intl/uconv/nsTextToSubURI.cpp

   1 /* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
   2 /* This Source Code Form is subject to the terms of the Mozilla Public
   3  * License, v. 2.0. If a copy of the MPL was not distributed with this
   4  * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
   5 #include "nsString.h"
   6 #include "nsITextToSubURI.h"
   7 #include "nsEscape.h"
   8 #include "nsTextToSubURI.h"
   9 #include "nsCRT.h"
  10 #include "mozilla/ArrayUtils.h"
  11 #include "mozilla/Encoding.h"
  12 #include "mozilla/Preferences.h"
  13 #include "mozilla/TextUtils.h"
  14 #include "mozilla/Utf8.h"
  15
  16 using namespace mozilla;
  17
  18 nsTextToSubURI::~nsTextToSubURI() = default;
  19
  20 NS_IMPL_ISUPPORTS(nsTextToSubURI, nsITextToSubURI)
  21
  22 NS_IMETHODIMP
  23 nsTextToSubURI::ConvertAndEscape(const nsACString& aCharset,
  24                                  const nsAString& aText, nsACString& aOut) {
  25   auto encoding = Encoding::ForLabelNoReplacement(aCharset);
  26   if (!encoding) {
  27     aOut.Truncate();
  28     return NS_ERROR_UCONV_NOCONV;
  29   }
  30   nsresult rv;
  31   const Encoding* actualEncoding;
  32   nsAutoCString intermediate;
  33   Tie(rv, actualEncoding) = encoding->Encode(aText, intermediate);
  34   Unused << actualEncoding;
  35   if (NS_FAILED(rv)) {
  36     aOut.Truncate();
  37     return rv;
  38   }
  39   bool ok = NS_Escape(intermediate, aOut, url_XPAlphas);
  40   if (!ok) {
  41     aOut.Truncate();
  42     return NS_ERROR_OUT_OF_MEMORY;
  43   }
  44   return NS_OK;
  45 }
  46
  47 NS_IMETHODIMP
  48 nsTextToSubURI::UnEscapeAndConvert(const nsACString& aCharset,
  49                                    const nsACString& aText, nsAString& aOut) {
  50   auto encoding = Encoding::ForLabelNoReplacement(aCharset);
  51   if (!encoding) {
  52     aOut.Truncate();
  53     return NS_ERROR_UCONV_NOCONV;
  54   }
  55   nsAutoCString unescaped(aText);
  56   NS_UnescapeURL(unescaped);
  57   auto rv = encoding->DecodeWithoutBOMHandling(unescaped, aOut);
  58   if (NS_SUCCEEDED(rv)) {
  59     return NS_OK;
  60   }
  61   return rv;
  62 }
  63
  64 static bool statefulCharset(const char* charset) {
  65   // HZ, UTF-7 and the CN and KR ISO-2022 variants are no longer in
  66   // mozilla-central but keeping them here just in case for the benefit of
  67   // comm-central.
  68   if (!nsCRT::strncasecmp(charset, "ISO-2022-", sizeof("ISO-2022-") - 1) ||
  69       !nsCRT::strcasecmp(charset, "UTF-7") ||
  70       !nsCRT::strcasecmp(charset, "HZ-GB-2312"))
  71     return true;
  72
  73   return false;
  74 }
  75
  76 nsresult nsTextToSubURI::convertURItoUnicode(const nsCString& aCharset,
  77                                              const nsCString& aURI,
  78                                              nsAString& aOut) {
  79   // check for 7bit encoding the data may not be ASCII after we decode
  80   bool isStatefulCharset = statefulCharset(aCharset.get());
  81
  82   if (!isStatefulCharset) {
  83     if (IsAscii(aURI)) {
  84       CopyASCIItoUTF16(aURI, aOut);
  85       return NS_OK;
  86     }
  87     if (IsUtf8(aURI)) {
  88       CopyUTF8toUTF16(aURI, aOut);
  89       return NS_OK;
  90     }
  91   }
  92
  93   // empty charset could indicate UTF-8, but aURI turns out not to be UTF-8.
  94   NS_ENSURE_FALSE(aCharset.IsEmpty(), NS_ERROR_INVALID_ARG);
  95
  96   auto encoding = Encoding::ForLabelNoReplacement(aCharset);
  97   if (!encoding) {
  98     aOut.Truncate();
  99     return NS_ERROR_UCONV_NOCONV;
 100   }
 101   return encoding->DecodeWithoutBOMHandlingAndWithoutReplacement(aURI, aOut);
 102 }
 103
 104 NS_IMETHODIMP nsTextToSubURI::UnEscapeURIForUI(const nsACString& aURIFragment,
 105                                                nsAString& _retval) {
 106   nsAutoCString unescapedSpec;
 107   // skip control octets (0x00 - 0x1f and 0x7f) when unescaping
 108   NS_UnescapeURL(PromiseFlatCString(aURIFragment),
 109                  esc_SkipControl | esc_AlwaysCopy, unescapedSpec);
 110
 111   // in case of failure, return escaped URI
 112   // Test for != NS_OK rather than NS_FAILED, because incomplete multi-byte
 113   // sequences are also considered failure in this context
 114   if (convertURItoUnicode(NS_LITERAL_CSTRING("UTF-8"), unescapedSpec,
 115                           _retval) != NS_OK) {
 116     // assume UTF-8 instead of ASCII  because hostname (IDN) may be in UTF-8
 117     CopyUTF8toUTF16(aURIFragment, _retval);
 118   }
 119
 120   // If there are any characters that are unsafe for URIs, reescape those.
 121   if (mIDNBlocklist.IsEmpty()) {
 122     mozilla::net::InitializeBlocklist(mIDNBlocklist);
 123     // we allow SPACE and IDEOGRAPHIC SPACE in this method
 124     mozilla::net::RemoveCharFromBlocklist(u' ', mIDNBlocklist);
 125     mozilla::net::RemoveCharFromBlocklist(0x3000, mIDNBlocklist);
 126   }
 127
 128   MOZ_ASSERT(!mIDNBlocklist.IsEmpty());
 129   const nsPromiseFlatString& unescapedResult = PromiseFlatString(_retval);
 130   nsString reescapedSpec;
 131   _retval = NS_EscapeURL(
 132       unescapedResult,
 133       [&](char16_t aChar) -> bool {
 134         return mozilla::net::CharInBlocklist(aChar, mIDNBlocklist);
 135       },
 136       reescapedSpec);
 137
 138   return NS_OK;
 139 }
 140
 141 NS_IMETHODIMP
 142 nsTextToSubURI::UnEscapeNonAsciiURI(const nsACString& aCharset,
 143                                     const nsACString& aURIFragment,
 144                                     nsAString& _retval) {
 145   nsAutoCString unescapedSpec;
 146   NS_UnescapeURL(PromiseFlatCString(aURIFragment),
 147                  esc_AlwaysCopy | esc_OnlyNonASCII, unescapedSpec);
 148   // leave the URI as it is if it's not UTF-8 and aCharset is not a ASCII
 149   // superset since converting "http:" with such an encoding is always a bad
 150   // idea.
 151   if (!IsUtf8(unescapedSpec) &&
 152       (aCharset.LowerCaseEqualsLiteral("utf-16") ||
 153        aCharset.LowerCaseEqualsLiteral("utf-16be") ||
 154        aCharset.LowerCaseEqualsLiteral("utf-16le") ||
 155        aCharset.LowerCaseEqualsLiteral("utf-7") ||
 156        aCharset.LowerCaseEqualsLiteral("x-imap4-modified-utf7"))) {
 157     CopyASCIItoUTF16(aURIFragment, _retval);
 158     return NS_OK;
 159   }
 160
 161   nsresult rv =
 162       convertURItoUnicode(PromiseFlatCString(aCharset), unescapedSpec, _retval);
 163   // NS_OK_UDEC_MOREINPUT is a success code, so caller can't catch the error
 164   // if the string ends with a valid (but incomplete) sequence.
 165   return rv == NS_OK_UDEC_MOREINPUT ? NS_ERROR_UDEC_ILLEGALINPUT : rv;
 166 }
 167
 168 //----------------------------------------------------------------------