gecko-dev/intl/uconv/nsTextToSubURI.cpp

179 строки
6.1 KiB
C++

/* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
#include "nsString.h"
#include "nsITextToSubURI.h"
#include "nsEscape.h"
#include "nsTextToSubURI.h"
#include "nsCRT.h"
#include "mozilla/ArrayUtils.h"
#include "mozilla/Encoding.h"
#include "mozilla/Preferences.h"
#include "mozilla/TextUtils.h"
#include "mozilla/Utf8.h"
using namespace mozilla;
nsTextToSubURI::~nsTextToSubURI() = default;
NS_IMPL_ISUPPORTS(nsTextToSubURI, nsITextToSubURI)
NS_IMETHODIMP
nsTextToSubURI::ConvertAndEscape(const nsACString& aCharset,
const nsAString& aText, nsACString& aOut) {
auto encoding = Encoding::ForLabelNoReplacement(aCharset);
if (!encoding) {
aOut.Truncate();
return NS_ERROR_UCONV_NOCONV;
}
nsresult rv;
nsAutoCString intermediate;
std::tie(rv, std::ignore) = encoding->Encode(aText, intermediate);
if (NS_FAILED(rv)) {
aOut.Truncate();
return rv;
}
bool ok = NS_Escape(intermediate, aOut, url_XPAlphas);
if (!ok) {
aOut.Truncate();
return NS_ERROR_OUT_OF_MEMORY;
}
return NS_OK;
}
NS_IMETHODIMP
nsTextToSubURI::UnEscapeAndConvert(const nsACString& aCharset,
const nsACString& aText, nsAString& aOut) {
auto encoding = Encoding::ForLabelNoReplacement(aCharset);
if (!encoding) {
aOut.Truncate();
return NS_ERROR_UCONV_NOCONV;
}
nsAutoCString unescaped(aText);
NS_UnescapeURL(unescaped);
auto rv = encoding->DecodeWithoutBOMHandling(unescaped, aOut);
if (NS_SUCCEEDED(rv)) {
return NS_OK;
}
return rv;
}
static bool statefulCharset(const char* charset) {
// HZ, UTF-7 and the CN and KR ISO-2022 variants are no longer in
// mozilla-central but keeping them here just in case for the benefit of
// comm-central.
if (!nsCRT::strncasecmp(charset, "ISO-2022-", sizeof("ISO-2022-") - 1) ||
!nsCRT::strcasecmp(charset, "UTF-7") ||
!nsCRT::strcasecmp(charset, "HZ-GB-2312"))
return true;
return false;
}
// static
nsresult nsTextToSubURI::convertURItoUnicode(const nsCString& aCharset,
const nsCString& aURI,
nsAString& aOut) {
// check for 7bit encoding the data may not be ASCII after we decode
bool isStatefulCharset = statefulCharset(aCharset.get());
if (!isStatefulCharset) {
if (IsAscii(aURI)) {
CopyASCIItoUTF16(aURI, aOut);
return NS_OK;
}
if (IsUtf8(aURI)) {
CopyUTF8toUTF16(aURI, aOut);
return NS_OK;
}
}
// empty charset could indicate UTF-8, but aURI turns out not to be UTF-8.
NS_ENSURE_FALSE(aCharset.IsEmpty(), NS_ERROR_INVALID_ARG);
auto encoding = Encoding::ForLabelNoReplacement(aCharset);
if (!encoding) {
aOut.Truncate();
return NS_ERROR_UCONV_NOCONV;
}
return encoding->DecodeWithoutBOMHandlingAndWithoutReplacement(aURI, aOut);
}
NS_IMETHODIMP nsTextToSubURI::UnEscapeURIForUI(const nsACString& aURIFragment,
bool aDontEscape,
nsAString& _retval) {
nsAutoCString unescapedSpec;
// skip control octets (0x00 - 0x1f and 0x7f) when unescaping
NS_UnescapeURL(PromiseFlatCString(aURIFragment),
esc_SkipControl | esc_AlwaysCopy, unescapedSpec);
// in case of failure, return escaped URI
// Test for != NS_OK rather than NS_FAILED, because incomplete multi-byte
// sequences are also considered failure in this context
if (convertURItoUnicode("UTF-8"_ns, unescapedSpec, _retval) != NS_OK) {
// assume UTF-8 instead of ASCII because hostname (IDN) may be in UTF-8
CopyUTF8toUTF16(aURIFragment, _retval);
}
if (aDontEscape) {
return NS_OK;
}
// If there are any characters that are unsafe for URIs, reescape those.
if (mIDNBlocklist.IsEmpty()) {
mozilla::net::InitializeBlocklist(mIDNBlocklist);
// we allow SPACE and IDEOGRAPHIC SPACE in this method
mozilla::net::RemoveCharFromBlocklist(u' ', mIDNBlocklist);
mozilla::net::RemoveCharFromBlocklist(0x3000, mIDNBlocklist);
}
MOZ_ASSERT(!mIDNBlocklist.IsEmpty());
const nsPromiseFlatString& unescapedResult = PromiseFlatString(_retval);
nsString reescapedSpec;
_retval = NS_EscapeURL(
unescapedResult,
[&](char16_t aChar) -> bool {
return mozilla::net::CharInBlocklist(aChar, mIDNBlocklist);
},
reescapedSpec);
return NS_OK;
}
NS_IMETHODIMP
nsTextToSubURI::UnEscapeNonAsciiURIJS(const nsACString& aCharset,
const nsACString& aURIFragment,
nsAString& _retval) {
return UnEscapeNonAsciiURI(aCharset, aURIFragment, _retval);
}
// static
nsresult nsTextToSubURI::UnEscapeNonAsciiURI(const nsACString& aCharset,
const nsACString& aURIFragment,
nsAString& _retval) {
nsAutoCString unescapedSpec;
NS_UnescapeURL(PromiseFlatCString(aURIFragment),
esc_AlwaysCopy | esc_OnlyNonASCII, unescapedSpec);
// leave the URI as it is if it's not UTF-8 and aCharset is not a ASCII
// superset since converting "http:" with such an encoding is always a bad
// idea.
if (!IsUtf8(unescapedSpec) &&
(aCharset.LowerCaseEqualsLiteral("utf-16") ||
aCharset.LowerCaseEqualsLiteral("utf-16be") ||
aCharset.LowerCaseEqualsLiteral("utf-16le") ||
aCharset.LowerCaseEqualsLiteral("utf-7") ||
aCharset.LowerCaseEqualsLiteral("x-imap4-modified-utf7"))) {
CopyASCIItoUTF16(aURIFragment, _retval);
return NS_OK;
}
nsresult rv =
convertURItoUnicode(PromiseFlatCString(aCharset), unescapedSpec, _retval);
// NS_OK_UDEC_MOREINPUT is a success code, so caller can't catch the error
// if the string ends with a valid (but incomplete) sequence.
return rv == NS_OK_UDEC_MOREINPUT ? NS_ERROR_UDEC_ILLEGALINPUT : rv;
}
//----------------------------------------------------------------------