// validate a given octet sequence for compliance with the specified
// encoding
bool IsValidOctetSequenceForCharset(nsACString& aCharset, const char *aOctets)
{
nsCOMPtr<nsIUTF8ConverterService> cvtUTF8(do_GetService
(NS_UTF8CONVERTERSERVICE_CONTRACTID));
if (!cvtUTF8) {
NS_WARNING("Can't get UTF8ConverterService\n");
return false;
}
nsCAutoString tmpRaw;
tmpRaw.Assign(aOctets);
nsCAutoString tmpDecoded;
nsresult rv = cvtUTF8->ConvertStringToUTF8(tmpRaw,
PromiseFlatCString(aCharset).get(),
false, false, 1, tmpDecoded);
if (rv != NS_OK) {
// we can't decode; charset may be unsupported, or the octet sequence
// is broken (illegal or incomplete octet sequence contained)
NS_WARNING("RFC2231/5987 parameter value does not decode according to specified charset\n");
return false;
}
return true;
}
// XXX : aTryLocaleCharset is not yet effective.
nsresult
nsMIMEHeaderParamImpl::DoGetParameter(const nsACString& aHeaderVal,
const char *aParamName,
ParamDecoding aDecoding,
const nsACString& aFallbackCharset,
bool aTryLocaleCharset,
char **aLang, nsAString& aResult)
{
aResult.Truncate();
nsresult rv;
// get parameter (decode RFC 2231/5987 when applicable, as specified by
// aDecoding (5987 being a subset of 2231) and return charset.)
nsXPIDLCString med;
nsXPIDLCString charset;
rv = DoParameterInternal(PromiseFlatCString(aHeaderVal).get(), aParamName,
aDecoding, getter_Copies(charset), aLang,
getter_Copies(med));
if (NS_FAILED(rv))
return rv;
// convert to UTF-8 after charset conversion and RFC 2047 decoding
// if necessary.
nsCAutoString str1;
rv = DecodeParameter(med, charset.get(), nullptr, false, str1);
NS_ENSURE_SUCCESS(rv, rv);
if (!aFallbackCharset.IsEmpty())
{
nsCAutoString str2;
nsCOMPtr<nsIUTF8ConverterService>
cvtUTF8(do_GetService(NS_UTF8CONVERTERSERVICE_CONTRACTID));
if (cvtUTF8 &&
NS_SUCCEEDED(cvtUTF8->ConvertStringToUTF8(str1,
PromiseFlatCString(aFallbackCharset).get(), false, true,
1, str2))) {
CopyUTF8toUTF16(str2, aResult);
return NS_OK;
}
}
if (IsUTF8(str1)) {
CopyUTF8toUTF16(str1, aResult);
return NS_OK;
}
if (aTryLocaleCharset && !NS_IsNativeUTF8())
return NS_CopyNativeToUnicode(str1, aResult);
CopyASCIItoUTF16(str1, aResult);
return NS_OK;
}
NS_IMETHODIMP
nsMIMEHeaderParamImpl::DecodeParameter(const nsACString& aParamValue,
const char* aCharset,
const char* aDefaultCharset,
PRBool aOverrideCharset,
nsACString& aResult)
{
aResult.Truncate();
nsresult rv;
// If aCharset is given, aParamValue was obtained from RFC2231
// encoding and we're pretty sure that it's in aCharset.
if (aCharset && *aCharset)
{
nsCOMPtr<nsIUTF8ConverterService> cvtUTF8(do_GetService(NS_UTF8CONVERTERSERVICE_CONTRACTID, &rv));
if (NS_SUCCEEDED(rv))
// skip ASCIIness/UTF8ness test if aCharset is 7bit non-ascii charset.
return cvtUTF8->ConvertStringToUTF8(aParamValue, aCharset,
IS_7BIT_NON_ASCII_CHARSET(aCharset), aResult);
}
const nsAFlatCString& param = PromiseFlatCString(aParamValue);
nsCAutoString unQuoted;
nsACString::const_iterator s, e;
param.BeginReading(s);
param.EndReading(e);
// strip '\' when used to quote CR, LF, '"' and '\'
for ( ; s != e; ++s) {
if ((*s == '\\')) {
if (++s == e) {
--s; // '\' is at the end. move back and append '\'.
}
else if (*s != nsCRT::CR && *s != nsCRT::LF && *s != '"' && *s != '\\') {
--s; // '\' is not foll. by CR,LF,'"','\'. move back and append '\'
}
// else : skip '\' and append the quoted character.
}
unQuoted.Append(*s);
}
aResult = unQuoted;
nsCAutoString decoded;
// Try RFC 2047 encoding, instead.
rv = DecodeRFC2047Header(unQuoted.get(), aDefaultCharset,
aOverrideCharset, PR_TRUE, decoded);
if (NS_SUCCEEDED(rv) && !decoded.IsEmpty())
aResult = decoded;
return rv;
}
// XXX : aTryLocaleCharset is not yet effective.
NS_IMETHODIMP
nsMIMEHeaderParamImpl::GetParameter(const nsACString& aHeaderVal,
const char *aParamName,
const nsACString& aFallbackCharset,
PRBool aTryLocaleCharset,
char **aLang, nsAString& aResult)
{
aResult.Truncate();
nsresult rv;
// get parameter (decode RFC 2231 if it's RFC 2231-encoded and
// return charset.)
nsXPIDLCString med;
nsXPIDLCString charset;
rv = GetParameterInternal(PromiseFlatCString(aHeaderVal).get(), aParamName,
getter_Copies(charset), aLang, getter_Copies(med));
if (NS_FAILED(rv))
return rv;
// convert to UTF-8 after charset conversion and RFC 2047 decoding
// if necessary.
nsCAutoString str1;
rv = DecodeParameter(med, charset.get(), nsnull, PR_FALSE, str1);
NS_ENSURE_SUCCESS(rv, rv);
if (!aFallbackCharset.IsEmpty())
{
nsCAutoString str2;
nsCOMPtr<nsIUTF8ConverterService>
cvtUTF8(do_GetService(NS_UTF8CONVERTERSERVICE_CONTRACTID));
if (cvtUTF8 &&
NS_SUCCEEDED(cvtUTF8->ConvertStringToUTF8(str1,
PromiseFlatCString(aFallbackCharset).get(), PR_FALSE, str2))) {
CopyUTF8toUTF16(str2, aResult);
return NS_OK;
}
}
if (IsUTF8(str1)) {
CopyUTF8toUTF16(str1, aResult);
return NS_OK;
}
if (aTryLocaleCharset && !NS_IsNativeUTF8())
return NS_CopyNativeToUnicode(str1, aResult);
CopyASCIItoUTF16(str1, aResult);
return NS_OK;
}
// copy 'raw' sequences of octets in aInput to aOutput.
// If aDefaultCharset is specified, the input is assumed to be in the
// charset and converted to UTF-8. Otherwise, a blind copy is made.
// If aDefaultCharset is specified, but the conversion to UTF-8
// is not successful, each octet is replaced by Unicode replacement
// chars. *aOutput is advanced by the number of output octets.
// static
void CopyRawHeader(const char *aInput, PRUint32 aLen,
const char *aDefaultCharset, nsACString &aOutput)
{
PRInt32 c;
// If aDefaultCharset is not specified, make a blind copy.
if (!aDefaultCharset || !*aDefaultCharset) {
aOutput.Append(aInput, aLen);
return;
}
// Copy as long as it's US-ASCII. An ESC may indicate ISO 2022
// A ~ may indicate it is HZ
while (aLen && (c = PRUint8(*aInput++)) != 0x1B && c != '~' && !(c & 0x80)) {
aOutput.Append(char(c));
aLen--;
}
if (!aLen) {
return;
}
aInput--;
// skip ASCIIness/UTF8ness test if aInput is supected to be a 7bit non-ascii
// string and aDefaultCharset is a 7bit non-ascii charset.
bool skipCheck = (c == 0x1B || c == '~') &&
IS_7BIT_NON_ASCII_CHARSET(aDefaultCharset);
// If not UTF-8, treat as default charset
nsCOMPtr<nsIUTF8ConverterService>
cvtUTF8(do_GetService(NS_UTF8CONVERTERSERVICE_CONTRACTID));
nsCAutoString utf8Text;
if (cvtUTF8 &&
NS_SUCCEEDED(
cvtUTF8->ConvertStringToUTF8(Substring(aInput, aInput + aLen),
aDefaultCharset, skipCheck, true, 1,
utf8Text))) {
aOutput.Append(utf8Text);
} else { // replace each octet with Unicode replacement char in UTF-8.
for (PRUint32 i = 0; i < aLen; i++) {
c = PRUint8(*aInput++);
if (c & 0x80)
aOutput.Append(REPLACEMENT_CHAR);
else
aOutput.Append(char(c));
}
}
}
// |decode_mime_part2_str| taken from comi18n.c
// Decode RFC2047-encoded words in the input and convert the result to UTF-8.
// If aOverrideCharset is true, charset in RFC2047-encoded words is
// ignored and aDefaultCharset is assumed, instead. aDefaultCharset
// is also used to convert raw octets (without RFC 2047 encoding) to UTF-8.
//static
nsresult DecodeRFC2047Str(const char *aHeader, const char *aDefaultCharset,
PRBool aOverrideCharset, nsACString &aResult)
{
const char *p, *q, *r;
char *decodedText;
const char *begin; // tracking pointer for where we are in the input buffer
PRInt32 isLastEncodedWord = 0;
const char *charsetStart, *charsetEnd;
char charset[80];
// initialize charset name to an empty string
charset[0] = '\0';
begin = aHeader;
// To avoid buffer realloc, if possible, set capacity in advance. No
// matter what, more than 3x expansion can never happen for all charsets
// supported by Mozilla. SCSU/BCSU with the sliding window set to a
// non-BMP block may be exceptions, but Mozilla does not support them.
// Neither any known mail/news program use them. Even if there's, we're
// safe because we don't use a raw *char any more.
aResult.SetCapacity(3 * strlen(aHeader));
while ((p = PL_strstr(begin, "=?")) != 0) {
if (isLastEncodedWord) {
// See if it's all whitespace.
for (q = begin; q < p; ++q) {
if (!PL_strchr(" \t\r\n", *q)) break;
}
}
if (!isLastEncodedWord || q < p) {
// copy the part before the encoded-word
CopyRawHeader(begin, p - begin, aDefaultCharset, aResult);
begin = p;
}
p += 2;
// Get charset info
charsetStart = p;
charsetEnd = 0;
for (q = p; *q != '?'; q++) {
if (*q <= ' ' || PL_strchr(especials, *q)) {
goto badsyntax;
}
// RFC 2231 section 5
if (!charsetEnd && *q == '*') {
charsetEnd = q;
}
}
if (!charsetEnd) {
charsetEnd = q;
}
// Check for too-long charset name
if (PRUint32(charsetEnd - charsetStart) >= sizeof(charset))
goto badsyntax;
memcpy(charset, charsetStart, charsetEnd - charsetStart);
charset[charsetEnd - charsetStart] = 0;
q++;
if (*q != 'Q' && *q != 'q' && *q != 'B' && *q != 'b')
goto badsyntax;
if (q[1] != '?')
goto badsyntax;
r = q;
for (r = q + 2; *r != '?'; r++) {
if (*r < ' ') goto badsyntax;
}
if (r[1] != '=')
goto badsyntax;
else if (r == q + 2) {
// it's empty, skip
begin = r + 2;
isLastEncodedWord = 1;
continue;
}
if(*q == 'Q' || *q == 'q')
decodedText = DecodeQ(q + 2, r - (q + 2));
else {
// bug 227290. ignore an extraneous '=' at the end.
// (# of characters in B-encoded part has to be a multiple of 4)
PRInt32 n = r - (q + 2);
n -= (n % 4 == 1 && !PL_strncmp(r - 3, "===", 3)) ? 1 : 0;
decodedText = PL_Base64Decode(q + 2, n, nsnull);
}
if (decodedText == nsnull)
goto badsyntax;
// Override charset if requested. Never override labeled UTF-8.
// Use default charset instead of UNKNOWN-8BIT
if ((aOverrideCharset && 0 != nsCRT::strcasecmp(charset, "UTF-8")) ||
(aDefaultCharset && 0 == nsCRT::strcasecmp(charset, "UNKNOWN-8BIT"))) {
PL_strncpy(charset, aDefaultCharset, sizeof(charset) - 1);
charset[sizeof(charset) - 1] = '\0';
}
{
nsCOMPtr<nsIUTF8ConverterService>
cvtUTF8(do_GetService(NS_UTF8CONVERTERSERVICE_CONTRACTID));
nsCAutoString utf8Text;
// skip ASCIIness/UTF8ness test if aCharset is 7bit non-ascii charset.
if (cvtUTF8 &&
NS_SUCCEEDED(
cvtUTF8->ConvertStringToUTF8(nsDependentCString(decodedText),
charset, IS_7BIT_NON_ASCII_CHARSET(charset), utf8Text))) {
aResult.Append(utf8Text);
} else {
aResult.Append(REPLACEMENT_CHAR);
}
}
PR_Free(decodedText);
begin = r + 2;
isLastEncodedWord = 1;
continue;
badsyntax:
// copy the part before the encoded-word
aResult.Append(begin, p - begin);
begin = p;
isLastEncodedWord = 0;
}
// put the tail back
CopyRawHeader(begin, strlen(begin), aDefaultCharset, aResult);
nsCAutoString tempStr(aResult);
tempStr.ReplaceChar('\t', ' ');
aResult = tempStr;
return NS_OK;
}