log/b.txt?ofs=6&amp;follow=1.>\[next\]" tmp
 '
 
+test_expect_success 'a cut subject keeps whole characters' '
+	git -C repos/foo commit --allow-empty -m "日本語の説明文です日本語の説明文です" &&
+	{
+		echo "max-message-length=20" &&
+		sed "s/^cache-size=.*/cache-size=0/" cgitrc
+	} >subjectrc &&
+	CGIT_CONFIG="$PWD/subjectrc" QUERY_STRING="url=foo/log/" cgit >tmp &&
+	grep ">日本語の説\.\.\.</a>" tmp
+'
+
 test_done