+// © 2016 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
/*
*******************************************************************************
*
- * Copyright (C) 2003-2007, International Business Machines
+ * Copyright (C) 2003-2014, International Business Machines
* Corporation and others. All Rights Reserved.
*
*******************************************************************************
* file name: uidna.cpp
- * encoding: US-ASCII
+ * encoding: UTF-8
* tab size: 8 (not used)
* indentation:4
*
#include "punycode.h"
#include "ustr_imp.h"
#include "cmemory.h"
+#include "uassert.h"
#include "sprpimpl.h"
/* it is official IDNA ACE Prefix is "xn--" */
#define ACE_PREFIX_LENGTH 4
#define MAX_LABEL_LENGTH 63
-#define HYPHEN 0x002D
-/* The Max length of the labels should not be more than 64 */
-#define MAX_LABEL_BUFFER_SIZE 100
-#define MAX_IDN_BUFFER_SIZE 300
+/* The Max length of the labels should not be more than MAX_LABEL_LENGTH */
+#define MAX_LABEL_BUFFER_SIZE 100
+
+#define MAX_DOMAIN_NAME_LENGTH 255
+/* The Max length of the domain names should not be more than MAX_DOMAIN_NAME_LENGTH */
+#define MAX_IDN_BUFFER_SIZE MAX_DOMAIN_NAME_LENGTH+1
-#define CAPITAL_A 0x0041
-#define CAPITAL_Z 0x005A
#define LOWER_CASE_DELTA 0x0020
+#define HYPHEN 0x002D
#define FULL_STOP 0x002E
-#define DATA_FILE_NAME "uidna"
+#define CAPITAL_A 0x0041
+#define CAPITAL_Z 0x005A
inline static UChar
toASCIILower(UChar ch){
// if *limit == separator then the length returned does not include
// the separtor.
static inline int32_t
-getNextSeparator(UChar *src,int32_t srcLength,
- UChar **limit,
- UBool *done){
+getNextSeparator(UChar *src, int32_t srcLength,
+ UChar **limit, UBool *done){
if(srcLength == -1){
int32_t i;
for(i=0 ; ;i++){
int32_t options,
UStringPrepProfile* nameprep,
UParseError* parseError,
- UErrorCode* status){
+ UErrorCode* status)
+{
+ // TODO Revisit buffer handling. The label should not be over 63 ASCII characters. ICU4J may need to be updated too.
UChar b1Stack[MAX_LABEL_BUFFER_SIZE], b2Stack[MAX_LABEL_BUFFER_SIZE];
//initialize pointers to stack buffers
UChar *b1 = b1Stack, *b2 = b2Stack;
// Step 4: if the source is ASCII then proceed to step 8
if(srcIsASCII){
if(b1Len <= destCapacity){
- uprv_memmove(dest, b1, b1Len * U_SIZEOF_UCHAR);
+ u_memmove(dest, b1, b1Len);
reqLength = b1Len;
}else{
reqLength = b1Len;
goto CLEANUP;
}
//Step 7: prepend the ACE prefix
- uprv_memcpy(dest,ACE_PREFIX,ACE_PREFIX_LENGTH * U_SIZEOF_UCHAR);
+ u_memcpy(dest, ACE_PREFIX, ACE_PREFIX_LENGTH);
//Step 6: copy the contents in b2 into dest
- uprv_memcpy(dest+ACE_PREFIX_LENGTH, b2, b2Len * U_SIZEOF_UCHAR);
+ u_memcpy(dest+ACE_PREFIX_LENGTH, b2, b2Len);
}else{
*status = U_IDNA_ACE_PREFIX_ERROR;
goto CLEANUP;
}
}
- // step 8: verify the length of lable
+ // step 8: verify the length of label
if(reqLength > MAX_LABEL_LENGTH){
*status = U_IDNA_LABEL_TOO_LONG_ERROR;
}
int32_t options,
UStringPrepProfile* nameprep,
UParseError* parseError,
- UErrorCode* status){
+ UErrorCode* status)
+{
//get the options
- UBool useSTD3ASCIIRules = (UBool)((options & UIDNA_USE_STD3_RULES) != 0);
+ //UBool useSTD3ASCIIRules = (UBool)((options & UIDNA_USE_STD3_RULES) != 0);
int32_t namePrepOptions = ((options & UIDNA_ALLOW_UNASSIGNED) != 0) ? USPREP_ALLOW_UNASSIGNED: 0;
-
+
+ // TODO Revisit buffer handling. The label should not be over 63 ASCII characters. ICU4J may need to be updated too.
UChar b1Stack[MAX_LABEL_BUFFER_SIZE], b2Stack[MAX_LABEL_BUFFER_SIZE], b3Stack[MAX_LABEL_BUFFER_SIZE];
//initialize pointers to stack buffers
UChar *b1 = b1Stack, *b2 = b2Stack, *b1Prime=NULL, *b3=b3Stack;
- int32_t b1Len, b2Len, b1PrimeLen, b3Len,
+ int32_t b1Len = 0, b2Len, b1PrimeLen, b3Len,
b1Capacity = MAX_LABEL_BUFFER_SIZE,
b2Capacity = MAX_LABEL_BUFFER_SIZE,
b3Capacity = MAX_LABEL_BUFFER_SIZE,
reqLength=0;
- b1Len = 0;
UBool* caseFlags = NULL;
UBool srcIsASCII = TRUE;
- UBool srcIsLDH = TRUE;
- int32_t failPos =0;
+ /*UBool srcIsLDH = TRUE;
+ int32_t failPos =0;*/
// step 1: find out if all the codepoints in src are ASCII
if(srcLength==-1){
for(;src[srcLength]!=0;){
if(src[srcLength]> 0x7f){
srcIsASCII = FALSE;
- }else if(isLDHChar(src[srcLength])==FALSE){
+ }/*else if(isLDHChar(src[srcLength])==FALSE){
// here we do not assemble surrogates
// since we know that LDH code points
// are in the ASCII range only
srcIsLDH = FALSE;
failPos = srcLength;
- }
+ }*/
srcLength++;
}
}else if(srcLength > 0){
for(int32_t j=0; j<srcLength; j++){
if(src[j]> 0x7f){
srcIsASCII = FALSE;
- }else if(isLDHChar(src[j])==FALSE){
+ }/*else if(isLDHChar(src[j])==FALSE){
// here we do not assemble surrogates
// since we know that LDH code points
// are in the ASCII range only
srcIsLDH = FALSE;
failPos = j;
- }
+ }*/
}
}else{
return 0;
}
-
+
if(srcIsASCII == FALSE){
// step 2: process the string
b1Len = usprep_prepare(nameprep, src, srcLength, b1, b1Capacity, namePrepOptions, parseError, status);
b1Len = srcLength;
}
+ // The RFC states that
+ // <quote>
+ // ToUnicode never fails. If any step fails, then the original input
+ // is returned immediately in that step.
+ // </quote>
+
//step 3: verify ACE Prefix
- if(startsWithPrefix(src,srcLength)){
-
+ if(startsWithPrefix(b1,b1Len)){
+
//step 4: Remove the ACE Prefix
b1Prime = b1 + ACE_PREFIX_LENGTH;
b1PrimeLen = b1Len - ACE_PREFIX_LENGTH;
//step 5: Decode using punycode
b2Len = u_strFromPunycode(b1Prime, b1PrimeLen, b2, b2Capacity, caseFlags,status);
-
+
if(*status == U_BUFFER_OVERFLOW_ERROR){
// redo processing of string
/* we do not have enough room so grow the buffer*/
}
*status = U_ZERO_ERROR; // reset error
-
+
b2Len = u_strFromPunycode(b1Prime, b1PrimeLen, b2, b2Len, caseFlags, status);
-
}
-
+
+
//step 6:Apply toASCII
- b3Len = uidna_toASCII(b2, b2Len, b3, b3Capacity,options,parseError, status);
-
+ b3Len = uidna_toASCII(b2, b2Len, b3, b3Capacity, options, parseError, status);
+
if(*status == U_BUFFER_OVERFLOW_ERROR){
// redo processing of string
/* we do not have enough room so grow the buffer*/
}
*status = U_ZERO_ERROR; // reset error
-
+
b3Len = uidna_toASCII(b2,b2Len,b3,b3Len,options,parseError, status);
-
+
}
-
//bail out on error
if(U_FAILURE(*status)){
goto CLEANUP;
//step 7: verify
if(compareCaseInsensitiveASCII(b1, b1Len, b3, b3Len) !=0){
- *status = U_IDNA_VERIFICATION_ERROR;
+ // Cause the original to be returned.
+ *status = U_IDNA_VERIFICATION_ERROR;
goto CLEANUP;
}
//step 8: return output of step 5
reqLength = b2Len;
if(b2Len <= destCapacity) {
- uprv_memmove(dest, b2, b2Len * U_SIZEOF_UCHAR);
+ u_memmove(dest, b2, b2Len);
}
- }else{
+ }
+ else{
+ // See the start of this if statement for why this is commented out.
// verify that STD3 ASCII rules are satisfied
- if(useSTD3ASCIIRules == TRUE){
- if( srcIsLDH == FALSE /* source contains some non-LDH characters */
+ /*if(useSTD3ASCIIRules == TRUE){
+ if( srcIsLDH == FALSE // source contains some non-LDH characters
|| src[0] == HYPHEN || src[srcLength-1] == HYPHEN){
*status = U_IDNA_STD3_ASCII_RULES_ERROR;
- /* populate the parseError struct */
+ // populate the parseError struct
if(srcIsLDH==FALSE){
// failPos is always set the index of failure
uprv_syntaxError(src,failPos, srcLength,parseError);
goto CLEANUP;
}
- }
+ }*/
+ // just return the source
//copy the source to destination
if(srcLength <= destCapacity){
- uprv_memmove(dest,src,srcLength * U_SIZEOF_UCHAR);
+ u_memmove(dest, src, srcLength);
}
reqLength = srcLength;
}
+
CLEANUP:
if(b1 != b1Stack && b1!=src){
uprv_free(b2);
}
uprv_free(caseFlags);
-
-
+
// The RFC states that
// <quote>
// ToUnicode never fails. If any step fails, then the original input
if(U_FAILURE(*status)){
//copy the source to destination
if(dest && srcLength <= destCapacity){
- if(srcLength == -1) {
- uprv_memmove(dest,src,u_strlen(src)* U_SIZEOF_UCHAR);
- } else {
- uprv_memmove(dest,src,srcLength * U_SIZEOF_UCHAR);
- }
+ // srcLength should have already been set earlier.
+ U_ASSERT(srcLength >= 0);
+ u_memmove(dest, src, srcLength);
}
reqLength = srcLength;
+ *status = U_ZERO_ERROR;
}
return u_terminateUChars(dest, destCapacity, reqLength, status);
return 0;
}
- UStringPrepProfile* nameprep = usprep_open(NULL,DATA_FILE_NAME, status);
+ UStringPrepProfile* nameprep = usprep_openByType(USPREP_RFC3491_NAMEPREP, status);
if(U_FAILURE(*status)){
return -1;
*status = U_ILLEGAL_ARGUMENT_ERROR;
return 0;
}
-
- UStringPrepProfile* nameprep = usprep_open(NULL, DATA_FILE_NAME, status);
+
+ UStringPrepProfile* nameprep = usprep_openByType(USPREP_RFC3491_NAMEPREP, status);
if(U_FAILURE(*status)){
return -1;
int32_t reqLength = 0;
- UStringPrepProfile* nameprep = usprep_open(NULL, DATA_FILE_NAME, status);
+ UStringPrepProfile* nameprep = usprep_openByType(USPREP_RFC3491_NAMEPREP, status);
if(U_FAILURE(*status)){
return 0;
// should never occur
remainingDestCapacity = 0;
}
+
if(done == TRUE){
break;
}
*currentDest++ = FULL_STOP;
remainingDestCapacity--;
}
- reqLength++;
+ reqLength++;
labelStart = delimiter;
if(remainingLen >0 ){
}
}
-
+
+ if(reqLength > MAX_DOMAIN_NAME_LENGTH){
+ *status = U_IDNA_DOMAIN_NAME_TOO_LONG_ERROR;
+ }
+
usprep_close(nameprep);
return u_terminateUChars(dest, destCapacity, reqLength, status);
int32_t reqLength = 0;
- UStringPrepProfile* nameprep = usprep_open(NULL, DATA_FILE_NAME, status);
+ UStringPrepProfile* nameprep = usprep_openByType(USPREP_RFC3491_NAMEPREP, status);
if(U_FAILURE(*status)){
return 0;
int32_t labelLen = 0, labelReqLength = 0;
UBool done = FALSE;
-
for(;;){
labelLen = getNextSeparator(labelStart,remainingLen, &delimiter,&done);
- if(labelLen==0 && done==FALSE){
+ // The RFC states that
+ // <quote>
+ // ToUnicode never fails. If any step fails, then the original input
+ // is returned immediately in that step.
+ // </quote>
+ // _internal_toUnicode will copy the label.
+ /*if(labelLen==0 && done==FALSE){
*status = U_IDNA_ZERO_LENGTH_LABEL_ERROR;
- }
+ break;
+ }*/
+
labelReqLength = _internal_toUnicode(labelStart, labelLen,
currentDest, remainingDestCapacity,
options, nameprep,
parseError, status);
if(*status == U_BUFFER_OVERFLOW_ERROR){
-
*status = U_ZERO_ERROR; // reset error
remainingDestCapacity = 0;
}
-
if(U_FAILURE(*status)){
break;
}
}
// add the label separator
+ // Unlike the ToASCII operation we don't normalize the label separators
if(remainingDestCapacity > 0){
- *currentDest++ = FULL_STOP;
+ *currentDest++ = *(labelStart + labelLen);
remainingDestCapacity--;
}
- reqLength++;
+ reqLength++;
labelStart = delimiter;
if(remainingLen >0 ){
}
}
-
+
+ if(reqLength > MAX_DOMAIN_NAME_LENGTH){
+ *status = U_IDNA_DOMAIN_NAME_TOO_LONG_ERROR;
+ }
+
usprep_close(nameprep);
return u_terminateUChars(dest, destCapacity, reqLength, status);
projects / apple / icu.git / blobdiff