From 68c33866cd283542b8962bdfec27e5c4a26bffd5 Mon Sep 17 00:00:00 2001
From: df <fieldhouse@gmx.net>
Date: Thu, 14 Jan 2021 23:55:08 +0000
Subject: [PATCH 1/3] Add more decoding of XML character entities

---
 tvdb.c | 28 +++++++++++++++++-----------
 1 file changed, 17 insertions(+), 11 deletions(-)

diff --git a/tvdb.c b/tvdb.c
index 7e68e56..5210bed 100644
--- a/tvdb.c
+++ b/tvdb.c
@@ -48,26 +48,32 @@ struct episode {
 void
 unescape(char *txt)
 {
-	char *p = txt;
 	int l = strlen(txt);
+	char *p = txt;
 
-	while ((p = strchr(p, '&')))
+	for (; p = strchr(p, '&'); p++)
 	{
-		HANDLE("&quot;", '"', 5);
-		HANDLE("&amp;", '&', 4);
-		HANDLE("&#xD;", '\n', 4);
-		HANDLE("&#xA;", '\r', 4);
-		p++;
+		size_t ll;
+		unsigned char icode;
+		if (1 == sscanf( p, "&#%hhu;%n", &icode, &ll) ||
+		    1 == sscanf( p, "&#%*[xX]%hhx;%n", &icode, &ll)) {
+			/*  &#x<hex>;, &#<decimal>; */
+			HANDLE(p, (char)icode, ll-1);
+		} else {
+			HANDLE("&amp;", '&', 4);
+			HANDLE("&quot;", '"', 5);
+			HANDLE("&apos;", '\'', 5);
+			HANDLE("&lt;", '<', 3);
+			HANDLE("&gt;", '>', 3);
+		}
 	}
 
-	p = txt;
-	while ((p = memchr(p, '\xe2', l - (p - txt))))
-	{
+	for (p = txt; p = memchr(p, '\xe2', l - (p - txt)); p++)
+	{	/* curly apostrophe, en dash, curly quotes */
 		HANDLE("\xe2\x80\x99", '\'', 2);
 		HANDLE("\xe2\x80\x93", '-', 2);
 		HANDLE("\xe2\x80\x9c", '"', 2);
 		HANDLE("\xe2\x80\x9d", '"', 2);
-		p++;
 	}
 
 	if ((p = strpbrk(txt, "\n\r")))
-- 
2.46.2


From 5aeab9da9d1428ee484351f048573c28218c7b22 Mon Sep 17 00:00:00 2001
From: df <fieldhouse@gmx.net>
Date: Fri, 15 Jan 2021 00:31:20 +0000
Subject: [PATCH 2/3] Build without warnings on-box

---
 Makefile | 2 +-
 tvdb.c   | 8 ++++----
 2 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/Makefile b/Makefile
index aebbbe4..b04f625 100644
--- a/Makefile
+++ b/Makefile
@@ -9,7 +9,7 @@ HDRS=
 OBJS=	$(SRCS:.c=.o)
 CC=gcc
 #CC=mipsel-linux-gcc
-CFLAGS=-g
+CFLAGS=-g -std=c99 -D_XOPEN_SOURCE=700
 INCS=
 LIBS=-lsqlite3
 WARN=-pedantic -Wall -W -Wnested-externs -Wpointer-arith -Wno-long-long
diff --git a/tvdb.c b/tvdb.c
index 5210bed..53bb8a5 100644
--- a/tvdb.c
+++ b/tvdb.c
@@ -48,12 +48,12 @@ struct episode {
 void
 unescape(char *txt)
 {
-	int l = strlen(txt);
 	char *p = txt;
+	int l = strlen(txt);
 
-	for (; p = strchr(p, '&'); p++)
+	for (; (p = strchr(p, '&')); p++)
 	{
-		size_t ll;
+		int ll;
 		unsigned char icode;
 		if (1 == sscanf( p, "&#%hhu;%n", &icode, &ll) ||
 		    1 == sscanf( p, "&#%*[xX]%hhx;%n", &icode, &ll)) {
@@ -68,7 +68,7 @@ unescape(char *txt)
 		}
 	}
 
-	for (p = txt; p = memchr(p, '\xe2', l - (p - txt)); p++)
+	for (p = txt; (p = memchr(p, '\xe2', l - (p - txt))); p++)
 	{	/* curly apostrophe, en dash, curly quotes */
 		HANDLE("\xe2\x80\x99", '\'', 2);
 		HANDLE("\xe2\x80\x93", '-', 2);
-- 
2.46.2


From cf86aaebca5d69b8d6f87cb0be6ea710bf5b94f7 Mon Sep 17 00:00:00 2001
From: df <fieldhouse@gmx.net>
Date: Fri, 15 Jan 2021 00:58:18 +0000
Subject: [PATCH 3/3] Ensure ; immediately follows the code in &#<hex/dec>;

---
 tvdb.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/tvdb.c b/tvdb.c
index 53bb8a5..ab3922c 100644
--- a/tvdb.c
+++ b/tvdb.c
@@ -53,10 +53,12 @@ unescape(char *txt)
 
 	for (; (p = strchr(p, '&')); p++)
 	{
-		int ll;
+		int ll = 0;
 		unsigned char icode;
-		if (1 == sscanf( p, "&#%hhu;%n", &icode, &ll) ||
-		    1 == sscanf( p, "&#%*[xX]%hhx;%n", &icode, &ll)) {
+		/* sscanf -> 1: the code was read; ll>0: ';' came next */
+		if ((1 == sscanf( p, "&#%hhu;%n", &icode, &ll) ||
+		     1 == sscanf( p, "&#%*[xX]%hhx;%n", &icode, &ll)) && 
+		     ll > 0) {
 			/*  &#x<hex>;, &#<decimal>; */
 			HANDLE(p, (char)icode, ll-1);
 		} else {
-- 
2.46.2