Bug 693624: Ensure that windows copes with utf8 filenames

When running under Windows, replace fopen with our own fopen_utf8
that converts from utf8 to unicode before calling the unicode
version of fopen.

3 files changed, 46 insertions, 6 deletions

diff --git a/fitz/base_string.c b/fitz/base_string.c
index 7385b500..d61f8cf3 100644
--- a/fitz/base_string.c
+++ b/fitz/base_string.c
@@ -101,7 +101,7 @@ enum
 };
 
 int
-fz_chartorune(int *rune, char *str)
+fz_chartorune(int *rune, const char *str)
 {
 	int c, c1, c2, c3;
 	long l;
@@ -110,7 +110,7 @@ fz_chartorune(int *rune, char *str)
 	 * one character sequence
 	 *	00000-0007F => T1
 	 */
-	c = *(unsigned char*)str;
+	c = *(const unsigned char*)str;
 	if(c < Tx) {
 		*rune = c;
 		return 1;
@@ -120,7 +120,7 @@ fz_chartorune(int *rune, char *str)
 	 * two character sequence
 	 *	0080-07FF => T2 Tx
 	 */
-	c1 = *(unsigned char*)(str+1) ^ Tx;
+	c1 = *(const unsigned char*)(str+1) ^ Tx;
 	if(c1 & Testx)
 		goto bad;
 	if(c < T3) {
@@ -137,7 +137,7 @@ fz_chartorune(int *rune, char *str)
 	 * three character sequence
 	 *	0800-FFFF => T3 Tx Tx
 	 */
-	c2 = *(unsigned char*)(str+2) ^ Tx;
+	c2 = *(const unsigned char*)(str+2) ^ Tx;
 	if(c2 & Testx)
 		goto bad;
 	if(c < T4) {
@@ -152,7 +152,7 @@ fz_chartorune(int *rune, char *str)
 	 * four character sequence (21-bit value)
 	 *	10000-1FFFFF => T4 Tx Tx Tx
 	 */
-	c3 = *(unsigned char*)(str+3) ^ Tx;
+	c3 = *(const unsigned char*)(str+3) ^ Tx;
 	if (c3 & Testx)
 		goto bad;
 	if (c < T5) {
diff --git a/fitz/base_time.c b/fitz/base_time.c
index 476749a7..01877a30 100644
--- a/fitz/base_time.c
+++ b/fitz/base_time.c
@@ -5,6 +5,7 @@
 #include <winsock2.h>
 #endif
 #include <windows.h>
+#include "fitz.h"
 
 #if defined(_MSC_VER) || defined(_MSC_EXTENSIONS)
 #define DELTA_EPOCH_IN_MICROSECS 11644473600000000Ui64
@@ -37,6 +38,40 @@ int gettimeofday(struct timeval *tv, struct timezone *tz)
 	return 0;
 }
 
+FILE *fopen_utf8(const char *name, const char *mode)
+{
+	wchar_t *wname, *wmode, *d;
+	const char *s;
+	int c;
+	FILE *file;
+
+	d = wname = malloc((strlen(name)+1) * sizeof(wchar_t));
+	if (d == NULL)
+		return NULL;
+	s = name;
+	while (*s) {
+		s += fz_chartorune(&c, s);
+		*d++ = c;
+	}
+	*d = 0;
+	d = wmode = malloc((strlen(mode)+1) * sizeof(wchar_t));
+	if (d == NULL)
+	{
+		free(wname);
+		return NULL;
+	}
+	s = mode;
+	while (*s) {
+		s += fz_chartorune(&c, s);
+		*d++ = c;
+	}
+	*d = 0;
+	file = _wfopen(wname, wmode);
+	free(wname);
+	free(wmode);
+	return file;
+}
+
 #else
 
 void fz_gettimeofday_dummy() { }
diff --git a/fitz/fitz.h b/fitz/fitz.h
index 29b9cee4..f915d308 100644
--- a/fitz/fitz.h
+++ b/fitz/fitz.h
@@ -59,12 +59,17 @@ int gettimeofday(struct timeval *tv, struct timezone *tz);
 #define isnan _isnan
 #define hypotf _hypotf
 
+#define fopen fopen_utf8
+
+FILE *fopen_utf8(const char *name, const char *mode);
+
 #else /* Unix or close enough */
 
 #include <unistd.h>
 
 #ifndef O_BINARY
 #define O_BINARY 0
+
 #endif
 
 #endif
@@ -628,7 +633,7 @@ int fz_strlcat(char *dst, const char *src, int n);
 
 	Returns the number of bytes consumed. Does not throw exceptions.
 */
-int fz_chartorune(int *rune, char *str);
+int fz_chartorune(int *rune, const char *str);
 
 /*
 	fz_runetochar: UTF8 encode a rune to a sequence of chars.