michael@0: /* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
michael@0: /* This Source Code Form is subject to the terms of the Mozilla Public
michael@0:  * License, v. 2.0. If a copy of the MPL was not distributed with this
michael@0:  * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
michael@0: #include "nsString.h"
michael@0: #include "nsIUnicodeEncoder.h"
michael@0: #include "nsICharsetConverterManager.h"
michael@0: #include "nsITextToSubURI.h"
michael@0: #include "nsEscape.h"
michael@0: #include "nsTextToSubURI.h"
michael@0: #include "nsCRT.h"
michael@0: #include "nsServiceManagerUtils.h"
michael@0: 
michael@0: static NS_DEFINE_CID(kCharsetConverterManagerCID, NS_ICHARSETCONVERTERMANAGER_CID);
michael@0: 
michael@0: nsTextToSubURI::nsTextToSubURI()
michael@0: {
michael@0: }
michael@0: nsTextToSubURI::~nsTextToSubURI()
michael@0: {
michael@0: }
michael@0: 
michael@0: NS_IMPL_ISUPPORTS(nsTextToSubURI, nsITextToSubURI)
michael@0: 
michael@0: NS_IMETHODIMP  nsTextToSubURI::ConvertAndEscape(
michael@0:   const char *charset, const char16_t *text, char **_retval) 
michael@0: {
michael@0:   if(nullptr == _retval)
michael@0:     return NS_ERROR_NULL_POINTER;
michael@0:   *_retval = nullptr;
michael@0:   nsresult rv = NS_OK;
michael@0:   
michael@0:   // Get Charset, get the encoder.
michael@0:   nsICharsetConverterManager *ccm;
michael@0:   rv = CallGetService(kCharsetConverterManagerCID, &ccm);
michael@0:   if(NS_SUCCEEDED(rv)) {
michael@0:      nsIUnicodeEncoder *encoder;
michael@0:      rv = ccm->GetUnicodeEncoder(charset, &encoder);
michael@0:      NS_RELEASE(ccm);
michael@0:      if (NS_SUCCEEDED(rv)) {
michael@0:        rv = encoder->SetOutputErrorBehavior(nsIUnicodeEncoder::kOnError_Replace, nullptr, (char16_t)'?');
michael@0:        if(NS_SUCCEEDED(rv))
michael@0:        {
michael@0:           char buf[256];
michael@0:           char *pBuf = buf;
michael@0:           int32_t ulen = text ? NS_strlen(text) : 0;
michael@0:           int32_t outlen = 0;
michael@0:           if(NS_SUCCEEDED(rv = encoder->GetMaxLength(text, ulen, &outlen))) 
michael@0:           {
michael@0:              if(outlen >= 256) {
michael@0:                 pBuf = (char*)NS_Alloc(outlen+1);
michael@0:              }
michael@0:              if(nullptr == pBuf) {
michael@0:                 outlen = 255;
michael@0:                 pBuf = buf;
michael@0:              }
michael@0:              int32_t bufLen = outlen;
michael@0:              if(NS_SUCCEEDED(rv = encoder->Convert(text,&ulen, pBuf, &outlen))) {
michael@0:                 // put termination characters (e.g. ESC(B of ISO-2022-JP) if necessary
michael@0:                 int32_t finLen = bufLen - outlen;
michael@0:                 if (finLen > 0) {
michael@0:                   if (NS_SUCCEEDED(encoder->Finish((char *)(pBuf+outlen), &finLen)))
michael@0:                     outlen += finLen;
michael@0:                 }
michael@0:                 pBuf[outlen] = '\0';
michael@0:                 *_retval = nsEscape(pBuf, url_XPAlphas);
michael@0:                 if(nullptr == *_retval)
michael@0:                   rv = NS_ERROR_OUT_OF_MEMORY;
michael@0:              }
michael@0:           }
michael@0:           if(pBuf != buf)
michael@0:              NS_Free(pBuf);
michael@0:        }
michael@0:        NS_RELEASE(encoder);
michael@0:      }
michael@0:   }
michael@0:   
michael@0:   return rv;
michael@0: }
michael@0: 
michael@0: NS_IMETHODIMP  nsTextToSubURI::UnEscapeAndConvert(
michael@0:   const char *charset, const char *text, char16_t **_retval) 
michael@0: {
michael@0:   if(nullptr == _retval)
michael@0:     return NS_ERROR_NULL_POINTER;
michael@0:   if(nullptr == text) {
michael@0:     // set empty string instead of returning error
michael@0:     // due to compatibility for old version
michael@0:     text = "";
michael@0:   }
michael@0:   *_retval = nullptr;
michael@0:   nsresult rv = NS_OK;
michael@0:   
michael@0:   // unescape the string, unescape changes the input
michael@0:   char *unescaped = NS_strdup(text);
michael@0:   if (nullptr == unescaped)
michael@0:     return NS_ERROR_OUT_OF_MEMORY;
michael@0:   unescaped = nsUnescape(unescaped);
michael@0:   NS_ASSERTION(unescaped, "nsUnescape returned null");
michael@0: 
michael@0:   // Convert from the charset to unicode
michael@0:   nsCOMPtr<nsICharsetConverterManager> ccm = 
michael@0:            do_GetService(kCharsetConverterManagerCID, &rv); 
michael@0:   if (NS_SUCCEEDED(rv)) {
michael@0:     nsIUnicodeDecoder *decoder;
michael@0:     rv = ccm->GetUnicodeDecoder(charset, &decoder);
michael@0:     if (NS_SUCCEEDED(rv)) {
michael@0:       char16_t *pBuf = nullptr;
michael@0:       int32_t len = strlen(unescaped);
michael@0:       int32_t outlen = 0;
michael@0:       if (NS_SUCCEEDED(rv = decoder->GetMaxLength(unescaped, len, &outlen))) {
michael@0:         pBuf = (char16_t *) NS_Alloc((outlen+1)*sizeof(char16_t));
michael@0:         if (nullptr == pBuf)
michael@0:           rv = NS_ERROR_OUT_OF_MEMORY;
michael@0:         else {
michael@0:           if (NS_SUCCEEDED(rv = decoder->Convert(unescaped, &len, pBuf, &outlen))) {
michael@0:             pBuf[outlen] = 0;
michael@0:             *_retval = pBuf;
michael@0:           }
michael@0:           else
michael@0:             NS_Free(pBuf);
michael@0:         }
michael@0:       }
michael@0:       NS_RELEASE(decoder);
michael@0:     }
michael@0:   }
michael@0:   NS_Free(unescaped);
michael@0: 
michael@0:   return rv;
michael@0: }
michael@0: 
michael@0: static bool statefulCharset(const char *charset)
michael@0: {
michael@0:   if (!nsCRT::strncasecmp(charset, "ISO-2022-", sizeof("ISO-2022-")-1) ||
michael@0:       !nsCRT::strcasecmp(charset, "UTF-7") ||
michael@0:       !nsCRT::strcasecmp(charset, "HZ-GB-2312"))
michael@0:     return true;
michael@0: 
michael@0:   return false;
michael@0: }
michael@0: 
michael@0: nsresult nsTextToSubURI::convertURItoUnicode(const nsAFlatCString &aCharset,
michael@0:                                              const nsAFlatCString &aURI, 
michael@0:                                              bool aIRI, 
michael@0:                                              nsAString &_retval)
michael@0: {
michael@0:   nsresult rv = NS_OK;
michael@0: 
michael@0:   // check for 7bit encoding the data may not be ASCII after we decode
michael@0:   bool isStatefulCharset = statefulCharset(aCharset.get());
michael@0: 
michael@0:   if (!isStatefulCharset && IsASCII(aURI)) {
michael@0:     CopyASCIItoUTF16(aURI, _retval);
michael@0:     return rv;
michael@0:   }
michael@0: 
michael@0:   if (!isStatefulCharset && aIRI) {
michael@0:     if (IsUTF8(aURI)) {
michael@0:       CopyUTF8toUTF16(aURI, _retval);
michael@0:       return rv;
michael@0:     }
michael@0:   }
michael@0: 
michael@0:   // empty charset could indicate UTF-8, but aURI turns out not to be UTF-8.
michael@0:   NS_ENSURE_FALSE(aCharset.IsEmpty(), NS_ERROR_INVALID_ARG);
michael@0: 
michael@0:   nsCOMPtr<nsICharsetConverterManager> charsetConverterManager;
michael@0: 
michael@0:   charsetConverterManager = do_GetService(NS_CHARSETCONVERTERMANAGER_CONTRACTID, &rv);
michael@0:   NS_ENSURE_SUCCESS(rv, rv);
michael@0: 
michael@0:   nsCOMPtr<nsIUnicodeDecoder> unicodeDecoder;
michael@0:   rv = charsetConverterManager->GetUnicodeDecoder(aCharset.get(), 
michael@0:                                                   getter_AddRefs(unicodeDecoder));
michael@0:   NS_ENSURE_SUCCESS(rv, rv);
michael@0:   unicodeDecoder->SetInputErrorBehavior(nsIUnicodeDecoder::kOnError_Signal);
michael@0: 
michael@0:   int32_t srcLen = aURI.Length();
michael@0:   int32_t dstLen;
michael@0:   rv = unicodeDecoder->GetMaxLength(aURI.get(), srcLen, &dstLen);
michael@0:   NS_ENSURE_SUCCESS(rv, rv);
michael@0: 
michael@0:   char16_t *ustr = (char16_t *) NS_Alloc(dstLen * sizeof(char16_t));
michael@0:   NS_ENSURE_TRUE(ustr, NS_ERROR_OUT_OF_MEMORY);
michael@0: 
michael@0:   rv = unicodeDecoder->Convert(aURI.get(), &srcLen, ustr, &dstLen);
michael@0: 
michael@0:   if (NS_SUCCEEDED(rv))
michael@0:     _retval.Assign(ustr, dstLen);
michael@0:   
michael@0:   NS_Free(ustr);
michael@0: 
michael@0:   return rv;
michael@0: }
michael@0: 
michael@0: NS_IMETHODIMP  nsTextToSubURI::UnEscapeURIForUI(const nsACString & aCharset, 
michael@0:                                                 const nsACString &aURIFragment, 
michael@0:                                                 nsAString &_retval)
michael@0: {
michael@0:   nsAutoCString unescapedSpec;
michael@0:   // skip control octets (0x00 - 0x1f and 0x7f) when unescaping
michael@0:   NS_UnescapeURL(PromiseFlatCString(aURIFragment), 
michael@0:                  esc_SkipControl | esc_AlwaysCopy, unescapedSpec);
michael@0: 
michael@0:   // in case of failure, return escaped URI
michael@0:   // Test for != NS_OK rather than NS_FAILED, because incomplete multi-byte
michael@0:   // sequences are also considered failure in this context
michael@0:   if (convertURItoUnicode(
michael@0:                 PromiseFlatCString(aCharset), unescapedSpec, true, _retval)
michael@0:       != NS_OK)
michael@0:     // assume UTF-8 instead of ASCII  because hostname (IDN) may be in UTF-8
michael@0:     CopyUTF8toUTF16(aURIFragment, _retval); 
michael@0:   return NS_OK;
michael@0: }
michael@0: 
michael@0: NS_IMETHODIMP  nsTextToSubURI::UnEscapeNonAsciiURI(const nsACString & aCharset, 
michael@0:                                                    const nsACString & aURIFragment, 
michael@0:                                                    nsAString &_retval)
michael@0: {
michael@0:   nsAutoCString unescapedSpec;
michael@0:   NS_UnescapeURL(PromiseFlatCString(aURIFragment),
michael@0:                  esc_AlwaysCopy | esc_OnlyNonASCII, unescapedSpec);
michael@0:   // leave the URI as it is if it's not UTF-8 and aCharset is not a ASCII
michael@0:   // superset since converting "http:" with such an encoding is always a bad 
michael@0:   // idea.
michael@0:   if (!IsUTF8(unescapedSpec) && 
michael@0:       (aCharset.LowerCaseEqualsLiteral("utf-16") ||
michael@0:        aCharset.LowerCaseEqualsLiteral("utf-16be") ||
michael@0:        aCharset.LowerCaseEqualsLiteral("utf-16le") ||
michael@0:        aCharset.LowerCaseEqualsLiteral("utf-7") ||
michael@0:        aCharset.LowerCaseEqualsLiteral("x-imap4-modified-utf7"))){
michael@0:     CopyASCIItoUTF16(aURIFragment, _retval);
michael@0:     return NS_OK;
michael@0:   }
michael@0: 
michael@0:   return convertURItoUnicode(PromiseFlatCString(aCharset), unescapedSpec, true, _retval);
michael@0: }
michael@0: 
michael@0: //----------------------------------------------------------------------