static int
toIDNA2003(const UStringPrepProfile *prep, UChar32 c, icu::UnicodeString &destString) {
UChar src[2];
int32_t srcLength=0;
U16_APPEND_UNSAFE(src, srcLength, c);
UChar *dest;
int32_t destLength;
dest=destString.getBuffer(32);
if(dest==NULL) {
return FALSE;
}
UErrorCode errorCode=U_ZERO_ERROR;
destLength=usprep_prepare(prep, src, srcLength,
dest, destString.getCapacity(),
USPREP_DEFAULT, NULL, &errorCode);
destString.releaseBuffer(destLength);
if(errorCode==U_STRINGPREP_PROHIBITED_ERROR) {
return -1;
} else {
// Returns FALSE=0 for U_STRINGPREP_UNASSIGNED_ERROR and processing errors,
// TRUE=1 if c is valid or mapped.
return U_SUCCESS(errorCode);
}
}
char *
_mongoc_sasl_prep_impl (const char *name,
const char *in_utf8,
int in_utf8_len,
bson_error_t *err)
{
/* The flow is in_utf8 -> in_utf16 -> SASLPrep -> out_utf16 -> out_utf8. */
UChar *in_utf16, *out_utf16;
char *out_utf8;
int32_t in_utf16_len, out_utf16_len, out_utf8_len;
UErrorCode error_code = U_ZERO_ERROR;
UStringPrepProfile *prep;
#define SASL_PREP_ERR_RETURN(msg) \
do { \
bson_set_error (err, \
MONGOC_ERROR_SCRAM, \
MONGOC_ERROR_SCRAM_PROTOCOL_ERROR, \
(msg), \
name); \
return NULL; \
} while (0)
/* 1. convert str to UTF-16. */
/* preflight to get the destination length. */
(void) u_strFromUTF8 (
NULL, 0, &in_utf16_len, in_utf8, in_utf8_len, &error_code);
if (error_code != U_BUFFER_OVERFLOW_ERROR) {
SASL_PREP_ERR_RETURN ("could not calculate UTF-16 length of %s");
}
/* convert to UTF-16. */
error_code = U_ZERO_ERROR;
in_utf16 = bson_malloc (sizeof (UChar) *
(in_utf16_len + 1)); /* add one for null byte. */
(void) u_strFromUTF8 (
in_utf16, in_utf16_len + 1, NULL, in_utf8, in_utf8_len, &error_code);
if (error_code) {
bson_free (in_utf16);
SASL_PREP_ERR_RETURN ("could not convert %s to UTF-16");
}
/* 2. perform SASLPrep. */
prep = usprep_openByType (USPREP_RFC4013_SASLPREP, &error_code);
if (error_code) {
bson_free (in_utf16);
SASL_PREP_ERR_RETURN ("could not start SASLPrep for %s");
}
/* preflight. */
out_utf16_len = usprep_prepare (
prep, in_utf16, in_utf16_len, NULL, 0, USPREP_DEFAULT, NULL, &error_code);
if (error_code != U_BUFFER_OVERFLOW_ERROR) {
bson_free (in_utf16);
usprep_close (prep);
SASL_PREP_ERR_RETURN ("could not calculate SASLPrep length of %s");
}
/* convert. */
error_code = U_ZERO_ERROR;
out_utf16 = bson_malloc (sizeof (UChar) * (out_utf16_len + 1));
(void) usprep_prepare (prep,
in_utf16,
in_utf16_len,
out_utf16,
out_utf16_len + 1,
USPREP_DEFAULT,
NULL,
&error_code);
if (error_code) {
bson_free (in_utf16);
bson_free (out_utf16);
usprep_close (prep);
SASL_PREP_ERR_RETURN ("could not execute SASLPrep for %s");
}
bson_free (in_utf16);
usprep_close (prep);
/* 3. convert back to UTF-8. */
/* preflight. */
(void) u_strToUTF8 (
NULL, 0, &out_utf8_len, out_utf16, out_utf16_len, &error_code);
if (error_code != U_BUFFER_OVERFLOW_ERROR) {
bson_free (out_utf16);
SASL_PREP_ERR_RETURN ("could not calculate UTF-8 length of %s");
}
/* convert. */
error_code = U_ZERO_ERROR;
out_utf8 = (char *) bson_malloc (
sizeof (char) * (out_utf8_len + 1)); /* add one for null byte. */
(void) u_strToUTF8 (
out_utf8, out_utf8_len + 1, NULL, out_utf16, out_utf16_len, &error_code);
if (error_code) {
bson_free (out_utf8);
bson_free (out_utf16);
SASL_PREP_ERR_RETURN ("could not convert %s back to UTF-8");
}
bson_free (out_utf16);
return out_utf8;
#undef SASL_PREP_ERR_RETURN
}
static int32_t
_internal_toUnicode(const UChar* src, int32_t srcLength,
UChar* dest, int32_t destCapacity,
int32_t options,
UStringPrepProfile* nameprep,
UParseError* parseError,
UErrorCode* status)
{
//get the options
//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,
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;*/
// step 1: find out if all the codepoints in src are ASCII
if(srcLength==-1){
srcLength = 0;
for(;src[srcLength]!=0;){
if(src[srcLength]> 0x7f){
srcIsASCII = 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){
// 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);
if(*status == U_BUFFER_OVERFLOW_ERROR){
// redo processing of string
/* we do not have enough room so grow the buffer*/
b1 = (UChar*) uprv_malloc(b1Len * U_SIZEOF_UCHAR);
if(b1==NULL){
*status = U_MEMORY_ALLOCATION_ERROR;
goto CLEANUP;
}
*status = U_ZERO_ERROR; // reset error
b1Len = usprep_prepare(nameprep, src, srcLength, b1, b1Len, namePrepOptions, parseError, status);
}
//bail out on error
if(U_FAILURE(*status)){
goto CLEANUP;
}
}else{
//just point src to b1
b1 = (UChar*) src;
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(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*/
b2 = (UChar*) uprv_malloc(b2Len * U_SIZEOF_UCHAR);
if(b2==NULL){
*status = U_MEMORY_ALLOCATION_ERROR;
goto CLEANUP;
}
*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);
if(*status == U_BUFFER_OVERFLOW_ERROR){
// redo processing of string
/* we do not have enough room so grow the buffer*/
b3 = (UChar*) uprv_malloc(b3Len * U_SIZEOF_UCHAR);
if(b3==NULL){
*status = U_MEMORY_ALLOCATION_ERROR;
goto CLEANUP;
}
*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){
// 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);
}
}
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
|| src[0] == HYPHEN || src[srcLength-1] == HYPHEN){
*status = U_IDNA_STD3_ASCII_RULES_ERROR;
// populate the parseError struct
if(srcIsLDH==FALSE){
// failPos is always set the index of failure
uprv_syntaxError(src,failPos, srcLength,parseError);
}else if(src[0] == HYPHEN){
// fail position is 0
uprv_syntaxError(src,0,srcLength,parseError);
}else{
// the last index in the source is always length-1
uprv_syntaxError(src, (srcLength>0) ? srcLength-1 : srcLength, srcLength,parseError);
}
goto CLEANUP;
}
}*/
// just return the source
//copy the source to destination
if(srcLength <= destCapacity){
uprv_memmove(dest,src,srcLength * U_SIZEOF_UCHAR);
}
reqLength = srcLength;
}
CLEANUP:
if(b1 != b1Stack && b1!=src){
uprv_free(b1);
}
if(b2 != b2Stack){
uprv_free(b2);
}
uprv_free(caseFlags);
// The RFC states that
// <quote>
// ToUnicode never fails. If any step fails, then the original input
// is returned immediately in that step.
// </quote>
// So if any step fails lets copy source to destination
if(U_FAILURE(*status)){
//copy the source to destination
if(dest && srcLength <= destCapacity){
// srcLength should have already been set earlier.
U_ASSERT(srcLength >= 0);
uprv_memmove(dest,src,srcLength * U_SIZEOF_UCHAR);
}
reqLength = srcLength;
*status = U_ZERO_ERROR;
}
return u_terminateUChars(dest, destCapacity, reqLength, status);
}
static int32_t
_internal_toASCII(const UChar* src, int32_t srcLength,
UChar* dest, int32_t destCapacity,
int32_t options,
UStringPrepProfile* nameprep,
UParseError* parseError,
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;
int32_t b1Len=0, b2Len,
b1Capacity = MAX_LABEL_BUFFER_SIZE,
b2Capacity = MAX_LABEL_BUFFER_SIZE ,
reqLength=0;
int32_t namePrepOptions = ((options & UIDNA_ALLOW_UNASSIGNED) != 0) ? USPREP_ALLOW_UNASSIGNED: 0;
UBool* caseFlags = NULL;
// the source contains all ascii codepoints
UBool srcIsASCII = TRUE;
// assume the source contains all LDH codepoints
UBool srcIsLDH = TRUE;
int32_t j=0;
//get the options
UBool useSTD3ASCIIRules = (UBool)((options & UIDNA_USE_STD3_RULES) != 0);
int32_t failPos = -1;
if(srcLength == -1){
srcLength = u_strlen(src);
}
if(srcLength > b1Capacity){
b1 = (UChar*) uprv_malloc(srcLength * U_SIZEOF_UCHAR);
if(b1==NULL){
*status = U_MEMORY_ALLOCATION_ERROR;
goto CLEANUP;
}
b1Capacity = srcLength;
}
// step 1
for( j=0;j<srcLength;j++){
if(src[j] > 0x7F){
srcIsASCII = FALSE;
}
b1[b1Len++] = src[j];
}
// step 2 is performed only if the source contains non ASCII
if(srcIsASCII == FALSE){
// step 2
b1Len = usprep_prepare(nameprep, src, srcLength, b1, b1Capacity, namePrepOptions, parseError, status);
if(*status == U_BUFFER_OVERFLOW_ERROR){
// redo processing of string
// we do not have enough room so grow the buffer
if(b1 != b1Stack){
uprv_free(b1);
}
b1 = (UChar*) uprv_malloc(b1Len * U_SIZEOF_UCHAR);
if(b1==NULL){
*status = U_MEMORY_ALLOCATION_ERROR;
goto CLEANUP;
}
*status = U_ZERO_ERROR; // reset error
b1Len = usprep_prepare(nameprep, src, srcLength, b1, b1Len, namePrepOptions, parseError, status);
}
}
// error bail out
if(U_FAILURE(*status)){
goto CLEANUP;
}
if(b1Len == 0){
*status = U_IDNA_ZERO_LENGTH_LABEL_ERROR;
goto CLEANUP;
}
// for step 3 & 4
srcIsASCII = TRUE;
for( j=0;j<b1Len;j++){
// check if output of usprep_prepare is all ASCII
if(b1[j] > 0x7F){
srcIsASCII = FALSE;
}else if(isLDHChar(b1[j])==FALSE){ // if the char is in ASCII range verify that it is an LDH character
srcIsLDH = FALSE;
failPos = j;
}
}
if(useSTD3ASCIIRules == TRUE){
// verify 3a and 3b
// 3(a) Verify the absence of non-LDH ASCII code points; that is, the
// absence of 0..2C, 2E..2F, 3A..40, 5B..60, and 7B..7F.
// 3(b) Verify the absence of leading and trailing hyphen-minus; that
// is, the absence of U+002D at the beginning and end of the
// sequence.
if( srcIsLDH == FALSE /* source at this point should not contain anyLDH characters */
|| b1[0] == HYPHEN || b1[b1Len-1] == HYPHEN){
*status = U_IDNA_STD3_ASCII_RULES_ERROR;
/* populate the parseError struct */
if(srcIsLDH==FALSE){
// failPos is always set the index of failure
uprv_syntaxError(b1,failPos, b1Len,parseError);
}else if(b1[0] == HYPHEN){
// fail position is 0
uprv_syntaxError(b1,0,b1Len,parseError);
}else{
// the last index in the source is always length-1
uprv_syntaxError(b1, (b1Len>0) ? b1Len-1 : b1Len, b1Len,parseError);
}
goto CLEANUP;
}
}
// 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);
reqLength = b1Len;
}else{
reqLength = b1Len;
goto CLEANUP;
}
}else{
// step 5 : verify the sequence does not begin with ACE prefix
if(!startsWithPrefix(b1,b1Len)){
//step 6: encode the sequence with punycode
// do not preserve the case flags for now!
// TODO: Preserve the case while implementing the RFE
// caseFlags = (UBool*) uprv_malloc(b1Len * sizeof(UBool));
// uprv_memset(caseFlags,TRUE,b1Len);
b2Len = u_strToPunycode(b1,b1Len,b2,b2Capacity,caseFlags, status);
if(*status == U_BUFFER_OVERFLOW_ERROR){
// redo processing of string
/* we do not have enough room so grow the buffer*/
b2 = (UChar*) uprv_malloc(b2Len * U_SIZEOF_UCHAR);
if(b2 == NULL){
*status = U_MEMORY_ALLOCATION_ERROR;
goto CLEANUP;
}
*status = U_ZERO_ERROR; // reset error
b2Len = u_strToPunycode(b1,b1Len,b2,b2Len,caseFlags, status);
}
//error bail out
if(U_FAILURE(*status)){
goto CLEANUP;
}
// TODO : Reconsider while implementing the case preserve RFE
// convert all codepoints to lower case ASCII
// toASCIILower(b2,b2Len);
reqLength = b2Len+ACE_PREFIX_LENGTH;
if(reqLength > destCapacity){
*status = U_BUFFER_OVERFLOW_ERROR;
goto CLEANUP;
}
//Step 7: prepend the ACE prefix
uprv_memcpy(dest,ACE_PREFIX,ACE_PREFIX_LENGTH * U_SIZEOF_UCHAR);
//Step 6: copy the contents in b2 into dest
uprv_memcpy(dest+ACE_PREFIX_LENGTH, b2, b2Len * U_SIZEOF_UCHAR);
}else{
*status = U_IDNA_ACE_PREFIX_ERROR;
//position of failure is 0
uprv_syntaxError(b1,0,b1Len,parseError);
goto CLEANUP;
}
}
// step 8: verify the length of label
if(reqLength > MAX_LABEL_LENGTH){
*status = U_IDNA_LABEL_TOO_LONG_ERROR;
}
CLEANUP:
if(b1 != b1Stack){
uprv_free(b1);
}
if(b2 != b2Stack){
uprv_free(b2);
}
uprv_free(caseFlags);
return u_terminateUChars(dest, destCapacity, reqLength, status);
}