Quelle nsContentList.cpp Sprache: 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/. */

/*
* nsBaseContentList is a basic list of content nodes; nsContentList
* is a commonly used NodeList implementation (used for
* getElementsByTagName, some properties on HTMLDocument/Document, etc).
*/

#include "nsContentList.h"
#include "nsIContent.h"
#include "mozilla/dom/Document.h"
#include "mozilla/ContentIterator.h"
#include "mozilla/dom/Element.h"
#include "nsWrapperCacheInlines.h"
#include "nsContentUtils.h"
#include "nsCCUncollectableMarker.h"
#include "nsGkAtoms.h"
#include "mozilla/dom/HTMLCollectionBinding.h"
#include "mozilla/dom/NodeListBinding.h"
#include "mozilla/Likely.h"
#include "nsGenericHTMLElement.h"
#include "jsfriendapi.h"
#include <algorithm>
#include "mozilla/dom/NodeInfoInlines.h"
#include "mozilla/MruCache.h"
#include "mozilla/StaticPtr.h"

#include "PLDHashTable.h"
#include "nsTHashtable.h"

#ifdef DEBUG_CONTENT_LIST
#  define ASSERT_IN_SYNC AssertInSync()
#else
#  define ASSERT_IN_SYNC PR_BEGIN_MACRO PR_END_MACRO
#endif

using namespace mozilla;
using namespace mozilla::dom;

nsBaseContentList::~nsBaseContentList() = default;

NS_IMPL_CYCLE_COLLECTION_WRAPPERCACHE_CLASS(nsBaseContentList)
NS_IMPL_CYCLE_COLLECTION_UNLINK_BEGIN(nsBaseContentList)
  NS_IMPL_CYCLE_COLLECTION_UNLINK(mElements)
  NS_IMPL_CYCLE_COLLECTION_UNLINK_PRESERVED_WRAPPER
  tmp->RemoveFromCaches();
NS_IMPL_CYCLE_COLLECTION_UNLINK_END
NS_IMPL_CYCLE_COLLECTION_TRAVERSE_BEGIN(nsBaseContentList)
  NS_IMPL_CYCLE_COLLECTION_TRAVERSE(mElements)
NS_IMPL_CYCLE_COLLECTION_TRAVERSE_END

NS_IMPL_CYCLE_COLLECTION_CAN_SKIP_BEGIN(nsBaseContentList)
  if (nsCCUncollectableMarker::sGeneration && tmp->HasKnownLiveWrapper()) {
    for (uint32_t i = 0; i < tmp->mElements.Length(); ++i) {
      nsIContent* c = tmp->mElements[i];
      if (c->IsPurple()) {
        c->RemovePurple();
      }
      Element::MarkNodeChildren(c);
    }
    return true;
  }
NS_IMPL_CYCLE_COLLECTION_CAN_SKIP_END

NS_IMPL_CYCLE_COLLECTION_CAN_SKIP_IN_CC_BEGIN(nsBaseContentList)
  return nsCCUncollectableMarker::sGeneration && tmp->HasKnownLiveWrapper();
NS_IMPL_CYCLE_COLLECTION_CAN_SKIP_IN_CC_END

NS_IMPL_CYCLE_COLLECTION_CAN_SKIP_THIS_BEGIN(nsBaseContentList)
  return nsCCUncollectableMarker::sGeneration && tmp->HasKnownLiveWrapper();
NS_IMPL_CYCLE_COLLECTION_CAN_SKIP_THIS_END

// QueryInterface implementation for nsBaseContentList
NS_INTERFACE_TABLE_HEAD(nsBaseContentList)
  NS_WRAPPERCACHE_INTERFACE_TABLE_ENTRY
  NS_INTERFACE_TABLE(nsBaseContentList, nsINodeList)
  NS_INTERFACE_TABLE_TO_MAP_SEGUE_CYCLE_COLLECTION(nsBaseContentList)
NS_INTERFACE_MAP_END

NS_IMPL_CYCLE_COLLECTING_ADDREF(nsBaseContentList)
NS_IMPL_CYCLE_COLLECTING_RELEASE_WITH_LAST_RELEASE(nsBaseContentList,
                                                   LastRelease())

nsIContent* nsBaseContentList::Item(uint32_t aIndex) {
  return mElements.SafeElementAt(aIndex);
}

int32_t nsBaseContentList::IndexOf(nsIContent* aContent, bool aDoFlush) {
  return mElements.IndexOf(aContent);
}

int32_t nsBaseContentList::IndexOf(nsIContent* aContent) {
  return IndexOf(aContent, true);
}

size_t nsBaseContentList::SizeOfIncludingThis(
    MallocSizeOf aMallocSizeOf) const {
  size_t n = aMallocSizeOf(this);
  n += mElements.ShallowSizeOfExcludingThis(aMallocSizeOf);
  return n;
}

NS_IMPL_CYCLE_COLLECTION_INHERITED(nsSimpleContentList, nsBaseContentList,
                                   mRoot)

NS_INTERFACE_MAP_BEGIN_CYCLE_COLLECTION(nsSimpleContentList)
NS_INTERFACE_MAP_END_INHERITING(nsBaseContentList)

NS_IMPL_ADDREF_INHERITED(nsSimpleContentList, nsBaseContentList)
NS_IMPL_RELEASE_INHERITED(nsSimpleContentList, nsBaseContentList)

JSObject* nsSimpleContentList::WrapObject(JSContext* cx,
                                          JS::Handle<JSObject*> aGivenProto) {
  return NodeList_Binding::Wrap(cx, this, aGivenProto);
}

NS_IMPL_CYCLE_COLLECTION_INHERITED(nsEmptyContentList, nsBaseContentList, mRoot)

NS_INTERFACE_MAP_BEGIN_CYCLE_COLLECTION(nsEmptyContentList)
  NS_INTERFACE_MAP_ENTRY(nsIHTMLCollection)
NS_INTERFACE_MAP_END_INHERITING(nsBaseContentList)

NS_IMPL_ADDREF_INHERITED(nsEmptyContentList, nsBaseContentList)
NS_IMPL_RELEASE_INHERITED(nsEmptyContentList, nsBaseContentList)

JSObject* nsEmptyContentList::WrapObject(JSContext* cx,
                                         JS::Handle<JSObject*> aGivenProto) {
  return HTMLCollection_Binding::Wrap(cx, this, aGivenProto);
}

mozilla::dom::Element* nsEmptyContentList::GetElementAt(uint32_t index) {
  return nullptr;
}

mozilla::dom::Element* nsEmptyContentList::GetFirstNamedElement(
    const nsAString& aName, bool& aFound) {
  aFound = false;
  return nullptr;
}

void nsEmptyContentList::GetSupportedNames(nsTArray<nsString>& aNames) {}

nsIContent* nsEmptyContentList::Item(uint32_t aIndex) { return nullptr; }

struct ContentListCache
    : public MruCache<nsContentListKey, nsContentList*, ContentListCache> {
  static HashNumber Hash(const nsContentListKey& aKey) {
    return aKey.GetHash();
  }
  static bool Match(const nsContentListKey& aKey, const nsContentList* aVal) {
    return aVal->MatchesKey(aKey);
  }
};

static ContentListCache sRecentlyUsedContentLists;

class nsContentList::HashEntry : public PLDHashEntryHdr {
public:
  using KeyType = const nsContentListKey*;
  using KeyTypePointer = KeyType;

  // Note that this is creating a blank entry, so you'll have to manually
  // initialize it after it has been inserted into the hash table.
  explicit HashEntry(KeyTypePointer aKey) : mContentList(nullptr) {}

  HashEntry(HashEntry&& aEnt) : mContentList(std::move(aEnt.mContentList)) {}

  ~HashEntry() {
    if (mContentList) {
      MOZ_RELEASE_ASSERT(mContentList->mInHashtable);
      mContentList->mInHashtable = false;
    }
  }

  bool KeyEquals(KeyTypePointer aKey) const {
    return mContentList->MatchesKey(*aKey);
  }

  static KeyTypePointer KeyToPointer(KeyType aKey) { return aKey; }

  static PLDHashNumber HashKey(KeyTypePointer aKey) { return aKey->GetHash(); }

  nsContentList* GetContentList() const { return mContentList; }
  void SetContentList(nsContentList* aContentList) {
    MOZ_RELEASE_ASSERT(!mContentList);
    MOZ_ASSERT(aContentList);
    MOZ_RELEASE_ASSERT(!aContentList->mInHashtable);
    mContentList = aContentList;
    mContentList->mInHashtable = true;
  }

  enum { ALLOW_MEMMOVE = true };

private:
  nsContentList* MOZ_UNSAFE_REF(
      "This entry will be removed in nsContentList::RemoveFromHashtable "
      "before mContentList is destroyed") mContentList;
};

// Hashtable for storing nsContentLists
static StaticAutoPtr<nsTHashtable<nsContentList::HashEntry>>
    gContentListHashTable;

already_AddRefed<nsContentList> NS_GetContentList(nsINode* aRootNode,
                                                  int32_t aMatchNameSpaceId,
                                                  const nsAString& aTagname) {
  NS_ASSERTION(aRootNode, "content list has to have a root");

  RefPtr<nsContentList> list;
  nsContentListKey hashKey(aRootNode, aMatchNameSpaceId, aTagname,
                           aRootNode->OwnerDoc()->IsHTMLDocument());
  auto p = sRecentlyUsedContentLists.Lookup(hashKey);
  if (p) {
    list = p.Data();
    return list.forget();
  }

  // Initialize the hashtable if needed.
  if (!gContentListHashTable) {
    gContentListHashTable = new nsTHashtable<nsContentList::HashEntry>();
  }

  // First we look in our hashtable.  Then we create a content list if needed
  auto entry = gContentListHashTable->PutEntry(&hashKey, fallible);
  if (entry) {
    list = entry->GetContentList();
  }

  if (!list) {
    // We need to create a ContentList and add it to our new entry, if
    // we have an entry
    RefPtr<nsAtom> xmlAtom = NS_Atomize(aTagname);
    RefPtr<nsAtom> htmlAtom;
    if (aMatchNameSpaceId == kNameSpaceID_Unknown) {
      nsAutoString lowercaseName;
      nsContentUtils::ASCIIToLower(aTagname, lowercaseName);
      htmlAtom = NS_Atomize(lowercaseName);
    } else {
      htmlAtom = xmlAtom;
    }
    list = new nsContentList(aRootNode, aMatchNameSpaceId, htmlAtom, xmlAtom);
    if (entry) {
      entry->SetContentList(list);
    }
  }

  p.Set(list);
  return list.forget();
}

#ifdef DEBUG
const nsCacheableFuncStringContentList::ContentListType
    nsCachableElementsByNameNodeList::sType =
        nsCacheableFuncStringContentList::eNodeList;
const nsCacheableFuncStringContentList::ContentListType
    nsCacheableFuncStringHTMLCollection::sType =
        nsCacheableFuncStringContentList::eHTMLCollection;
#endif

class nsCacheableFuncStringContentList::HashEntry : public PLDHashEntryHdr {
public:
  using KeyType = const nsFuncStringCacheKey*;
  using KeyTypePointer = KeyType;

  // Note that this is creating a blank entry, so you'll have to manually
  // initialize it after it has been inserted into the hash table.
  explicit HashEntry(KeyTypePointer aKey) : mContentList(nullptr) {}

  HashEntry(HashEntry&& aEnt) : mContentList(std::move(aEnt.mContentList)) {}

  ~HashEntry() {
    if (mContentList) {
      MOZ_RELEASE_ASSERT(mContentList->mInHashtable);
      mContentList->mInHashtable = false;
    }
  }

  bool KeyEquals(KeyTypePointer aKey) const {
    return mContentList->Equals(aKey);
  }

  static KeyTypePointer KeyToPointer(KeyType aKey) { return aKey; }

  static PLDHashNumber HashKey(KeyTypePointer aKey) { return aKey->GetHash(); }

  nsCacheableFuncStringContentList* GetContentList() const {
    return mContentList;
  }
  void SetContentList(nsCacheableFuncStringContentList* aContentList) {
    MOZ_RELEASE_ASSERT(!mContentList);
    MOZ_ASSERT(aContentList);
    MOZ_RELEASE_ASSERT(!aContentList->mInHashtable);
    mContentList = aContentList;
    mContentList->mInHashtable = true;
  }

  enum { ALLOW_MEMMOVE = true };

private:
  nsCacheableFuncStringContentList* MOZ_UNSAFE_REF(
      "This entry will be removed in "
      "nsCacheableFuncStringContentList::RemoveFromFuncStringHashtable "
      "before mContentList is destroyed") mContentList;
};

// Hashtable for storing nsCacheableFuncStringContentList
static StaticAutoPtr<nsTHashtable<nsCacheableFuncStringContentList::HashEntry>>
    gFuncStringContentListHashTable;

template <class ListType>
already_AddRefed<nsContentList> GetFuncStringContentList(
    nsINode* aRootNode, nsContentListMatchFunc aFunc,
    nsContentListDestroyFunc aDestroyFunc,
    nsFuncStringContentListDataAllocator aDataAllocator,
    const nsAString& aString) {
  NS_ASSERTION(aRootNode, "content list has to have a root");

  RefPtr<nsCacheableFuncStringContentList> list;

  // Initialize the hashtable if needed.
  if (!gFuncStringContentListHashTable) {
    gFuncStringContentListHashTable =
        new nsTHashtable<nsCacheableFuncStringContentList::HashEntry>();
  }

  nsCacheableFuncStringContentList::HashEntry* entry = nullptr;
  // First we look in our hashtable.  Then we create a content list if needed
  if (gFuncStringContentListHashTable) {
    nsFuncStringCacheKey hashKey(aRootNode, aFunc, aString);

    entry = gFuncStringContentListHashTable->PutEntry(&hashKey, fallible);
    if (entry) {
      list = entry->GetContentList();
#ifdef DEBUG
      MOZ_ASSERT_IF(list, list->mType == ListType::sType);
#endif
    }
  }

  if (!list) {
    // We need to create a ContentList and add it to our new entry, if
    // we have an entry
    list =
        new ListType(aRootNode, aFunc, aDestroyFunc, aDataAllocator, aString);
    if (entry) {
      entry->SetContentList(list);
    }
  }

  // Don't cache these lists globally

  return list.forget();
}

// Explicit instantiations to avoid link errors
template already_AddRefed<nsContentList>
GetFuncStringContentList<nsCachableElementsByNameNodeList>(
    nsINode* aRootNode, nsContentListMatchFunc aFunc,
    nsContentListDestroyFunc aDestroyFunc,
    nsFuncStringContentListDataAllocator aDataAllocator,
    const nsAString& aString);
template already_AddRefed<nsContentList>
GetFuncStringContentList<nsCacheableFuncStringHTMLCollection>(
    nsINode* aRootNode, nsContentListMatchFunc aFunc,
    nsContentListDestroyFunc aDestroyFunc,
    nsFuncStringContentListDataAllocator aDataAllocator,
    const nsAString& aString);

//-----------------------------------------------------
// nsContentList implementation

nsContentList::nsContentList(nsINode* aRootNode, int32_t aMatchNameSpaceId,
                             nsAtom* aHTMLMatchAtom, nsAtom* aXMLMatchAtom,
                             bool aDeep, bool aLiveList)
    : nsBaseContentList(),
      mRootNode(aRootNode),
      mMatchNameSpaceId(aMatchNameSpaceId),
      mHTMLMatchAtom(aHTMLMatchAtom),
      mXMLMatchAtom(aXMLMatchAtom),
      mState(State::Dirty),
      mDeep(aDeep),
      mFuncMayDependOnAttr(false),
      mIsHTMLDocument(aRootNode->OwnerDoc()->IsHTMLDocument()),
      mNamedItemsCacheValid(false),
      mIsLiveList(aLiveList),
      mInHashtable(false) {
  NS_ASSERTION(mRootNode, "Must have root");
  if (nsGkAtoms::_asterisk == mHTMLMatchAtom) {
    NS_ASSERTION(mXMLMatchAtom == nsGkAtoms::_asterisk,
                 "HTML atom and XML atom are not both asterisk?");
    mMatchAll = true;
  } else {
    mMatchAll = false;
  }
  // This is aLiveList instead of mIsLiveList to avoid Valgrind errors.
  if (aLiveList) {
    SetEnabledCallbacks(nsIMutationObserver::kNodeWillBeDestroyed);
    mRootNode->AddMutationObserver(this);
  }

  // We only need to flush if we're in an non-HTML document, since the
  // HTML5 parser doesn't need flushing.  Further, if we're not in a
  // document at all right now (in the GetUncomposedDoc() sense), we're
  // not parser-created and don't need to be flushing stuff under us
  // to get our kids right.
  Document* doc = mRootNode->GetUncomposedDoc();
  mFlushesNeeded = doc && !doc->IsHTMLDocument();
}

nsContentList::nsContentList(nsINode* aRootNode, nsContentListMatchFunc aFunc,
                             nsContentListDestroyFunc aDestroyFunc, void* aData,
                             bool aDeep, nsAtom* aMatchAtom,
                             int32_t aMatchNameSpaceId,
                             bool aFuncMayDependOnAttr, bool aLiveList)
    : nsBaseContentList(),
      mRootNode(aRootNode),
      mMatchNameSpaceId(aMatchNameSpaceId),
      mHTMLMatchAtom(aMatchAtom),
      mXMLMatchAtom(aMatchAtom),
      mFunc(aFunc),
      mDestroyFunc(aDestroyFunc),
      mData(aData),
      mState(State::Dirty),
      mMatchAll(false),
      mDeep(aDeep),
      mFuncMayDependOnAttr(aFuncMayDependOnAttr),
      mIsHTMLDocument(false),
      mNamedItemsCacheValid(false),
      mIsLiveList(aLiveList),
      mInHashtable(false) {
  NS_ASSERTION(mRootNode, "Must have root");
  // This is aLiveList instead of mIsLiveList to avoid Valgrind errors.
  if (aLiveList) {
    SetEnabledCallbacks(nsIMutationObserver::kNodeWillBeDestroyed);
    mRootNode->AddMutationObserver(this);
  }

  // We only need to flush if we're in an non-HTML document, since the
  // HTML5 parser doesn't need flushing.  Further, if we're not in a
  // document at all right now (in the GetUncomposedDoc() sense), we're
  // not parser-created and don't need to be flushing stuff under us
  // to get our kids right.
  Document* doc = mRootNode->GetUncomposedDoc();
  mFlushesNeeded = doc && !doc->IsHTMLDocument();
}

nsContentList::~nsContentList() {
  RemoveFromHashtable();
  if (mIsLiveList && mRootNode) {
    mRootNode->RemoveMutationObserver(this);
  }

  if (mDestroyFunc) {
    // Clean up mData
    (*mDestroyFunc)(mData);
  }
}

JSObject* nsContentList::WrapObject(JSContext* cx,
                                    JS::Handle<JSObject*> aGivenProto) {
  return HTMLCollection_Binding::Wrap(cx, this, aGivenProto);
}

NS_IMPL_ISUPPORTS_INHERITED(nsContentList, nsBaseContentList, nsIHTMLCollection,
                            nsIMutationObserver)

uint32_t nsContentList::Length(bool aDoFlush) {
  BringSelfUpToDate(aDoFlush);

  return mElements.Length();
}

nsIContent* nsContentList::Item(uint32_t aIndex, bool aDoFlush) {
  if (mRootNode && aDoFlush && mFlushesNeeded) {
    // XXX sXBL/XBL2 issue
    Document* doc = mRootNode->GetUncomposedDoc();
    if (doc) {
      // Flush pending content changes Bug 4891.
      doc->FlushPendingNotifications(FlushType::ContentAndNotify);
    }
  }

  if (mState != State::UpToDate) {
    PopulateSelf(std::min(aIndex, UINT32_MAX - 1) + 1);
  }

  ASSERT_IN_SYNC;
  NS_ASSERTION(!mRootNode || mState != State::Dirty,
               "PopulateSelf left the list in a dirty (useless) state!");

  return mElements.SafeElementAt(aIndex);
}

inline void nsContentList::InsertElementInNamedItemsCache(
    nsIContent& aContent) {
  const bool hasName = aContent.HasName();
  const bool hasId = aContent.HasID();
  if (!hasName && !hasId) {
    return;
  }

  Element* el = aContent.AsElement();
  MOZ_ASSERT_IF(hasName, el->IsHTMLElement());

  uint32_t i = 0;
  while (BorrowedAttrInfo info = el->GetAttrInfoAt(i++)) {
    const bool valid = (info.mName->Equals(nsGkAtoms::name) && hasName) ||
                       (info.mName->Equals(nsGkAtoms::id) && hasId);
    if (!valid) {
      continue;
    }

    if (!mNamedItemsCache) {
      mNamedItemsCache = MakeUnique<NamedItemsCache>();
    }

    nsAtom* name = info.mValue->GetAtomValue();
    // NOTE: LookupOrInsert makes sure we keep the first element we find for a
    // given name.
    mNamedItemsCache->LookupOrInsert(name, el);
  }
}

inline void nsContentList::InvalidateNamedItemsCacheForAttributeChange(
    int32_t aNamespaceID, nsAtom* aAttribute) {
  if (!mNamedItemsCacheValid) {
    return;
  }
  if ((aAttribute == nsGkAtoms::id || aAttribute == nsGkAtoms::name) &&
      aNamespaceID == kNameSpaceID_None) {
    InvalidateNamedItemsCache();
  }
}

inline void nsContentList::InvalidateNamedItemsCacheForInsertion(
    Element& aElement) {
  if (!mNamedItemsCacheValid) {
    return;
  }

  InsertElementInNamedItemsCache(aElement);
}

inline void nsContentList::InvalidateNamedItemsCacheForDeletion(
    Element& aElement) {
  if (!mNamedItemsCacheValid) {
    return;
  }
  if (aElement.HasName() || aElement.HasID()) {
    InvalidateNamedItemsCache();
  }
}

void nsContentList::EnsureNamedItemsCacheValid(bool aDoFlush) {
  BringSelfUpToDate(aDoFlush);

  if (mNamedItemsCacheValid) {
    return;
  }

  MOZ_ASSERT(!mNamedItemsCache);

  // https://dom.spec.whatwg.org/#dom-htmlcollection-nameditem-key
  // XXX: Blink/WebKit don't follow the spec here, and searches first-by-id,
  // then by name.
  for (const nsCOMPtr<nsIContent>& content : mElements) {
    InsertElementInNamedItemsCache(*content);
  }

  mNamedItemsCacheValid = true;
}

Element* nsContentList::NamedItem(const nsAString& aName, bool aDoFlush) {
  if (aName.IsEmpty()) {
    return nullptr;
  }

  EnsureNamedItemsCacheValid(aDoFlush);

  if (!mNamedItemsCache) {
    return nullptr;
  }

  // Typically IDs and names are atomized
  RefPtr<nsAtom> name = NS_Atomize(aName);
  NS_ENSURE_TRUE(name, nullptr);

  return mNamedItemsCache->Get(name);
}

void nsContentList::GetSupportedNames(nsTArray<nsString>& aNames) {
  BringSelfUpToDate(true);

  AutoTArray<nsAtom*, 8> atoms;
  for (uint32_t i = 0; i < mElements.Length(); ++i) {
    nsIContent* content = mElements.ElementAt(i);
    if (content->HasID()) {
      nsAtom* id = content->GetID();
      MOZ_ASSERT(id != nsGkAtoms::_empty, "Empty ids don't get atomized");
      if (!atoms.Contains(id)) {
        atoms.AppendElement(id);
      }
    }

    nsGenericHTMLElement* el = nsGenericHTMLElement::FromNode(content);
    if (el) {
      // XXXbz should we be checking for particular tags here?  How
      // stable is this part of the spec?
      // Note: nsINode::HasName means the name is exposed on the document,
      // which is false for options, so we don't check it here.
      const nsAttrValue* val = el->GetParsedAttr(nsGkAtoms::name);
      if (val && val->Type() == nsAttrValue::eAtom) {
        nsAtom* name = val->GetAtomValue();
        MOZ_ASSERT(name != nsGkAtoms::_empty, "Empty names don't get atomized");
        if (!atoms.Contains(name)) {
          atoms.AppendElement(name);
        }
      }
    }
  }

  uint32_t atomsLen = atoms.Length();
  nsString* names = aNames.AppendElements(atomsLen);
  for (uint32_t i = 0; i < atomsLen; ++i) {
    atoms[i]->ToString(names[i]);
  }
}

int32_t nsContentList::IndexOf(nsIContent* aContent, bool aDoFlush) {
  BringSelfUpToDate(aDoFlush);

  return mElements.IndexOf(aContent);
}

int32_t nsContentList::IndexOf(nsIContent* aContent) {
  return IndexOf(aContent, true);
}

void nsContentList::NodeWillBeDestroyed(nsINode* aNode) {
  // We shouldn't do anything useful from now on

  RemoveFromCaches();
  mRootNode = nullptr;

  // We will get no more updates, so we can never know we're up to
  // date
  SetDirty();
}

void nsContentList::LastRelease() {
  RemoveFromCaches();
  if (mIsLiveList && mRootNode) {
    mRootNode->RemoveMutationObserver(this);
    mRootNode = nullptr;
  }
  SetDirty();
}

Element* nsContentList::GetElementAt(uint32_t aIndex) {
  return static_cast<Element*>(Item(aIndex, true));
}

nsIContent* nsContentList::Item(uint32_t aIndex) {
  return GetElementAt(aIndex);
}

void nsContentList::AttributeChanged(Element* aElement, int32_t aNameSpaceID,
                                     nsAtom* aAttribute, int32_t aModType,
                                     const nsAttrValue* aOldValue) {
  MOZ_ASSERT(aElement, "Must have a content node to work with");

  if (mState == State::Dirty ||
      !MayContainRelevantNodes(aElement->GetParentNode()) ||
      !nsContentUtils::IsInSameAnonymousTree(mRootNode, aElement)) {
    // Either we're already dirty or aElement will never match us.
    return;
  }

  InvalidateNamedItemsCacheForAttributeChange(aNameSpaceID, aAttribute);

  if (!mFunc || !mFuncMayDependOnAttr) {
    // aElement might be relevant but the attribute change doesn't affect
    // whether we match it.
    return;
  }

  if (Match(aElement)) {
    if (mElements.IndexOf(aElement) == mElements.NoIndex) {
      // We match aElement now, and it's not in our list already.  Just dirty
      // ourselves; this is simpler than trying to figure out where to insert
      // aElement.
      SetDirty();
    }
  } else {
    // We no longer match aElement.  Remove it from our list.  If it's
    // already not there, this is a no-op (though a potentially
    // expensive one).  Either way, no change of mState is required
    // here.
    if (mElements.RemoveElement(aElement)) {
      InvalidateNamedItemsCacheForDeletion(*aElement);
    }
  }
}

void nsContentList::ContentAppended(nsIContent* aFirstNewContent) {
  nsIContent* container = aFirstNewContent->GetParent();
  MOZ_ASSERT(container, "Can't get at the new content if no container!");

  /*
   * If the state is State::Dirty then we have no useful information in our list
   * and we want to put off doing work as much as possible.
   *
   * Also, if container is anonymous from our point of view, we know that we
   * can't possibly be matching any of the kids.
   *
   * Optimize out also the common case when just one new node is appended and
   * it doesn't match us.
   */
  if (mState == State::Dirty ||
      !nsContentUtils::IsInSameAnonymousTree(mRootNode, container) ||
      !MayContainRelevantNodes(container) ||
      (!aFirstNewContent->HasChildren() &&
       !aFirstNewContent->GetNextSibling() && !MatchSelf(aFirstNewContent))) {
    MaybeMarkDirty();
    return;
  }

  /*
   * We want to handle the case of ContentAppended by sometimes
   * appending the content to our list, not just setting state to
   * State::Dirty, since most of our ContentAppended notifications
   * should come during pageload and be at the end of the document.
   * Do a bit of work to see whether we could just append to what we
   * already have.
   */

  uint32_t ourCount = mElements.Length();
  const bool appendingToList = [&] {
    if (ourCount == 0) {
      return true;
    }
    if (mRootNode == container) {
      return true;
    }
    return nsContentUtils::PositionIsBefore(mElements.LastElement(),
                                            aFirstNewContent);
  }();

  if (!appendingToList) {
    // The new stuff is somewhere in the middle of our list; check
    // whether we need to invalidate
    for (nsIContent* cur = aFirstNewContent; cur; cur = cur->GetNextSibling()) {
      if (MatchSelf(cur)) {
        // Uh-oh.  We're gonna have to add elements into the middle
        // of our list. That's not worth the effort.
        SetDirty();
        break;
      }
    }

    ASSERT_IN_SYNC;
    return;
  }

  /*
   * At this point we know we could append.  If we're not up to
   * date, however, that would be a bad idea -- it could miss some
   * content that we never picked up due to being lazy.  Further, we
   * may never get asked for this content... so don't grab it yet.
   */
  if (mState == State::Lazy) {
    return;
  }

  /*
   * We're up to date.  That means someone's actively using us; we
   * may as well grab this content....
   */
  if (mDeep) {
    for (nsIContent* cur = aFirstNewContent; cur;
         cur = cur->GetNextNode(container)) {
      if (cur->IsElement() && Match(cur->AsElement())) {
        mElements.AppendElement(cur);
        InvalidateNamedItemsCacheForInsertion(*cur->AsElement());
      }
    }
  } else {
    for (nsIContent* cur = aFirstNewContent; cur; cur = cur->GetNextSibling()) {
      if (cur->IsElement() && Match(cur->AsElement())) {
        mElements.AppendElement(cur);
        InvalidateNamedItemsCacheForInsertion(*cur->AsElement());
      }
    }
  }

  ASSERT_IN_SYNC;
}

void nsContentList::ContentInserted(nsIContent* aChild) {
  // Note that aChild->GetParentNode() can be null here if we are inserting into
  // the document itself; any attempted optimizations to this method should deal
  // with that.
  if (mState != State::Dirty &&
      MayContainRelevantNodes(aChild->GetParentNode()) &&
      nsContentUtils::IsInSameAnonymousTree(mRootNode, aChild) &&
      MatchSelf(aChild)) {
    SetDirty();
  }

  ASSERT_IN_SYNC;
}

void nsContentList::ContentWillBeRemoved(nsIContent* aChild,
                                         const BatchRemovalState*) {
  if (mState != State::Dirty &&
      MayContainRelevantNodes(aChild->GetParentNode()) &&
      nsContentUtils::IsInSameAnonymousTree(mRootNode, aChild) &&
      MatchSelf(aChild)) {
    SetDirty();
  }

  ASSERT_IN_SYNC;
}

bool nsContentList::Match(Element* aElement) {
  if (mFunc) {
    return (*mFunc)(aElement, mMatchNameSpaceId, mXMLMatchAtom, mData);
  }

  if (!mXMLMatchAtom) return false;

  NodeInfo* ni = aElement->NodeInfo();

  bool unknown = mMatchNameSpaceId == kNameSpaceID_Unknown;
  bool wildcard = mMatchNameSpaceId == kNameSpaceID_Wildcard;
  bool toReturn = mMatchAll;
  if (!unknown && !wildcard) toReturn &= ni->NamespaceEquals(mMatchNameSpaceId);

  if (toReturn) return toReturn;

  bool matchHTML =
      mIsHTMLDocument && aElement->GetNameSpaceID() == kNameSpaceID_XHTML;

  if (unknown) {
    return matchHTML ? ni->QualifiedNameEquals(mHTMLMatchAtom)
                     : ni->QualifiedNameEquals(mXMLMatchAtom);
  }

  if (wildcard) {
    return matchHTML ? ni->Equals(mHTMLMatchAtom) : ni->Equals(mXMLMatchAtom);
  }

  return matchHTML ? ni->Equals(mHTMLMatchAtom, mMatchNameSpaceId)
                   : ni->Equals(mXMLMatchAtom, mMatchNameSpaceId);
}

bool nsContentList::MatchSelf(nsIContent* aContent) {
  MOZ_ASSERT(aContent, "Can't match null stuff, you know");
  MOZ_ASSERT(mDeep || aContent->GetParentNode() == mRootNode,
             "MatchSelf called on a node that we can't possibly match");

  if (!aContent->IsElement()) {
    return false;
  }

  if (Match(aContent->AsElement())) return true;

  if (!mDeep) return false;

  for (nsIContent* cur = aContent->GetFirstChild(); cur;
       cur = cur->GetNextNode(aContent)) {
    if (cur->IsElement() && Match(cur->AsElement())) {
      return true;
    }
  }

  return false;
}

void nsContentList::PopulateSelf(uint32_t aNeededLength,
                                 uint32_t aExpectedElementsIfDirty) {
  if (!mRootNode) {
    return;
  }

  ASSERT_IN_SYNC;

  uint32_t count = mElements.Length();
  NS_ASSERTION(mState != State::Dirty || count == aExpectedElementsIfDirty,
               "Reset() not called when setting state to State::Dirty?");

  if (count >= aNeededLength)  // We're all set
    return;

  uint32_t elementsToAppend = aNeededLength - count;
#ifdef DEBUG
  uint32_t invariant = elementsToAppend + mElements.Length();
#endif

  if (mDeep) {
    // If we already have nodes start searching at the last one, otherwise
    // start searching at the root.
    nsINode* cur = count ? mElements[count - 1].get() : mRootNode;
    do {
      cur = cur->GetNextNode(mRootNode);
      if (!cur) {
        break;
      }
      if (cur->IsElement() && Match(cur->AsElement())) {
        // Append AsElement() to get nsIContent instead of nsINode
        mElements.AppendElement(cur->AsElement());
        --elementsToAppend;
      }
    } while (elementsToAppend);
  } else {
    nsIContent* cur = count ? mElements[count - 1]->GetNextSibling()
                            : mRootNode->GetFirstChild();
    for (; cur && elementsToAppend; cur = cur->GetNextSibling()) {
      if (cur->IsElement() && Match(cur->AsElement())) {
        mElements.AppendElement(cur);
        --elementsToAppend;
      }
    }
  }

  NS_ASSERTION(elementsToAppend + mElements.Length() == invariant,
               "Something is awry!");

  if (elementsToAppend != 0) {
    mState = State::UpToDate;
  } else {
    mState = State::Lazy;
  }

  SetEnabledCallbacks(nsIMutationObserver::kAll);

  ASSERT_IN_SYNC;
}

void nsContentList::RemoveFromHashtable() {
  if (mFunc) {
    // nsCacheableFuncStringContentList can be in a hash table without being
    // in gContentListHashTable, but it will have been removed from the hash
    // table in its dtor before it runs the nsContentList dtor.
    MOZ_RELEASE_ASSERT(!mInHashtable);

    // This can't be in gContentListHashTable.
    return;
  }

  nsDependentAtomString str(mXMLMatchAtom);
  nsContentListKey key(mRootNode, mMatchNameSpaceId, str, mIsHTMLDocument);
  sRecentlyUsedContentLists.Remove(key);

  if (gContentListHashTable) {
    gContentListHashTable->RemoveEntry(&key);

    if (gContentListHashTable->Count() == 0) {
      gContentListHashTable = nullptr;
    }
  }

  MOZ_RELEASE_ASSERT(!mInHashtable);
}

void nsContentList::BringSelfUpToDate(bool aDoFlush) {
  if (mFlushesNeeded && mRootNode && aDoFlush) {
    // XXX sXBL/XBL2 issue
    if (Document* doc = mRootNode->GetUncomposedDoc()) {
      // Flush pending content changes Bug 4891.
      doc->FlushPendingNotifications(FlushType::ContentAndNotify);
    }
  }

  if (mState != State::UpToDate) {
    PopulateSelf(uint32_t(-1));
  }

  mMissedUpdates = 0;

  ASSERT_IN_SYNC;
  NS_ASSERTION(!mRootNode || mState == State::UpToDate,
               "PopulateSelf dod not bring content list up to date!");
}

nsCacheableFuncStringContentList::~nsCacheableFuncStringContentList() {
  RemoveFromFuncStringHashtable();
}

void nsCacheableFuncStringContentList::RemoveFromFuncStringHashtable() {
  if (!gFuncStringContentListHashTable) {
    MOZ_RELEASE_ASSERT(!mInHashtable);
    return;
  }

  nsFuncStringCacheKey key(mRootNode, mFunc, mString);
  gFuncStringContentListHashTable->RemoveEntry(&key);

  if (gFuncStringContentListHashTable->Count() == 0) {
    gFuncStringContentListHashTable = nullptr;
  }

  MOZ_RELEASE_ASSERT(!mInHashtable);
}

#ifdef DEBUG_CONTENT_LIST
void nsContentList::AssertInSync() {
  if (mState == State::Dirty) {
    return;
  }

  if (!mRootNode) {
    NS_ASSERTION(mElements.Length() == 0 && mState == State::Dirty,
                 "Empty iterator isn't quite empty?");
    return;
  }

  // XXX This code will need to change if nsContentLists can ever match
  // elements that are outside of the document element.
  nsIContent* root = mRootNode->IsDocument()
                         ? mRootNode->AsDocument()->GetRootElement()
                         : mRootNode->AsContent();

  PreContentIterator preOrderIter;
  if (mDeep) {
    preOrderIter.Init(root);
    preOrderIter.First();
  }

  uint32_t cnt = 0, index = 0;
  while (true) {
    if (cnt == mElements.Length() && mState == State::Lazy) {
      break;
    }

    nsIContent* cur =
        mDeep ? preOrderIter.GetCurrentNode() : mRootNode->GetChildAt(index++);
    if (!cur) {
      break;
    }

    if (cur->IsElement() && Match(cur->AsElement())) {
      NS_ASSERTION(cnt < mElements.Length() && mElements[cnt] == cur,
                   "Elements is out of sync");
      ++cnt;
    }

    if (mDeep) {
      preOrderIter.Next();
    }
  }

  NS_ASSERTION(cnt == mElements.Length(), "Too few elements");
}
#endif

//-----------------------------------------------------
// nsCachableElementsByNameNodeList

JSObject* nsCachableElementsByNameNodeList::WrapObject(
    JSContext* cx, JS::Handle<JSObject*> aGivenProto) {
  return NodeList_Binding::Wrap(cx, this, aGivenProto);
}

void nsCachableElementsByNameNodeList::AttributeChanged(
    Element* aElement, int32_t aNameSpaceID, nsAtom* aAttribute,
    int32_t aModType, const nsAttrValue* aOldValue) {
  // No need to rebuild the list if the changed attribute is not the name
  // attribute.
  if (aAttribute != nsGkAtoms::name) {
    InvalidateNamedItemsCacheForAttributeChange(aNameSpaceID, aAttribute);
    return;
  }

  nsCacheableFuncStringContentList::AttributeChanged(
      aElement, aNameSpaceID, aAttribute, aModType, aOldValue);
}

//-----------------------------------------------------
// nsCacheableFuncStringHTMLCollection

JSObject* nsCacheableFuncStringHTMLCollection::WrapObject(
    JSContext* cx, JS::Handle<JSObject*> aGivenProto) {
  return HTMLCollection_Binding::Wrap(cx, this, aGivenProto);
}

//-----------------------------------------------------
// nsLabelsNodeList

JSObject* nsLabelsNodeList::WrapObject(JSContext* cx,
                                       JS::Handle<JSObject*> aGivenProto) {
  return NodeList_Binding::Wrap(cx, this, aGivenProto);
}

void nsLabelsNodeList::AttributeChanged(Element* aElement, int32_t aNameSpaceID,
                                        nsAtom* aAttribute, int32_t aModType,
                                        const nsAttrValue* aOldValue) {
  MOZ_ASSERT(aElement, "Must have a content node to work with");
  if (mState == State::Dirty ||
      !nsContentUtils::IsInSameAnonymousTree(mRootNode, aElement)) {
    return;
  }

  InvalidateNamedItemsCacheForAttributeChange(aNameSpaceID, aAttribute);

  // We need to handle input type changes to or from "hidden".
  if (aElement->IsHTMLElement(nsGkAtoms::input) &&
      aAttribute == nsGkAtoms::type && aNameSpaceID == kNameSpaceID_None) {
    SetDirty();
    return;
  }
}

void nsLabelsNodeList::ContentAppended(nsIContent* aFirstNewContent) {
  nsIContent* container = aFirstNewContent->GetParent();
  // If a labelable element is moved to outside or inside of
  // nested associated labels, we're gonna have to modify
  // the content list.
  if (mState != State::Dirty &&
      nsContentUtils::IsInSameAnonymousTree(mRootNode, container)) {
    SetDirty();
    return;
  }
}

void nsLabelsNodeList::ContentInserted(nsIContent* aChild) {
  // If a labelable element is moved to outside or inside of
  // nested associated labels, we're gonna have to modify
  // the content list.
  if (mState != State::Dirty &&
      nsContentUtils::IsInSameAnonymousTree(mRootNode, aChild)) {
    SetDirty();
    return;
  }
}

void nsLabelsNodeList::ContentWillBeRemoved(nsIContent* aChild,
                                            const BatchRemovalState* aState) {
  // If a labelable element is removed, we're gonna have to clean
  // the content list.
  if (mState != State::Dirty &&
      nsContentUtils::IsInSameAnonymousTree(mRootNode, aChild)) {
    SetDirty();
    return;
  }
}

void nsLabelsNodeList::MaybeResetRoot(nsINode* aRootNode) {
  MOZ_ASSERT(aRootNode, "Must have root");
  if (mRootNode == aRootNode) {
    return;
  }

  MOZ_ASSERT(mIsLiveList, "nsLabelsNodeList is always a live list");
  if (mRootNode) {
    mRootNode->RemoveMutationObserver(this);
  }
  mRootNode = aRootNode;
  mRootNode->AddMutationObserver(this);
  SetDirty();
}

void nsLabelsNodeList::PopulateSelf(uint32_t aNeededLength,
                                    uint32_t aExpectedElementsIfDirty) {
  if (!mRootNode) {
    return;
  }

  // Start searching at the root.
  nsINode* cur = mRootNode;
  if (mElements.IsEmpty() && cur->IsElement() && Match(cur->AsElement())) {
    mElements.AppendElement(cur->AsElement());
    ++aExpectedElementsIfDirty;
  }

  nsContentList::PopulateSelf(aNeededLength, aExpectedElementsIfDirty);
}

quality94%

¤ Diese beiden folgenden Angebotsgruppen bietet das Unternehmen0.23Angebot Wie Sie bei der Firma Beratungs- und Dienstleistungen beauftragen können ¤

*Eine klare Vorstellung vom Zielzustand

Wurzel

Suchen

Beweissystem der NASA

Beweissystem Isabelle

NIST Cobol Testsuite

Cephes Mathematical Library

Wiener Entwicklungsmethode

Haftungshinweis

Die Informationen auf dieser Webseite wurden nach bestem Wissen sorgfältig zusammengestellt. Es wird jedoch weder Vollständigkeit, noch Richtigkeit, noch Qualität der bereit gestellten Informationen zugesichert.

Bemerkung:

Die farbliche Syntaxdarstellung ist noch experimentell.