SF.net SVN: frontierkernel: [1338] Frontier/trunk/Common/source/strings.c

[<set...@us...>]

Revision: 1338
Author:   sethdill
Date:     2006-05-02 13:04:56 -0700 (Tue, 02 May 2006)
ViewCVS:  http://svn.sourceforge.net/frontierkernel/?rev=1338&view=rev

Log Message:
-----------
Slightly better error reporting on Windows when converting character sets. If the requested set doesn't exist, throw the same error as is seen on the mac. Also, getTextEncodingId sets an error if the id doesn't exist, so isTextEncodingAvailable now clears that error if it is returning false. (The point of the boolean function is to avoid the error.)

Modified Paths:
--------------
    Frontier/trunk/Common/source/strings.c
Modified: Frontier/trunk/Common/source/strings.c
===================================================================
--- Frontier/trunk/Common/source/strings.c	2006-05-02 16:49:55 UTC (rev 1337)
+++ Frontier/trunk/Common/source/strings.c	2006-05-02 20:04:56 UTC (rev 1338)
@@ -51,11 +51,21 @@
 	
 #define stringerrorlist 263
 
+/*
+	constants related to text-conversion functions
+*/
 #define cs_utf8			BIGSTRING( "\xA5" "utf-8" )
 #define cs_utf16		BIGSTRING( "\x06" "utf-16" )
 #define cs_iso88591		BIGSTRING( "\xA0" "iso-8859-1" )
 #define cs_macintosh	BIGSTRING( "\x09" "macintosh" )
 
+#ifdef WIN95VERSION
+	#define kTextUnsupportedEncodingErr	-8738
+	#define kTextMalformedInputErr		-8739
+	#define kTextUndefinedElementErr	-8740
+	#define kTECNoConversionPathErr		-8749
+	#define kTECPartialCharErr			-8753
+#endif
 
 byte zerostring [] = "\0"; /*use this to conserve constant space*/
 
@@ -74,8 +84,8 @@
 static boolean getTextEncodingID( bigstring, long * );
 
 #ifdef WIN95VERSION
-	static HRESULT UnicodeToAnsi(Handle, Handle, unsigned long);
-	static HRESULT AnsiToUnicode(Handle, Handle, unsigned long);
+	static boolean UnicodeToAnsi(Handle, Handle, unsigned long, long *);
+	static boolean AnsiToUnicode(Handle, Handle, unsigned long, long *);
 #endif
 
 
@@ -1724,14 +1734,17 @@
 boolean copyPStringToWide( bigstring bs, long * lenResult, wchar_t * outString )
 {
 	HRESULT err;
-	size_t lenOut, lenIn;
+	size_t lenIn;
+	bigstring bsCString;
 
 	lenIn = (size_t) bs[ 0 ];
-	convertpstring( bs );
+
+	copystring( bs, bsCString );
+	convertpstring( bsCString );
 	
-	lenOut = MultiByteToWideChar( 1252, 0, bs, lenIn, (LPWSTR) outString, 255 );
+	*lenResult = (long) MultiByteToWideChar( 1252, 0, bsCString, lenIn, (LPWSTR) outString, 255 );
 	
-	if ( lenOut == 0 )
+	if ( *lenResult == 0 )
 	{
 		err = GetLastError();
 		
@@ -1747,6 +1760,8 @@
 		}
 	}
 
+	outString[ *lenResult ] = L'\0'; /* the API doesn't seem to do this consistently */
+
 	return ( true );
 }
 #endif
@@ -2343,11 +2358,6 @@
 
 static short getTextEncodingErrorNumFromOSCode( long osStatusCode )
 {
-#ifdef WIN95VERSION
-	return 2;
-#endif
-
-#ifdef MACVERSION
 	switch ( osStatusCode )
 	{
 		case kTextUnsupportedEncodingErr:
@@ -2364,7 +2374,6 @@
 		default:
 			return 2;
 	}
-#endif
 }
 
 static void setTextEncodingConversionError( const bigstring encIn, const bigstring encOut, long osStatusCode )
@@ -2399,7 +2408,7 @@
 	sethandlesize (hresult, cbAnsi);
 
     // Convert to ANSI.
-    lentext = WideCharToMultiByte( encoding, WC_COMPOSITECHECK, (wchar_t *) *h, -1, *hresult, cbAnsi, NULL, NULL);
+    lentext = WideCharToMultiByte( encoding, 0, (wchar_t *) *h, -1, *hresult, cbAnsi, NULL, NULL);
 	
 	if ( lentext == 0 )  // report error
 	{
@@ -2424,7 +2433,6 @@
 
     ULONG cbUni, cCharacters;
 	long lentext;
-	// unsigned long ix;
 
     cCharacters = gethandlesize (h);
 
@@ -2490,7 +2498,7 @@
 
 		if ( equalstrings( bsEncodingName, BIGSTRING ("\x06" "utf-16") ) )
 		{
-			*encodingId = 0;
+			*encodingId = 1200;
 			
 			return (true);
 		}
@@ -2513,12 +2521,10 @@
 
 		copyPStringToWide( bsEncodingName, &lenIanaName, ianaName );
 
-		// pSetInfo = (PMIMECSETINFO)CoTaskMemAlloc( sizeof(MIMECSETINFO) );
-
 		err = pMultiLanguage->lpVtbl->GetCharsetInfo( pMultiLanguage, ianaName, &setInfo );
 		if ( FAILED( err ) )
 		{
-			setTextEncodingConversionError( bsEncodingName, BIGSTRING( "" ), err );
+			setTextEncodingConversionError( bsEncodingName, BIGSTRING( "" ), kTextUnsupportedEncodingErr );
 			
 			pMultiLanguage->lpVtbl->Release( pMultiLanguage );
 			
@@ -2543,7 +2549,15 @@
 	long encodingId;
 	
 	if ( ! getTextEncodingID( bsEncodingName, &encodingId ) )
+	{
+		/*
+			if the encoding isn't recognized then getTextEncodingId set an error
+			we need to clear that error
+		*/
+		langerrorclear();
+
 		return (false);
+	}
 	
 	#ifdef MACVERSION
 	
@@ -2554,7 +2568,7 @@
 	
 	#ifdef WIN95VERSION
 	
-		// we use 0 to refer to UTF-16 on Windows, as it's the default character set and has no code page
+		// we use 0 to refer to UTF-16 on Windows, as it's the default character set and seems to have no code page
 		return (true);
 	
 	#endif
@@ -2640,7 +2654,8 @@
 	
 	switch ( inputcharset ) {
 		
-		case 0: /* unicode with BOM, not handled as a code page. I Hatesk Windows. */
+		case 1200:
+		case 0: /* utf-16le, a.k.a. "unicode" on windows, not handled as a code page. */
 
 			sethandlesize (h,lentext + 2);
 
@@ -2649,15 +2664,20 @@
 
 			flResult = UnicodeToAnsi( h, hresult, outputcharset, OSStatusCode );
 
-			if ( flStatus && ( (*h) [0] == '\xFF') )
+			if ( flResult && ( (*h) [0] == '\xFF') )
 				pullfromhandle (hresult,0, 1, NULL);  /* Pop off the BOM */
 
+			break;
+
 		default: {
 
 			switch ( outputcharset ) {
-				case 0: /* unicode with BOM, not handled as a code page */
+				case 1200:
+				case 0: /* utf-16le, a.k.a. "unicode" on windows, not handled as a code page */
 
 					flResult = AnsiToUnicode( h, hresult, inputcharset, OSStatusCode );
+
+					break;
 				
 				default: {
 					Handle htemp;
@@ -2688,11 +2708,6 @@
 		
 		} /* switch inputcharset */
 	
-	/*
-	if ( ! flResult )
-		*OSStatusCode = (long) GetLastError();
-	*/
-	
 	return ( flResult );
 
 #endif
@@ -2805,14 +2820,14 @@
 
 		if ( ! AnsiToUnicode (h, htemp, CP_UTF8, &err ) ) /*Convert to UTF-16*/
 		{
-			setTextEncodingConversionError( utf8, utf16, err );
+			setTextEncodingConversionError( cs_utf8, cs_utf16, err );
 			
 			return ( false );
 		}
 
 		if ( ! UnicodeToAnsi( htemp, hresult, CP_ACP, &err ) ) /*Convert from UTF-16 to ANSI*/
 		{
-			setTextEncodingConversionError( utf16, iso88591, err);
+			setTextEncodingConversionError( cs_utf16, cs_iso88591, err);
 			
 			return ( false );
 		}


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.