From 5176b4a17ec9ca5151f1b8b503b2f237f3c83b27 Mon Sep 17 00:00:00 2001
From: John Cupitt <jcupitt@gmail.com>
Date: Thu, 22 Mar 2018 12:08:39 +0000
Subject: [PATCH] better header sniffing for small files

vips__get_bytes() used to fail if the file was too small for the
allocated buffer, which was a problem for svg, since files can be extremely
small.

This change makes vips__get_bytes() return the number of bytes read, so
the is_a testers can work on files smaller than the max header size.
---
 ChangeLog                      |  1 +
 libvips/foreign/gifload.c      |  2 +-
 libvips/foreign/jpeg2vips.c    |  2 +-
 libvips/foreign/matlab.c       |  2 +-
 libvips/foreign/openexr2vips.c |  2 +-
 libvips/foreign/pdfload.c      |  2 +-
 libvips/foreign/ppm.c          |  2 +-
 libvips/foreign/svgload.c      |  5 +++--
 libvips/foreign/tiff2vips.c    |  2 +-
 libvips/foreign/vipspng.c      |  2 +-
 libvips/foreign/webp2vips.c    |  2 +-
 libvips/include/vips/util.h    |  3 ++-
 libvips/iofuncs/util.c         | 27 +++++++++++----------------
 libvips/iofuncs/vips.c         |  4 ++--
 14 files changed, 28 insertions(+), 30 deletions(-)

diff --git a/ChangeLog b/ChangeLog
index 7cde3231..2f015aed 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -8,6 +8,7 @@
 - add vips_sobel() and vips_canny() edge detectors
 - add vips_rotate() ... a convenience method for vips_similarity()
 - svgload was missing is_a [lovell]
+- better header sniffing for small files
 
 12/3/18 started 8.6.4
 - better fitting of fonts with overhanging edges, thanks Adrià 
diff --git a/libvips/foreign/gifload.c b/libvips/foreign/gifload.c
index c43f9a05..4860a66e 100644
--- a/libvips/foreign/gifload.c
+++ b/libvips/foreign/gifload.c
@@ -383,7 +383,7 @@ vips_foreign_load_gif_is_a( const char *filename )
 {
 	unsigned char buf[4];
 
-	if( vips__get_bytes( filename, buf, 4 ) &&
+	if( vips__get_bytes( filename, buf, 4 ) == 4 &&
 		vips_foreign_load_gif_is_a_buffer( buf, 4 ) )
 		return( 1 );
 
diff --git a/libvips/foreign/jpeg2vips.c b/libvips/foreign/jpeg2vips.c
index 4acd6959..ceed5918 100644
--- a/libvips/foreign/jpeg2vips.c
+++ b/libvips/foreign/jpeg2vips.c
@@ -968,7 +968,7 @@ vips__isjpeg( const char *filename )
 {
 	unsigned char buf[2];
 
-	if( vips__get_bytes( filename, buf, 2 ) &&
+	if( vips__get_bytes( filename, buf, 2 ) == 2 &&
 		vips__isjpeg_buffer( buf, 2 ) )
 		return( 1 );
 
diff --git a/libvips/foreign/matlab.c b/libvips/foreign/matlab.c
index b2d6a01a..5335647b 100644
--- a/libvips/foreign/matlab.c
+++ b/libvips/foreign/matlab.c
@@ -323,7 +323,7 @@ vips__mat_ismat( const char *filename )
 {
 	unsigned char buf[15];
 
-	if( vips__get_bytes( filename, buf, 10 ) &&
+	if( vips__get_bytes( filename, buf, 10 ) == 10 &&
 		vips_isprefix( "MATLAB 5.0", (char *) buf ) )
 		return( 1 );
 
diff --git a/libvips/foreign/openexr2vips.c b/libvips/foreign/openexr2vips.c
index 7d472561..8e113634 100644
--- a/libvips/foreign/openexr2vips.c
+++ b/libvips/foreign/openexr2vips.c
@@ -107,7 +107,7 @@ vips__openexr_isexr( const char *filename )
 {
 	unsigned char buf[4];
 
-	if( vips__get_bytes( filename, buf, 4 ) )
+	if( vips__get_bytes( filename, buf, 4 ) == 4 )
 		if( buf[0] == 0x76 && buf[1] == 0x2f &&
 			buf[2] == 0x31 && buf[3] == 0x01 )
 			return( TRUE );
diff --git a/libvips/foreign/pdfload.c b/libvips/foreign/pdfload.c
index 289c937d..89f1c960 100644
--- a/libvips/foreign/pdfload.c
+++ b/libvips/foreign/pdfload.c
@@ -165,7 +165,7 @@ vips_foreign_load_pdf_is_a( const char *filename )
 {
 	unsigned char buf[4];
 
-	if( vips__get_bytes( filename, buf, 4 ) &&
+	if( vips__get_bytes( filename, buf, 4 ) == 4 &&
 		vips_foreign_load_pdf_is_a_buffer( buf, 4 ) )
 		return( 1 );
 
diff --git a/libvips/foreign/ppm.c b/libvips/foreign/ppm.c
index 9a7677f5..a5022e78 100644
--- a/libvips/foreign/ppm.c
+++ b/libvips/foreign/ppm.c
@@ -514,7 +514,7 @@ vips__ppm_isppm( const char *filename )
 {
 	VipsPel buf[3];
 
-	if( vips__get_bytes( filename, buf, 2 ) ) {
+	if( vips__get_bytes( filename, buf, 2 ) == 2 ) {
 		int i;
 
 		buf[2] = '\0';
diff --git a/libvips/foreign/svgload.c b/libvips/foreign/svgload.c
index 279ebc1e..fd4b9776 100644
--- a/libvips/foreign/svgload.c
+++ b/libvips/foreign/svgload.c
@@ -458,9 +458,10 @@ static gboolean
 vips_foreign_load_svg_file_is_a( const char *filename )
 {
 	unsigned char buf[300];
+	guint64 bytes;
 
-	return( vips__get_bytes( filename, buf, 300 ) &&
-		vips_foreign_load_svg_is_a( buf, 300 ) );
+	return( (bytes = vips__get_bytes( filename, buf, 300 )) > 0 &&
+		vips_foreign_load_svg_is_a( buf, bytes ) );
 }
 
 static int
diff --git a/libvips/foreign/tiff2vips.c b/libvips/foreign/tiff2vips.c
index ced03af5..49220307 100644
--- a/libvips/foreign/tiff2vips.c
+++ b/libvips/foreign/tiff2vips.c
@@ -2362,7 +2362,7 @@ vips__istiff( const char *filename )
 {
 	unsigned char buf[4];
 
-	if( vips__get_bytes( filename, buf, 4 ) &&
+	if( vips__get_bytes( filename, buf, 4 ) == 4 &&
 		vips__istiff_buffer( buf, 4 ) )
 		return( TRUE );
 
diff --git a/libvips/foreign/vipspng.c b/libvips/foreign/vipspng.c
index 65958eb9..b645da3a 100644
--- a/libvips/foreign/vipspng.c
+++ b/libvips/foreign/vipspng.c
@@ -677,7 +677,7 @@ vips__png_ispng( const char *filename )
 {
 	unsigned char buf[8];
 
-	return( vips__get_bytes( filename, buf, 8 ) &&
+	return( vips__get_bytes( filename, buf, 8 ) == 8 &&
 		vips__png_ispng_buffer( buf, 8 ) ); 
 }
 
diff --git a/libvips/foreign/webp2vips.c b/libvips/foreign/webp2vips.c
index 971f9238..b14a61b2 100644
--- a/libvips/foreign/webp2vips.c
+++ b/libvips/foreign/webp2vips.c
@@ -119,7 +119,7 @@ vips__iswebp( const char *filename )
 	 */
 	unsigned char header[12];
 
-	if( vips__get_bytes( filename, header, 12 ) &&
+	if( vips__get_bytes( filename, header, 12 ) == 12 &&
 		vips__iswebp_buffer( header, 12 ) )
 		return( 1 );
 
diff --git a/libvips/include/vips/util.h b/libvips/include/vips/util.h
index aa693183..f481e943 100644
--- a/libvips/include/vips/util.h
+++ b/libvips/include/vips/util.h
@@ -266,7 +266,8 @@ char *vips__file_read( FILE *fp, const char *name, size_t *length_out );
 char *vips__file_read_name( const char *name, const char *fallback_dir, 
 	size_t *length_out );
 int vips__file_write( void *data, size_t size, size_t nmemb, FILE *stream );
-int vips__get_bytes( const char *filename, unsigned char buf[], int len );
+guint64 vips__get_bytes( const char *filename, 
+	unsigned char buf[], guint64 len );
 int vips__fgetc( FILE *fp );
 
 GValue *vips__gvalue_ref_string_new( const char *text );
diff --git a/libvips/iofuncs/util.c b/libvips/iofuncs/util.c
index e22dca85..871a0af9 100644
--- a/libvips/iofuncs/util.c
+++ b/libvips/iofuncs/util.c
@@ -844,33 +844,28 @@ vips__file_write( void *data, size_t size, size_t nmemb, FILE *stream )
 	return( 0 );
 }
 
-/* Read a few bytes from the start of a file. For sniffing file types.
- * Filename may contain a mode. 
+/* Read a few bytes from the start of a file. This is used for sniffing file 
+ * types, so we must read binary. 
+ *
+ * Return the number of bytes actually read (the file might be shorter than
+ * len), or 0 for error.
  */
-int
-vips__get_bytes( const char *filename, unsigned char buf[], int len )
+guint64
+vips__get_bytes( const char *filename, unsigned char buf[], guint64 len )
 {
-	char name[FILENAME_MAX];
-	char mode[FILENAME_MAX];
 	int fd;
-
-	/* Split off the mode part.
-	 */
-	im_filename_split( filename, name, mode );
+	guint64 bytes_read;
 
 	/* File may not even exist (for tmp images for example!)
 	 * so no hasty messages. And the file might be truncated, so no error
 	 * on read either.
 	 */
-	if( (fd = vips__open_read( name )) == -1 )
+	if( (fd = vips__open_read( filename )) == -1 )
 		return( 0 );
-	if( read( fd, buf, len ) != len ) {
-		close( fd );
-		return( 0 );
-	}
+	bytes_read = read( fd, buf, len );
 	close( fd );
 
-	return( 1 );
+	return( bytes_read );
 }
 
 /* We try to support stupid DOS files too. These have \r\n (13, 10) as line
diff --git a/libvips/iofuncs/vips.c b/libvips/iofuncs/vips.c
index 50ed9d58..49e6614f 100644
--- a/libvips/iofuncs/vips.c
+++ b/libvips/iofuncs/vips.c
@@ -255,9 +255,9 @@ vips__file_magic( const char *filename )
 {
 	guint32 magic;
 
-	if( vips__get_bytes( filename, (unsigned char *) &magic, 4 ) &&
+	if( vips__get_bytes( filename, (unsigned char *) &magic, 4 ) == 4 &&
 		(magic == VIPS_MAGIC_INTEL || 
-		 magic == VIPS_MAGIC_SPARC ) )
+		 magic == VIPS_MAGIC_SPARC) )
 		return( magic );
 
 	return( 0 );