Convert T61String-labeled ISO-8859-1 to UTF-8: bug 53133 r=nelsonb a=wtc

4 files changed, 161 insertions, 29 deletions

diff --git a/security/nss/lib/certdb/secname.c b/security/nss/lib/certdb/secname.c
index 5a40249da..d1abab4fa 100644
--- a/security/nss/lib/certdb/secname.c
+++ b/security/nss/lib/certdb/secname.c
@@ -17,6 +17,7 @@
  * Rights Reserved.
  * 
  * Contributor(s):
+ *  John Gardiner Myers <jgmyers@speakeasy.net>
  * 
  * Alternatively, the contents of this file may be used under the
  * terms of the GNU General Public License Version 2 or later (the
@@ -603,8 +604,7 @@ CERT_DecodeAVAValue(const SECItem *derAVAValue)
 {
           SECItem          *retItem; 
     const SEC_ASN1Template *theTemplate       = NULL;
-          PRBool            convertUCS4toUTF8 = PR_FALSE;
-          PRBool            convertUCS2toUTF8 = PR_FALSE;
+          enum { conv_none, conv_ucs4, conv_ucs2, conv_iso88591 } convert = conv_none;
           SECItem           avaValue          = {siBuffer, 0}; 
           PLArenaPool      *newarena          = NULL;
 
@@ -614,7 +614,7 @@ CERT_DecodeAVAValue(const SECItem *derAVAValue)
 
     switch(derAVAValue->data[0]) {
 	case SEC_ASN1_UNIVERSAL_STRING:
-	    convertUCS4toUTF8 = PR_TRUE;
+	    convert = conv_ucs4;
 	    theTemplate = SEC_UniversalStringTemplate;
 	    break;
 	case SEC_ASN1_IA5_STRING:
@@ -624,10 +624,15 @@ CERT_DecodeAVAValue(const SECItem *derAVAValue)
 	    theTemplate = SEC_PrintableStringTemplate;
 	    break;
 	case SEC_ASN1_T61_STRING:
+	    /*
+	     * Per common practice, we're not decoding actual T.61, but instead
+	     * treating T61-labeled strings as containing ISO-8859-1.
+	     */
+	    convert = conv_iso88591;
 	    theTemplate = SEC_T61StringTemplate;
 	    break;
 	case SEC_ASN1_BMP_STRING:
-	    convertUCS2toUTF8 = PR_TRUE;
+	    convert = conv_ucs2;
 	    theTemplate = SEC_BMPStringTemplate;
 	    break;
 	case SEC_ASN1_UTF8_STRING:
@@ -649,36 +654,41 @@ CERT_DecodeAVAValue(const SECItem *derAVAValue)
 	return NULL;
     }
 
-    if (convertUCS4toUTF8) {
+    if (convert != conv_none) {
 	unsigned int   utf8ValLen = avaValue.len * 3;
 	unsigned char *utf8Val    = (unsigned char*)
 				    PORT_ArenaZAlloc(newarena, utf8ValLen);
 
-	if(avaValue.len % 4 != 0 ||
-	   !PORT_UCS4_UTF8Conversion(PR_FALSE, avaValue.data, avaValue.len,
-				     utf8Val, utf8ValLen, &utf8ValLen)) {
-            PORT_FreeArena(newarena, PR_FALSE);
-	    PORT_SetError(SEC_ERROR_INVALID_AVA);
-	    return NULL;
+        switch (convert) {
+        case conv_ucs4:
+           if(avaValue.len % 4 != 0 ||
+              !PORT_UCS4_UTF8Conversion(PR_FALSE, avaValue.data, avaValue.len,
+					utf8Val, utf8ValLen, &utf8ValLen)) {
+                PORT_FreeArena(newarena, PR_FALSE);
+                PORT_SetError(SEC_ERROR_INVALID_AVA);
+		return NULL;
+	   }
+	   break;
+	case conv_ucs2:
+           if(avaValue.len % 2 != 0 ||
+              !PORT_UCS2_UTF8Conversion(PR_FALSE, avaValue.data, avaValue.len,
+					utf8Val, utf8ValLen, &utf8ValLen)) {
+                PORT_FreeArena(newarena, PR_FALSE);
+                PORT_SetError(SEC_ERROR_INVALID_AVA);
+		return NULL;
+	   }
+	   break;
+	case conv_iso88591:
+           if(!PORT_ISO88591_UTF8Conversion(avaValue.data, avaValue.len,
+					utf8Val, utf8ValLen, &utf8ValLen)) {
+                PORT_FreeArena(newarena, PR_FALSE);
+                PORT_SetError(SEC_ERROR_INVALID_AVA);
+		return NULL;
+	   }
+	   break;
+	case conv_none: ;
 	}
-
-	avaValue.data = utf8Val;
-	avaValue.len = utf8ValLen;
-
-    } else if (convertUCS2toUTF8) {
-
-	unsigned int   utf8ValLen = avaValue.len * 3;
-	unsigned char *utf8Val    = (unsigned char*)
-				    PORT_ArenaZAlloc(newarena, utf8ValLen);
-
-	if(avaValue.len % 2 != 0 ||
-	   !PORT_UCS2_UTF8Conversion(PR_FALSE, avaValue.data, avaValue.len,
-				     utf8Val, utf8ValLen, &utf8ValLen)) {
-            PORT_FreeArena(newarena, PR_FALSE);
-	    PORT_SetError(SEC_ERROR_INVALID_AVA);
-	    return NULL;
-	}
-
+	  
 	avaValue.data = utf8Val;
 	avaValue.len = utf8ValLen;
     }
diff --git a/security/nss/lib/util/secport.c b/security/nss/lib/util/secport.c
index 7d276f9b7..1abf5c030 100644
--- a/security/nss/lib/util/secport.c
+++ b/security/nss/lib/util/secport.c
@@ -542,6 +542,15 @@ PORT_UCS2_UTF8Conversion(PRBool toUnicode, unsigned char *inBuf,
 }
 
 PRBool 
+PORT_ISO88591_UTF8Conversion(const unsigned char *inBuf,
+			 unsigned int inBufLen, unsigned char *outBuf,
+			 unsigned int maxOutBufLen, unsigned int *outBufLen)
+{
+    return sec_port_iso88591_utf8_conversion_function(inBuf, inBufLen,
+      outBuf, maxOutBufLen, outBufLen);
+}
+
+PRBool 
 PORT_UCS2_ASCIIConversion(PRBool toUnicode, unsigned char *inBuf,
 			  unsigned int inBufLen, unsigned char *outBuf,
 			  unsigned int maxOutBufLen, unsigned int *outBufLen,
diff --git a/security/nss/lib/util/secport.h b/security/nss/lib/util/secport.h
index 5cfdbd19d..693b35110 100644
--- a/security/nss/lib/util/secport.h
+++ b/security/nss/lib/util/secport.h
@@ -255,6 +255,11 @@ PRBool PORT_UCS2_UTF8Conversion(PRBool toUnicode, unsigned char *inBuf,
 			unsigned int inBufLen, unsigned char *outBuf,
 			unsigned int maxOutBufLen, unsigned int *outBufLen);
 
+/* One-way conversion from ISO-8859-1 to UTF-8 */
+PRBool PORT_ISO88591_UTF8Conversion(const unsigned char *inBuf,
+			unsigned int inBufLen, unsigned char *outBuf,
+			unsigned int maxOutBufLen, unsigned int *outBufLen);
+
 PR_EXTERN(PRBool)
 sec_port_ucs4_utf8_conversion_function
 (
@@ -277,6 +282,17 @@ sec_port_ucs2_utf8_conversion_function
   unsigned int *outBufLen
 );
 
+/* One-way conversion from ISO-8859-1 to UTF-8 */
+extern PRBool
+sec_port_iso88591_utf8_conversion_function
+(
+  const unsigned char *inBuf,
+  unsigned int inBufLen,
+  unsigned char *outBuf,
+  unsigned int maxOutBufLen,
+  unsigned int *outBufLen
+);
+
 extern int NSS_PutEnv(const char * envVarName, const char * envValue);
 
 SEC_END_PROTOS
diff --git a/security/nss/lib/util/utf8.c b/security/nss/lib/util/utf8.c
index 013f202a9..f418233eb 100644
--- a/security/nss/lib/util/utf8.c
+++ b/security/nss/lib/util/utf8.c
@@ -17,6 +17,7 @@
  * Rights Reserved.
  * 
  * Contributor(s):
+ *  John Gardiner Myers <jgmyers@speakeasy.net>
  * 
  * Alternatively, the contents of this file may be used under the
  * terms of the GNU General Public License Version 2 or later (the
@@ -553,6 +554,56 @@ sec_port_ucs2_utf8_conversion_function
   }
 }
 
+PRBool
+sec_port_iso88591_utf8_conversion_function
+(
+  const unsigned char *inBuf,
+  unsigned int inBufLen,
+  unsigned char *outBuf,
+  unsigned int maxOutBufLen,
+  unsigned int *outBufLen
+)
+{
+#ifndef TEST_UTF8
+  PORT_Assert((unsigned int *)NULL != outBufLen);
+#endif /* TEST_UTF8 */
+
+  unsigned int i, len = 0;
+
+  for( i = 0; i < inBufLen; i++) {
+    if( (inBuf[i] & 0x80) == 0x00 ) len += 1;
+    else len += 2;
+  }
+
+  if( len > maxOutBufLen ) {
+    *outBufLen = len;
+    return PR_FALSE;
+  }
+
+  len = 0;
+
+  for( i = 0; i < inBufLen; i++) {
+    if( (inBuf[i] & 0x80) == 0x00 ) {
+      /* 7F -> 0xxxxxx */
+      /* 0abcdefg -> 0abcdefg */
+
+      outBuf[len] = inBuf[i];
+      len += 1;
+    } else {
+      /* 80-FF <- 110xxxxx 10xxxxxx */
+      /* 00000000 abcdefgh -> 110000ab 10cdefgh */
+
+      outBuf[len+0] = 0xC0 | ((inBuf[i] & 0xC0) >> 6);
+      outBuf[len+1] = 0x80 | ((inBuf[i] & 0x3F) >> 0);
+
+      len += 2;
+    }
+  }
+
+  *outBufLen = len;
+  return PR_TRUE;
+}
+
 #ifdef TEST_UTF8
 
 #include <stdio.h>
@@ -1743,6 +1794,51 @@ test_utf16_chars
 #endif /* UTF16 */
 
 static PRBool
+test_iso88591_chars
+(
+  void
+)
+{
+  PRBool rv = PR_TRUE;
+  int i;
+
+  for( i = 0; i < sizeof(ucs2)/sizeof(ucs2[0]); i++ ) {
+    struct ucs2 *e = &ucs2[i];
+    PRBool result;
+    unsigned char iso88591;
+    unsigned char utf8[8];
+    unsigned int len = 0;
+
+    if (e->c > 0xFF) continue;
+
+    (void)memset(utf8, 0, sizeof(utf8));
+    iso88591 = e->c;
+    
+    result = sec_port_iso88591_utf8_conversion_function(&iso88591,
+      1, utf8, sizeof(utf8), &len);
+
+    if( !result ) {
+      fprintf(stdout, "Failed to convert ISO-8859-1 0x%02.2x to UTF-8\n", iso88591);
+      rv = PR_FALSE;
+      continue;
+    }
+
+    if( (len >= sizeof(utf8)) ||
+        (strlen(e->utf8) != len) ||
+        (utf8[len] = '\0', 0 != strcmp(e->utf8, utf8)) ) {
+      fprintf(stdout, "Wrong conversion of ISO-8859-1 0x%02.2x to UTF-8: ", iso88591);
+      dump_utf8("expected", e->utf8, ", ");
+      dump_utf8("received", utf8, "\n");
+      rv = PR_FALSE;
+      continue;
+    }
+
+  }
+
+  return rv;
+}
+
+static PRBool
 test_zeroes
 (
   void
@@ -2057,6 +2153,7 @@ main
 #ifdef UTF16
       test_utf16_chars() &&
 #endif /* UTF16 */
+      test_iso88591_chars() &&
       test_zeroes() &&
       test_multichars() &&
       PR_TRUE ) {