Various unicode fixes made.

git-svn-id: svn+ssh://svn.gna.org/svn/gnustep/libs/base/trunk@13371 72102866-910b-0410-8b05-ffd578937521

ChangeLog

@@ -1,3 +1,9 @@
+2002-04-06  Richard Frith-Macdonald <rfm@gnu.org>
+
+	* Source/NSString.m: Fix error in unicode BOM handling
+	* Source/GSString.m: Fix error in length of unicode data produced.
+	* Tools/cvtenc.m: Made easy to use and added instructions.
+
 Sat Apr  6 02:42:43 2002  Nicola Pero  <n.pero@mi.flashnet.it>
 
 	* Source/NSUser.m (NSSearchPathForDirectoriesInDomains): Fixed

Source/GSString.m

@@ -813,7 +813,7 @@ dataUsingEncoding_c(ivars self, NSStringEncoding encoding, BOOL flag)
 	{
 	  return nil;
 	}
-      return [NSDataClass dataWithBytesNoCopy: r length: l];
+      return [NSDataClass dataWithBytesNoCopy: r length: l * sizeof(unichar)];
     }
   else
     {

Source/NSString.m

@@ -1166,12 +1166,12 @@ handle_printf_atsign (FILE *stream,
 
 	  b = (unsigned char*)[data bytes];
 	  uptr = (unichar*)b;
-	  if (*uptr == 0xFFFE)
+	  if (*uptr == byteOrderMark)
 	    {
 	      b = (unsigned char*)++uptr;
 	      len -= sizeof(unichar);
 	    }
-	  else if (*uptr == 0xFEFF)
+	  else if (*uptr == byteOrderMarkSwapped)
 	    {
 	      b = (unsigned char*)++uptr;
 	      len -= sizeof(unichar);
@@ -2376,7 +2376,7 @@ handle_printf_atsign (FILE *stream,
 
       buff = (unichar*)NSZoneMalloc(NSDefaultMallocZone(),
 	sizeof(unichar)*(len+1));
-      buff[0] = 0xFEFF;
+      buff[0] = byteOrderMark;
       [self getCharacters: &buff[1]];
       return [NSDataClass dataWithBytesNoCopy: buff
 					length: sizeof(unichar)*(len+1)];

Tools/cvtenc.m

@@ -38,6 +38,9 @@ main(int argc, char** argv, char **env)
   NSProcessInfo		*proc;
   NSArray		*args;
   unsigned		i;
+  BOOL			found = NO;
+  NSString		*n;
+  NSStringEncoding	enc = 0;
 
 #ifdef GS_PASS_ARGUMENTS
   [NSProcessInfo initializeWithArguments: argv count: argc environment: env];
@@ -53,15 +56,6 @@ main(int argc, char** argv, char **env)
 
   args = [proc arguments];
 
-  if ([args count] <= 1)
-    {
-      NSLog(@"No file names given to convert.");
-    }
-  else
-    {
-      NSString		*n;
-      NSStringEncoding	enc = 0;
-
   n = [[NSUserDefaults standardUserDefaults] stringForKey: @"Encoding"];
   if (n == nil)
     {
@@ -95,7 +89,7 @@ main(int argc, char** argv, char **env)
 	}
     }
 
-      for (i = 1; i < [args count]; i++)
+  for (i = 1; found == NO && i < [args count]; i++)
     {
       NSString	*file = [args objectAtIndex: i];
 
@@ -104,17 +98,51 @@ main(int argc, char** argv, char **env)
 	  i++;
 	  continue;
 	}
+      found = YES;
       NS_DURING
 	{
 	  NSData	*myData;
-	      NSString	*myString;
 
-	      myString = [NSString stringWithContentsOfFile: file];
+	  myData = [[NSData alloc] initWithContentsOfFile: file];
-	      myData = [myString dataUsingEncoding: enc
-			      allowLossyConversion: NO];
 	  if (myData == nil)
 	    {
-		  NSLog(@"Encoding conversion failed.", file);
+	      NSLog(@"File read operation failed for %@.", file);
+	    }
+	  else
+	    {
+	      unsigned		l = [myData length];
+	      const unichar		*b = (const unichar*)[myData bytes];
+	      NSStringEncoding	iEnc;
+	      NSStringEncoding	oEnc;
+	      NSString		*myString;
+
+	      if (l > 1 && (*b == 0xFFFE || *b == 0xFEFF))
+		{
+		  iEnc = NSUnicodeStringEncoding;
+		  oEnc = enc;
+		}
+	      else
+		{
+		  iEnc = enc;
+		  oEnc = NSUnicodeStringEncoding;
+		}
+	      
+	      myString = [[NSString alloc] initWithData: myData
+					       encoding: iEnc];
+	      RELEASE(myData);
+	      if (myString == nil)
+		{
+		  NSLog(@"Encoding input conversion failed for %@.", file);
+		}
+	      else
+		{
+		  myData = [myString dataUsingEncoding: oEnc
+				  allowLossyConversion: NO];
+		  RELEASE(myString);
+		  if (myData == nil)
+		    {
+		      NSLog(@"Encoding output conversion failed for %@.",
+			file);
 		    }
 		  else
 		    {
@@ -125,12 +153,22 @@ main(int argc, char** argv, char **env)
 		      [out synchronizeFile];
 		    }
 		}
+	    }
+	}
       NS_HANDLER
 	{
 	  NSLog(@"Converting '%@' - %@", file, [localException reason]);
 	}
       NS_ENDHANDLER
     }
+
+  if (found == NO)
+    {
+      NSLog(@"\nThis utility expects a filename as an argument.\n"
+	@"It reads the file, and writes it to STDOUT after converting it\n"
+	@"to unicode from C string encoding or vice versa.\n"
+	@"You can supply a '-Encoding name' option to specify the C string\n"
+	@"encoding to be used, if you don't want to use the default.");
     }
   [pool release];
   return 0;