More UTF-8 in paths, from Alexey Kovalenko.

Change-Id: Ic5b75d640cd42fc4c23a30c9f2a721a95471df69
grke · May 27, 2017 · 9131186 · 9131186
1 parent f2eb4cf
commit 9131186
Show file tree

Hide file tree

Showing 8 changed files with 69 additions and 43 deletions.
diff --git a/src/burp.h b/src/burp.h
@@ -77,12 +77,12 @@
 // Local Burp includes. Be sure to put all the system includes before these.
 #ifdef HAVE_WIN32
 	#include <windows.h>
-	#include "win32/compat/compat.h"
 #endif
 
 #include "burpconfig.h"
 
 #ifdef HAVE_WIN32
+	#include "win32/compat/compat.h"
 	#include "win32/winapi.h"
 	#include "winhost.h"
 #endif

diff --git a/src/burpconfig.h b/src/burpconfig.h
@@ -6,7 +6,36 @@
 
 #define ASSERT(x)
 
+// MAX_PATH is Windows constatnt, usually 260, maybe changed in some Win10 update.
+// PATH_MAX is Posix constant
+// right method - it to call pathconf(), but such case lead to reworking lots of code;
+// and some buffers are better to allocate on stack, not on heap
+#ifndef MAX_PATH
+    #ifdef PATH_MAX
+        #define MAX_PATH           PATH_MAX
+    #else
+        #define MAX_PATH           260
+    #endif
+#endif
+
+// unicode enabling of win 32 needs some defines and functions
+
+// using an average of 3 bytes per character is probably fine in
+// practice but I believe that Windows actually uses UTF-16 encoding
+// as opposed to UCS2 which means characters 0x10000-0x10ffff are
+// valid and result in 4 byte UTF-8 encodings.
+#define MAX_PATH_UTF8    MAX_PATH*4  // strict upper bound on UTF-16 to UTF-8 conversion
+
+// from
+// http://msdn.microsoft.com/library/default.asp?url=/library/en-us/fileio/fs/getfileattributesex.asp
+// In the ANSI version of this function, the name is limited to
+// MAX_PATH characters. To extend this limit to 32,767 wide
+// characters, call the Unicode version of the function and prepend
+// "\\?\" to the path. For more information, see Naming a File.
+#define MAX_PATH_W 32767
+
 #ifdef HAVE_WIN32
+
 	#define WIN32_REPARSE_POINT  1 // Any odd dir except the next two.
 	#define WIN32_MOUNT_POINT    2 // Directory link to Volume.
 	#define WIN32_JUNCTION_POINT 3 // Directory link to a directory.
@@ -24,10 +53,6 @@
 	#endif
 #endif
 
-#ifndef S_ISLNK
-#define S_ISLNK(m) (((m) & S_IFM) == S_IFLNK)
-#endif
-
 #ifndef O_BINARY
 	#define O_BINARY 0
 #endif

diff --git a/src/server/protocol1/backup_phase2.c b/src/server/protocol1/backup_phase2.c
@@ -32,7 +32,10 @@ static int path_length_warn(struct iobuf *path, struct conf **cconfs)
 
 static int path_too_long(struct iobuf *path, struct conf **cconfs)
 {
-	static const char *cp;
+	const char *cp;
+	const char *cp1;
+	size_t len;
+
 	if(treepathlen+path->len+1>fs_full_path_max)
 	{
 		// FIX THIS:
@@ -43,9 +46,21 @@ static int path_too_long(struct iobuf *path, struct conf **cconfs)
 		// to be able to fix it.
 		return path_length_warn(path, cconfs);
 	}
-	if((cp=strrchr(path->buf, '/'))) cp++;
-	else cp=path->buf;
-	if(strlen(cp)>fs_name_max) return path_length_warn(path, cconfs);
+	// We have to check every part in the path to ensure it less then fs_name_max
+	// minimum windows case is c:/a, nix case is /
+	// usual: c:/users, /home
+	cp = strchr(path->buf, '/');
+	if( !cp ) // very strange
+	    return strlen(path->buf) > fs_name_max ?  path_length_warn(path, cconfs):0;
+	while(cp && *cp)
+	{
+	    cp++;
+	    cp1 = strchr(cp, '/');
+	    len = cp1? cp1-cp : strlen(cp);
+	    if( len > fs_name_max )
+		return path_length_warn(path, cconfs);
+	    cp = cp1;
+	}
 	return 0;
 }
 

diff --git a/src/server/protocol1/bedup.c b/src/server/protocol1/bedup.c
@@ -605,7 +605,7 @@ static int process_dir(const char *oldpath, const char *newpath,
 	}
 	ret=0;
 end:
-	closedir(dirp);
+	if(dirp) closedir(dirp);
 	free_w(&newfile.path);
 	free_w(&path);
 	return ret;

diff --git a/src/win32/Makefile b/src/win32/Makefile
@@ -61,8 +61,8 @@ $(DIRS):
 			exit 1 ; \
 		fi ; \
 	done ; \
-	ln -vsfT $${TOOLSDIR}/burp-cross-tools ../../burp-cross-tools ; \
-	ln -vsfT $${TOOLSDIR}/burp-depkgs ../../burp-depkgs
+	ln -vsf $${TOOLSDIR}/burp-cross-tools ../../burp-cross-tools ; \
+	ln -vsf $${TOOLSDIR}/burp-depkgs ../../burp-depkgs
 
 Makefile.inc: ../../burp-cross-tools
 	@echo Creating $@

diff --git a/src/win32/compat/compat.cpp b/src/win32/compat/compat.cpp
@@ -41,8 +41,6 @@
 #include "mem_pool.h"
 #include "berrno.h"
 
-#define MAX_PATHLENGTH  1024
-
 /* UTF-8 to UCS2 path conversion is expensive,
    so we cache the conversion. During backup the
    conversion is called 3 times (lstat, attribs, open),
@@ -1206,6 +1204,7 @@ DIR *opendir(const char *path)
 
 int closedir(DIR *dirp)
 {
+	if(!dirp) return 0;
 	_dir *dp=(_dir *)dirp;
 	FindClose(dp->dirh);
 	_dir_free(dp);
@@ -1215,7 +1214,7 @@ int closedir(DIR *dirp)
 static void copyin(struct dirent *entry, const char *fname)
 {
 	char *cp=entry->d_name;
-	while(*fname)
+	while( *fname && entry->d_reclen < (MAX_PATH_UTF8-1) )
 	{
 		*cp++=*fname++;
 		entry->d_reclen++;
@@ -1312,9 +1311,9 @@ long pathconf(const char *path, int name)
 	{
 		case _PC_PATH_MAX:
 			if(!strncmp(path, "\\\\?\\", 4))
-				return 32767;
+				return MAX_PATH_W;
 		case _PC_NAME_MAX:
-			return 255;
+			return MAX_PATH;
 	}
 	return system_error();
 }
@@ -1783,7 +1782,7 @@ const char * getArgv0(const char *cmdline)
 
 static int dwAltNameLength_ok(DWORD dwAltNameLength)
 {
-	return dwAltNameLength>0 && dwAltNameLength<MAX_PATHLENGTH;
+	return dwAltNameLength>0 && dwAltNameLength<MAX_PATH_UTF8;
 }
 
 /* Extracts the executable or script name from the first string in 
@@ -1883,11 +1882,11 @@ bool GetApplicationName(const char *cmdline, char **pexe, const char **pargs)
 	DWORD dwBasePathLength=pExeEnd-pExeStart;
 
 	DWORD dwAltNameLength=0;
-	char *pPathname=(char *)alloca(MAX_PATHLENGTH+1);
-	char *pAltPathname=(char *)alloca(MAX_PATHLENGTH+1);
+	char *pPathname=(char *)alloca(MAX_PATH_UTF8+1);
+	char *pAltPathname=(char *)alloca(MAX_PATH_UTF8+1);
 
-	pPathname[MAX_PATHLENGTH]='\0';
-	pAltPathname[MAX_PATHLENGTH]='\0';
+	pPathname[MAX_PATH_UTF8]='\0';
+	pAltPathname[MAX_PATH_UTF8]='\0';
 
 	memcpy(pPathname, pExeStart, dwBasePathLength);
 	pPathname[dwBasePathLength]='\0';
@@ -1904,7 +1903,7 @@ bool GetApplicationName(const char *cmdline, char **pexe, const char **pargs)
 				// There are no path separators, search in the
 				// standard locations
 				dwAltNameLength=SearchPath(NULL, pPathname,
-				  ExtensionList[index], MAX_PATHLENGTH,
+				  ExtensionList[index], MAX_PATH_UTF8,
 				  pAltPathname, NULL);
 				if(dwAltNameLength_ok(dwAltNameLength))
 				{
@@ -1917,7 +1916,7 @@ bool GetApplicationName(const char *cmdline, char **pexe, const char **pargs)
 			else
 			{
 				snprintf(&pPathname[dwBasePathLength],
-					MAX_PATHLENGTH-dwBasePathLength,
+					MAX_PATH_UTF8-dwBasePathLength,
 					"%s", ExtensionList[index]);
 				if(GetFileAttributes(pPathname)
 				  !=INVALID_FILE_ATTRIBUTES)
@@ -1931,7 +1930,7 @@ bool GetApplicationName(const char *cmdline, char **pexe, const char **pargs)
 		// There are no path separators, search in the standard
 		// locations.
 		dwAltNameLength=SearchPath(NULL, pPathname,
-			NULL, MAX_PATHLENGTH, pAltPathname, NULL);
+			NULL, MAX_PATH_UTF8, pAltPathname, NULL);
 		if(dwAltNameLength_ok(dwAltNameLength))
 		{
 			memcpy(pPathname, pAltPathname, dwAltNameLength);
@@ -1942,7 +1941,7 @@ bool GetApplicationName(const char *cmdline, char **pexe, const char **pargs)
 	if(strchr(pPathname, ' '))
 	{
 		dwAltNameLength=GetShortPathName(pPathname,
-			pAltPathname, MAX_PATHLENGTH);
+			pAltPathname, MAX_PATH_UTF8);
 
 		if(dwAltNameLength_ok(dwAltNameLength))
 		{

diff --git a/src/win32/compat/compat.h b/src/win32/compat/compat.h
@@ -137,10 +137,7 @@ struct dirent
 	uint64_t d_ino;
 	uint32_t d_off;
 	uint16_t d_reclen;
-	// MAX_PATH_UTF8 is MAX_PATH*4, but I have had problems trying to
-	// include winapi.h so that it is seen here. Do it manually here, and
-	// add a bit more just in case.
-	char d_name[(MAX_PATH*4)+16];
+	char d_name[MAX_PATH_UTF8];
 };
 typedef void DIR;
 

diff --git a/src/win32/winapi.h b/src/win32/winapi.h
@@ -41,19 +41,9 @@
 
 // unicode enabling of win 32 needs some defines and functions
 
-// using an average of 3 bytes per character is probably fine in
-// practice but I believe that Windows actually uses UTF-16 encoding
-// as opposed to UCS2 which means characters 0x10000-0x10ffff are
-// valid and result in 4 byte UTF-8 encodings.
-#define MAX_PATH_UTF8    MAX_PATH*4  // strict upper bound on UTF-16 to UTF-8 conversion
-
-// from
-// http://msdn.microsoft.com/library/default.asp?url=/library/en-us/fileio/fs/getfileattributesex.asp
-// In the ANSI version of this function, the name is limited to
-// MAX_PATH characters. To extend this limit to 32,767 wide
-// characters, call the Unicode version of the function and prepend
-// "\\?\" to the path. For more information, see Naming a File.
-#define MAX_PATH_W 32767
+// MAX_PATH_UTF8 moved to burpconfig.h
+// winapi.h used only in burp.h just after using burpconfig.h
+#include "burpconfig.h"
 
 int wchar_2_UTF8(char *pszUTF, const WCHAR *pszUCS, int cchChar = MAX_PATH_UTF8);
 int UTF8_2_wchar(char **pszUCS, const char *pszUTF);