Files
palemoon27/dom/base/nsHTMLContentSerializer.cpp
T
roytam1 8a0a002cf2 import changes from `dev' branch of rmottola/Arctic-Fox:
- Bug 1148708: Add missing 'override' annotations in DocAccessibleChild.h. rs=ehsan (d606358545)
- Bug 1210408 - make nsMaiInterfaceAction work with proxies, r=tbsaunde (f7c819c6ae)
- Bug 1210407 - teach nsMaiInterfaceTable to use proxies, r=tbsaunde (4ca4f10b5f)
- bug 1185157 make sure we don't send an event to a destroyed ipc document r=billm (23acf53f75)
- bug 1214864 - make SetCarretOffset() async r=davidb (e3079e9b2d)
- missing of Bug 1139972 - IPC Proxy for charAt, r=tbsaunde (e9593ed752)
- bug 1191598 - Pass MOZ_CURRENT_PROJECT in environment when running post-build automation steps for universal mac builds. r=gps (fc342c6ced)
- Bug 1164596 - Add mach android-emulator command; r=ahal (afeb9b27d1)
- Bug 1223149 - Add basic usage documentation for mach build; r=glandium (bfb802d175)
- Bug 1182301 - Improve 'mach build' notifications. r=gps (2c65a122d1)
- Bug 1184696 - Add clobber targets to |mach clobber|; Ability to clobber compiled python files, r=gps (35d8be292e)
- Bug 1117958 - Allow any debugging options to the run or gtest mach subcommands to automatically enable debugging. r=gps (32f986af4b)
- Bug 1180081 - Properly rebuild gtest/libxul before running gtests. r=gps (80db9a3d49)
- Bug 1171647, part 1 - Define a new function to convert the mode to a string. r=njn (61ad16f5ba)
- Bug 1171647, part 2 - Remove redundant assertion for dark matter mode. r=njn (b5ac9519f3)
- Bug 1058178, part 1 - Implement DMD heap scanning mode. r=njn (60e1079536)
- Bug 1058178, part 2 - Implement address clamping analysis for DMD scan logs. r=njn (45c0326b93)
- Bug 1102388 - Fix DMD static constructor ordering dependency. r=mccr8 (59b87897a1)
- Bug 1128705 - Don't redefine PAGE_SIZE in DMD if it's already defined. r=erahm (49216348ee)
- Bug 1179042 - Add a script for analyzing memory blocks using a heap scan DMD log. r=njn DONTBUILD (1c08d2d66e)
- Bug 1207245 - part 6 - rename nsRefPtr<T> to RefPtr<T>; r=ehsan; a=Tomcat (1c999d139b)
- Bug 1158772 - fix non-idiomatic memset call in nsDeque.cpp; r=erahm (de6b555245)
- Bug 1199400 - Part 1: Use CheckedInt when growing nsDeque capacity. r=froydnj (dfdf6814a3)
- Bug 1199400 - Part 0: Remove unused nsDequeIterator. r=froydnj (38d69d7f47)
- Bug 1199400 - Part 2: Add tests for possible nsDeque corner cases. r=froydnj (931baff195)
- Bug 1201997 - Part 1 - Converted compiled test to gtest for nsDeque class. Added tests to test untested methods. r=froydn (e893916651)
- Bug 1201997 - Part 2 - Removing unused methods from the nsDeque class. r=froydn (41595a90ac)
- Bug 1201997 - Part 3 - Make internally used methods private. r=froydn (8cd3afd96f)
- Bug 1201997 - Part 4 - Change size and offset variables to size_t.r=froydn (73eabc8d60)
- Bug 1215140 P1 Add an nsIConsoleReportCollector interface to support navigation channel logging. r=bz (8a41535e2b)
- Bug 1215140 P2 Make HttpBaseChannel implement nsIConsoleReportCollector. r=bz (75fca301f2)
- Bug 1197679 - If nsUnknownDecoder is involved in e10s DivertToParent can break. r=jduell (5d94a12504)
- Bug 1178991 - smartptr for http converter r=hurley (8e7fbc8443)
- bug 366559 - patch 2, fix nsHTTPCompressConv indentation r=bagder (ba762da587)
- bug 366559 - patch 3, fix nsHTTPCompressConv bracing style r=bagder (54195ab451)
- bits of  bug 366559 - patch 7, content-encoding brotli for http (f0b4051022)
- Bug 1205112 - Make PushEvent.data nullable. r=mt,smaug (775db32856)
- Bug 1193414 - SharedWorkers thread should be kept alive also when the SharedWorker object is CCed, r=khuey (b77ea8125c)
- Bug 1206520: Add about:config prefs to enable throwing on asm.js validation failures; r=bz (c42126665d)
- Bug 1193414 - Telemetry for SharedWorker spawning. r=bkelly (77984b7bcc)
- Bug 1205676 - Enable WPT service-worker/unregister-then-register-new-script.https.html in e10s, r=nsm (ec24939cf6)
- Bug 1193133 - Throw when calling postMessage from a Service Worker dom object with no global. r=bkelly (526dcacfab)
- Bug 1181871 P1 Only enforce Cache Context shared data destruction on target thread after init. r=ehsan (cdbf3ed3a8)
- Bug 1181871 P2 Fix ServiceWorkerManager usage of stack-based ErrorResult. r=ehsan (c449195d90)
- minor cleanup and missing bit of 1198230 (02f459db05)
- Bug 1143717 - Implement the ServiceWorkerMessageEvent interface. r=baku (027b3465f2)
- fix misspatch (708eee4e84)
- Bug 1188545 - Disentangle service workers from shared workers and refactor event dispatching code into a separate class. r=nsm,mrbkap (fb5b5341c9)
- Bug 1205228 - Change PackagedAppVerifier to notify the verification result asynchronously. r=valentin. (9edda0fa00)
- Bug 1178518 - Packaged App Utils. r=valentin (f60f3b7a93)
- Bug 1213150 - Part 1: Add a nsContentUtils::IsNonSubresourceRequest helper; r=jdm (b509cc3cc9)
- Bug 1213150 - Part 2: Rework ShouldPrepareForIntercept() in terms of subresource requests; r=jdm (2e92fe8780)
- Bug 1213150 - Part 3: Remove nsIInterceptedChannel.isNavigation; r=jdm (becf1cc12f)
- Bug 1213150 follow-up: fix build bustage (8d73d6ca73)
- Bug 1198394 - Part 1: Allow interception of HSTS upgraded connections in non-e10s mode; r=mcmanus (f504c5be08)
- Bug 1198394 - Part 2: Add a test for interception of HSTS upgraded connections; r=jdm (054e984eef)
- Bug 1187011 - Don't allow response body with null body status. r=bkelly (b1860741d1)
- missing bit of 1140788 (29d319712e)
- Bug 1213436 - Reject core dumps with node IDs that don't fit in an IEEE 754 double; r=sfink (3c1f6fdda0)
- Bug 1211006 - Add Debugger.Source.prototype.canonicalId; r=ejpbruel (eef7b79fce)
- Bug 1199218 - Implement JS::ubi::Node::size for js::LazyScript referents; r=sfink (098a48d240)
- Bug 1220031 - Add JS::ubi::Node::scriptFilename; r=sfink (6b824ae680)
- Bug 1143575. Remove unused MediaQueue::Empty. r=cpearce (de737f3433)
- Bug 1209933 - Make sure all parent runtime pointers are the topmost parent, r=billm. (fe824d967d)
- Bug 1197012 - Fix ThrowTypeError in Notification. r=mccr8 (0b1a097526)
- Bug 1197893 - Check the number of arguments for ThrowTypeError() and ThrowRangeError() at compile time. r=peterv (d98c7d78a0)
- Bug 1142083 - Add test for IDN Unicode domain redirect. r=mcmanus (0c8961fe17)
- Bug 1187159 - Add mochitest for loading packaged apps (iframe+fetch+mozapp) r=jduell (ce90ea561b)
- Bug 1186290 - Notify TabParent to switch process when loading a signed package. r=honzab, r=kanru. (c58a14554a)
- fix (15e2df75eb)
- Bug 1206124 P1 Fix "same-origin" CORS credentials in FetchDriver. r=ehsan (fae1bb6ab3)
- Bug 1206124 P2 Test fetch() with credentials and redirects. r=ehsan (ffc6254112)
- Bug 1211751: Remove nsIChannelEventSink-forwarding from EventSource and FetchDriver. It's never needed. r=smaug (adafe5737a)
- Bug 1212433 Fail fetch() calls that require preflight and also redirect. r=sicking a=abillings (c0d6742b9e)
- Bug 1193128 - Fix base64 decoding when fetching data URIs. r=baku (80bafa291a)
- Bug 1195167 part 1: Let necko handle all protocols. r=bkelly (bb932b0ada)
- Bug 1195167 part 2: Remove redundant aCORSFlag argument and instead use mCORSFlagEverSet. r=bkelly (beadafcad0)
- Bug 1195167 part 3: Remove more scheme-specific handling from FetchDriver. r=bkelly (d00b38db9e)
- Bug 1195167 part 4: Remove FetchDriver::BasicFetch since it is empty. r=bkelly (c5ed097267)
- Bug 1210413 P2 Test CORS credentials on cross-origin redirects. r=sicking a=dveditz (b4eeb8aac0)
- Bug 1210413 P1 Propagate new channel load flags from child to parent on redirect. r=jduell a=dveditz (8b329af4fa)
- Bug 1195167 part 5: Make FetchDriver use AsyncOpen2. r=bkelly (cc217c4cc1)
- Bug 1195167 part 6: Some code simplification since necko handles fetch recursion. r=bkelly (f3b6da2262)
- Bug 1195167: Followup to fix test which I forgot to change (81e7439a2e)
- Bug 1215746: Remove RequestMode::Cors_with_forced_preflight. r=bkelly (0336e812b6)
- Bug 1211000: Move CORS preflight logic from nsCORSListenerProxy to nsCORSPreflightListener. r=ehsan (bf2f71cf22)
- missing bit of Bug 1211443 - Drop scheduled update if decoder initialization isn't done yet. r=jya (f6bc074e33)
- Bug 1182571: Fix nsILoadInfo->GetContentPolicyType API to be less ambigious. Audit and fix all users of it. r=ckerschb (5af6fa7442)
- fix (e40c8e7625)
- Bug 1173811 - Part 1: Propagate the response URL to intercepted channels when necessary (non-e10s). r=mayhemer,bkelly (26f4f13c28)
- Bug 1173811 - Part 2: Propagate the response URL to intercepted channels when necessary (e10s). r=mayhemer,bkelly (a603fe1df2)
- Bug 1154309 - Add New Resource Timing Fields r=bz,hurley (1d14eb6bef)
- Bug 1175685 - add OriginAttribute to LoadInfo. r=jonas, r=ckerschb, r=michal (a5d18bb637)
- Bug 1175685 - add OriginAttribute to LoadInfo. r=jonas, r=ckerschb, r=michal (fb07d2c8aa)
- Bug 1212904 P1 Add a LoadTainting enumeration. r=jduell (a1db8a3e99)
- Bug 1212904 P2 Add LoadTainting information to nsILoadInfo. r=jduell (2482e5e334)
- Bug 1221151 - use [infallible] in nsILoadInfo.idl instead of manual %{C++ blocks; r=jduell (aae73129b6)
- Bug 1045891 - CSP 2 child-src implementation r=ckerschb (792920aeb9)
- Bug 1219931 - CSP: Don't allow removing a policy (r=sicking) (9daaab4186)
- Bug 1208661 - Dump client-side layer textures. r=BenWa (1f2d17d515)
2022-10-17 11:08:37 +08:00

600 lines
22 KiB
C++

/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* vim: set ts=8 sts=2 et sw=2 tw=80: */
/* 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/. */
/*
* nsIContentSerializer implementation that can be used with an
* nsIDocumentEncoder to convert an HTML (not XHTML!) DOM to an HTML
* string that could be parsed into more or less the original DOM.
*/
#include "nsHTMLContentSerializer.h"
#include "nsIDOMElement.h"
#include "nsIContent.h"
#include "nsIDocument.h"
#include "nsNameSpaceManager.h"
#include "nsString.h"
#include "nsUnicharUtils.h"
#include "nsXPIDLString.h"
#include "nsIServiceManager.h"
#include "nsIDocumentEncoder.h"
#include "nsGkAtoms.h"
#include "nsIURI.h"
#include "nsNetUtil.h"
#include "nsEscape.h"
#include "nsITextToSubURI.h"
#include "nsCRT.h"
#include "nsIParserService.h"
#include "nsContentUtils.h"
#include "nsLWBrkCIID.h"
#include "nsIScriptElement.h"
#include "nsAttrName.h"
#include "nsIDocShell.h"
#include "nsIEditor.h"
#include "nsIHTMLEditor.h"
#include "mozilla/dom/Element.h"
#include "nsParserConstants.h"
using namespace mozilla::dom;
nsresult
NS_NewHTMLContentSerializer(nsIContentSerializer** aSerializer)
{
RefPtr<nsHTMLContentSerializer> it = new nsHTMLContentSerializer();
it.forget(aSerializer);
return NS_OK;
}
nsHTMLContentSerializer::nsHTMLContentSerializer()
{
mIsHTMLSerializer = true;
}
nsHTMLContentSerializer::~nsHTMLContentSerializer()
{
}
NS_IMETHODIMP
nsHTMLContentSerializer::AppendDocumentStart(nsIDocument *aDocument,
nsAString& aStr)
{
return NS_OK;
}
bool
nsHTMLContentSerializer::SerializeHTMLAttributes(nsIContent* aContent,
nsIContent *aOriginalElement,
nsAString& aTagPrefix,
const nsAString& aTagNamespaceURI,
nsIAtom* aTagName,
int32_t aNamespace,
nsAString& aStr)
{
int32_t count = aContent->GetAttrCount();
if (!count)
return true;
nsresult rv;
nsAutoString valueStr;
NS_NAMED_LITERAL_STRING(_mozStr, "_moz");
for (int32_t index = count; index > 0;) {
--index;
const nsAttrName* name = aContent->GetAttrNameAt(index);
int32_t namespaceID = name->NamespaceID();
nsIAtom* attrName = name->LocalName();
// Filter out any attribute starting with [-|_]moz
nsDependentAtomString attrNameStr(attrName);
if (StringBeginsWith(attrNameStr, NS_LITERAL_STRING("_moz")) ||
StringBeginsWith(attrNameStr, NS_LITERAL_STRING("-moz"))) {
continue;
}
aContent->GetAttr(namespaceID, attrName, valueStr);
//
// Filter out special case of <br type="_moz"> or <br _moz*>,
// used by the editor. Bug 16988. Yuck.
//
if (aTagName == nsGkAtoms::br && aNamespace == kNameSpaceID_XHTML &&
attrName == nsGkAtoms::type && namespaceID == kNameSpaceID_None &&
StringBeginsWith(valueStr, _mozStr)) {
continue;
}
if (mIsCopying && mIsFirstChildOfOL &&
aTagName == nsGkAtoms::li && aNamespace == kNameSpaceID_XHTML &&
attrName == nsGkAtoms::value && namespaceID == kNameSpaceID_None){
// This is handled separately in SerializeLIValueAttribute()
continue;
}
bool isJS = IsJavaScript(aContent, attrName, namespaceID, valueStr);
if (((attrName == nsGkAtoms::href &&
(namespaceID == kNameSpaceID_None ||
namespaceID == kNameSpaceID_XLink)) ||
(attrName == nsGkAtoms::src && namespaceID == kNameSpaceID_None))) {
// Make all links absolute when converting only the selection:
if (mFlags & nsIDocumentEncoder::OutputAbsoluteLinks) {
// Would be nice to handle OBJECT and APPLET tags,
// but that gets more complicated since we have to
// search the tag list for CODEBASE as well.
// For now, just leave them relative.
nsCOMPtr<nsIURI> uri = aContent->GetBaseURI();
if (uri) {
nsAutoString absURI;
rv = NS_MakeAbsoluteURI(absURI, valueStr, uri);
if (NS_SUCCEEDED(rv)) {
valueStr = absURI;
}
}
}
// Need to escape URI.
nsAutoString tempURI(valueStr);
if (!isJS && NS_FAILED(EscapeURI(aContent, tempURI, valueStr)))
valueStr = tempURI;
}
if (mRewriteEncodingDeclaration && aTagName == nsGkAtoms::meta &&
aNamespace == kNameSpaceID_XHTML && attrName == nsGkAtoms::content
&& namespaceID == kNameSpaceID_None) {
// If we're serializing a <meta http-equiv="content-type">,
// use the proper value, rather than what's in the document.
nsAutoString header;
aContent->GetAttr(kNameSpaceID_None, nsGkAtoms::httpEquiv, header);
if (header.LowerCaseEqualsLiteral("content-type")) {
valueStr = NS_LITERAL_STRING("text/html; charset=") +
NS_ConvertASCIItoUTF16(mCharset);
}
}
nsDependentAtomString nameStr(attrName);
nsAutoString prefix;
if (namespaceID == kNameSpaceID_XML) {
prefix.AssignLiteral(MOZ_UTF16("xml"));
} else if (namespaceID == kNameSpaceID_XLink) {
prefix.AssignLiteral(MOZ_UTF16("xlink"));
}
// Expand shorthand attribute.
if (aNamespace == kNameSpaceID_XHTML &&
namespaceID == kNameSpaceID_None &&
IsShorthandAttr(attrName, aTagName) &&
valueStr.IsEmpty()) {
valueStr = nameStr;
}
NS_ENSURE_TRUE(SerializeAttr(prefix, nameStr, valueStr,
aStr, !isJS), false);
}
return true;
}
NS_IMETHODIMP
nsHTMLContentSerializer::AppendElementStart(Element* aElement,
Element* aOriginalElement,
nsAString& aStr)
{
NS_ENSURE_ARG(aElement);
nsIContent* content = aElement;
bool forceFormat = false;
nsresult rv = NS_OK;
if (!CheckElementStart(content, forceFormat, aStr, rv)) {
return rv;
}
NS_ENSURE_SUCCESS(rv, rv);
nsIAtom *name = content->NodeInfo()->NameAtom();
int32_t ns = content->GetNameSpaceID();
bool lineBreakBeforeOpen = LineBreakBeforeOpen(ns, name);
if ((mDoFormat || forceFormat) && !mDoRaw && !PreLevel()) {
if (mColPos && lineBreakBeforeOpen) {
NS_ENSURE_TRUE(AppendNewLineToString(aStr), NS_ERROR_OUT_OF_MEMORY);
}
else {
NS_ENSURE_TRUE(MaybeAddNewlineForRootNode(aStr), NS_ERROR_OUT_OF_MEMORY);
}
if (!mColPos) {
NS_ENSURE_TRUE(AppendIndentation(aStr), NS_ERROR_OUT_OF_MEMORY);
}
else if (mAddSpace) {
bool result = AppendToString(char16_t(' '), aStr);
mAddSpace = false;
NS_ENSURE_TRUE(result, NS_ERROR_OUT_OF_MEMORY);
}
}
else if (mAddSpace) {
bool result = AppendToString(char16_t(' '), aStr);
mAddSpace = false;
NS_ENSURE_TRUE(result, NS_ERROR_OUT_OF_MEMORY);
}
else {
NS_ENSURE_TRUE(MaybeAddNewlineForRootNode(aStr), NS_ERROR_OUT_OF_MEMORY);
}
// Always reset to avoid false newlines in case MaybeAddNewlineForRootNode wasn't
// called
mAddNewlineForRootNode = false;
NS_ENSURE_TRUE(AppendToString(kLessThan, aStr), NS_ERROR_OUT_OF_MEMORY);
NS_ENSURE_TRUE(AppendToString(nsDependentAtomString(name), aStr), NS_ERROR_OUT_OF_MEMORY);
MaybeEnterInPreContent(content);
// for block elements, we increase the indentation
if ((mDoFormat || forceFormat) && !mDoRaw && !PreLevel())
NS_ENSURE_TRUE(IncrIndentation(name), NS_ERROR_OUT_OF_MEMORY);
// Need to keep track of OL and LI elements in order to get ordinal number
// for the LI.
if (mIsCopying && name == nsGkAtoms::ol && ns == kNameSpaceID_XHTML){
// We are copying and current node is an OL;
// Store its start attribute value in olState->startVal.
nsAutoString start;
int32_t startAttrVal = 0;
aElement->GetAttr(kNameSpaceID_None, nsGkAtoms::start, start);
if (!start.IsEmpty()){
nsresult rv = NS_OK;
startAttrVal = start.ToInteger(&rv);
//If OL has "start" attribute, first LI element has to start with that value
//Therefore subtracting 1 as all the LI elements are incrementing it before using it;
//In failure of ToInteger(), default StartAttrValue to 0.
if (NS_SUCCEEDED(rv))
startAttrVal--;
else
startAttrVal = 0;
}
mOLStateStack.AppendElement(olState(startAttrVal, true));
}
if (mIsCopying && name == nsGkAtoms::li && ns == kNameSpaceID_XHTML) {
mIsFirstChildOfOL = IsFirstChildOfOL(aOriginalElement);
if (mIsFirstChildOfOL){
// If OL is parent of this LI, serialize attributes in different manner.
NS_ENSURE_TRUE(SerializeLIValueAttribute(aElement, aStr), NS_ERROR_OUT_OF_MEMORY);
}
}
// Even LI passed above have to go through this
// for serializing attributes other than "value".
nsAutoString dummyPrefix;
NS_ENSURE_TRUE(SerializeHTMLAttributes(content,
aOriginalElement,
dummyPrefix,
EmptyString(),
name,
ns,
aStr), NS_ERROR_OUT_OF_MEMORY);
NS_ENSURE_TRUE(AppendToString(kGreaterThan, aStr), NS_ERROR_OUT_OF_MEMORY);
if (ns == kNameSpaceID_XHTML &&
(name == nsGkAtoms::script ||
name == nsGkAtoms::style ||
name == nsGkAtoms::noscript ||
name == nsGkAtoms::noframes)) {
++mDisableEntityEncoding;
}
if ((mDoFormat || forceFormat) && !mDoRaw && !PreLevel() &&
LineBreakAfterOpen(ns, name)) {
NS_ENSURE_TRUE(AppendNewLineToString(aStr), NS_ERROR_OUT_OF_MEMORY);
}
NS_ENSURE_TRUE(AfterElementStart(content, aOriginalElement, aStr), NS_ERROR_OUT_OF_MEMORY);
return NS_OK;
}
NS_IMETHODIMP
nsHTMLContentSerializer::AppendElementEnd(Element* aElement,
nsAString& aStr)
{
NS_ENSURE_ARG(aElement);
nsIContent* content = aElement;
nsIAtom *name = content->NodeInfo()->NameAtom();
int32_t ns = content->GetNameSpaceID();
if (ns == kNameSpaceID_XHTML &&
(name == nsGkAtoms::script ||
name == nsGkAtoms::style ||
name == nsGkAtoms::noscript ||
name == nsGkAtoms::noframes)) {
--mDisableEntityEncoding;
}
bool forceFormat = !(mFlags & nsIDocumentEncoder::OutputIgnoreMozDirty) &&
content->HasAttr(kNameSpaceID_None, nsGkAtoms::mozdirty);
if ((mDoFormat || forceFormat) && !mDoRaw && !PreLevel()) {
DecrIndentation(name);
}
if (name == nsGkAtoms::script) {
nsCOMPtr<nsIScriptElement> script = do_QueryInterface(aElement);
if (ShouldMaintainPreLevel() && script && script->IsMalformed()) {
// We're looking at a malformed script tag. This means that the end tag
// was missing in the source. Imitate that here by not serializing the end
// tag.
--PreLevel();
return NS_OK;
}
}
else if (mIsCopying && name == nsGkAtoms::ol && ns == kNameSpaceID_XHTML) {
NS_ASSERTION((!mOLStateStack.IsEmpty()), "Cannot have an empty OL Stack");
/* Though at this point we must always have an state to be deleted as all
the OL opening tags are supposed to push an olState object to the stack*/
if (!mOLStateStack.IsEmpty()) {
mOLStateStack.RemoveElementAt(mOLStateStack.Length() -1);
}
}
if (ns == kNameSpaceID_XHTML) {
nsIParserService* parserService = nsContentUtils::GetParserService();
if (parserService) {
bool isContainer;
parserService->
IsContainer(parserService->HTMLCaseSensitiveAtomTagToId(name),
isContainer);
if (!isContainer) {
return NS_OK;
}
}
}
if ((mDoFormat || forceFormat) && !mDoRaw && !PreLevel()) {
bool lineBreakBeforeClose = LineBreakBeforeClose(ns, name);
if (mColPos && lineBreakBeforeClose) {
NS_ENSURE_TRUE(AppendNewLineToString(aStr), NS_ERROR_OUT_OF_MEMORY);
}
if (!mColPos) {
NS_ENSURE_TRUE(AppendIndentation(aStr), NS_ERROR_OUT_OF_MEMORY);
}
else if (mAddSpace) {
bool result = AppendToString(char16_t(' '), aStr);
mAddSpace = false;
NS_ENSURE_TRUE(result, NS_ERROR_OUT_OF_MEMORY);
}
}
else if (mAddSpace) {
bool result = AppendToString(char16_t(' '), aStr);
mAddSpace = false;
NS_ENSURE_TRUE(result, NS_ERROR_OUT_OF_MEMORY);
}
NS_ENSURE_TRUE(AppendToString(kEndTag, aStr), NS_ERROR_OUT_OF_MEMORY);
NS_ENSURE_TRUE(AppendToString(nsDependentAtomString(name), aStr), NS_ERROR_OUT_OF_MEMORY);
NS_ENSURE_TRUE(AppendToString(kGreaterThan, aStr), NS_ERROR_OUT_OF_MEMORY);
MaybeLeaveFromPreContent(content);
if ((mDoFormat || forceFormat)&& !mDoRaw && !PreLevel()
&& LineBreakAfterClose(ns, name)) {
NS_ENSURE_TRUE(AppendNewLineToString(aStr), NS_ERROR_OUT_OF_MEMORY);
}
else {
MaybeFlagNewlineForRootNode(aElement);
}
if (name == nsGkAtoms::body && ns == kNameSpaceID_XHTML) {
--mInBody;
}
return NS_OK;
}
static const uint16_t kValNBSP = 160;
static const char* kEntities[] = {
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, "&amp;", nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
"&lt;", nullptr, "&gt;", nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
"&nbsp;"
};
static const char* kAttrEntities[] = {
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, "&quot;", nullptr, nullptr, nullptr, "&amp;", nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
"&lt;", nullptr, "&gt;", nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr, nullptr,
"&nbsp;"
};
uint32_t FindNextBasicEntity(const nsAString& aStr,
const uint32_t aLen,
uint32_t aIndex,
const char** aEntityTable,
const char** aEntity)
{
for (; aIndex < aLen; ++aIndex) {
// for each character in this chunk, check if it
// needs to be replaced
char16_t val = aStr[aIndex];
if (val <= kValNBSP && aEntityTable[val]) {
*aEntity = aEntityTable[val];
return aIndex;
}
}
return aIndex;
}
bool
nsHTMLContentSerializer::AppendAndTranslateEntities(const nsAString& aStr,
nsAString& aOutputStr)
{
if (mBodyOnly && !mInBody) {
return true;
}
if (mDisableEntityEncoding) {
return aOutputStr.Append(aStr, mozilla::fallible);
}
bool nonBasicEntities =
!!(mFlags & (nsIDocumentEncoder::OutputEncodeLatin1Entities |
nsIDocumentEncoder::OutputEncodeHTMLEntities |
nsIDocumentEncoder::OutputEncodeW3CEntities));
if (!nonBasicEntities &&
(mFlags & (nsIDocumentEncoder::OutputEncodeBasicEntities))) {
const char **entityTable = mInAttribute ? kAttrEntities : kEntities;
uint32_t start = 0;
const uint32_t len = aStr.Length();
for (uint32_t i = 0; i < len; ++i) {
const char* entity = nullptr;
i = FindNextBasicEntity(aStr, len, i, entityTable, &entity);
uint32_t normalTextLen = i - start;
if (normalTextLen) {
NS_ENSURE_TRUE(aOutputStr.Append(Substring(aStr, start, normalTextLen),
mozilla::fallible), false);
}
if (entity) {
NS_ENSURE_TRUE(aOutputStr.AppendASCII(entity, mozilla::fallible), false);
start = i + 1;
}
}
return true;
} else if (nonBasicEntities) {
nsIParserService* parserService = nsContentUtils::GetParserService();
if (!parserService) {
NS_ERROR("Can't get parser service");
return true;
}
nsReadingIterator<char16_t> done_reading;
aStr.EndReading(done_reading);
// for each chunk of |aString|...
uint32_t advanceLength = 0;
nsReadingIterator<char16_t> iter;
const char **entityTable = mInAttribute ? kAttrEntities : kEntities;
nsAutoCString entityReplacement;
for (aStr.BeginReading(iter);
iter != done_reading;
iter.advance(int32_t(advanceLength))) {
uint32_t fragmentLength = iter.size_forward();
uint32_t lengthReplaced = 0; // the number of UTF-16 codepoints
// replaced by a particular entity
const char16_t* c = iter.get();
const char16_t* fragmentStart = c;
const char16_t* fragmentEnd = c + fragmentLength;
const char* entityText = nullptr;
const char* fullConstEntityText = nullptr;
char* fullEntityText = nullptr;
advanceLength = 0;
// for each character in this chunk, check if it
// needs to be replaced
for (; c < fragmentEnd; c++, advanceLength++) {
char16_t val = *c;
if (val <= kValNBSP && entityTable[val]) {
fullConstEntityText = entityTable[val];
break;
} else if (val > 127 &&
((val < 256 &&
mFlags & nsIDocumentEncoder::OutputEncodeLatin1Entities) ||
mFlags & nsIDocumentEncoder::OutputEncodeHTMLEntities)) {
entityReplacement.Truncate();
parserService->HTMLConvertUnicodeToEntity(val, entityReplacement);
if (!entityReplacement.IsEmpty()) {
entityText = entityReplacement.get();
break;
}
}
else if (val > 127 &&
mFlags & nsIDocumentEncoder::OutputEncodeW3CEntities &&
mEntityConverter) {
if (NS_IS_HIGH_SURROGATE(val) &&
c + 1 < fragmentEnd &&
NS_IS_LOW_SURROGATE(*(c + 1))) {
uint32_t valUTF32 = SURROGATE_TO_UCS4(val, *(++c));
if (NS_SUCCEEDED(mEntityConverter->ConvertUTF32ToEntity(valUTF32,
nsIEntityConverter::entityW3C, &fullEntityText))) {
lengthReplaced = 2;
break;
}
else {
advanceLength++;
}
}
else if (NS_SUCCEEDED(mEntityConverter->ConvertToEntity(val,
nsIEntityConverter::entityW3C,
&fullEntityText))) {
lengthReplaced = 1;
break;
}
}
}
bool result = aOutputStr.Append(fragmentStart, advanceLength, mozilla::fallible);
if (entityText) {
NS_ENSURE_TRUE(aOutputStr.Append(char16_t('&'), mozilla::fallible), false);
NS_ENSURE_TRUE(AppendASCIItoUTF16(entityText, aOutputStr, mozilla::fallible), false);
NS_ENSURE_TRUE(aOutputStr.Append(char16_t(';'), mozilla::fallible), false);
advanceLength++;
}
else if (fullConstEntityText) {
NS_ENSURE_TRUE(aOutputStr.AppendASCII(fullConstEntityText, mozilla::fallible), false);
++advanceLength;
}
// if it comes from nsIEntityConverter, it already has '&' and ';'
else if (fullEntityText) {
bool ok = AppendASCIItoUTF16(fullEntityText, aOutputStr, mozilla::fallible);
free(fullEntityText);
advanceLength += lengthReplaced;
NS_ENSURE_TRUE(ok, false);
}
NS_ENSURE_TRUE(result, false);
}
} else {
NS_ENSURE_TRUE(nsXMLContentSerializer::AppendAndTranslateEntities(aStr, aOutputStr), false);
}
return true;
}