зеркало из https://github.com/mozilla/gecko-dev.git
Bug 1706613: part 2) Change `NS_ASSERTION`s to `MOZ_ASSERT`s in `mozInlineSpellWordUtil`. r=smaug
Depends on D112958 Differential Revision: https://phabricator.services.mozilla.com/D112959
This commit is contained in:
Родитель
d45f47b9a0
Коммит
ff55282cb1
|
@ -167,8 +167,8 @@ static nsINode* FindNextNode(nsINode* aNode, const nsINode* aRoot,
|
|||
static nsINode* FindNextTextNode(nsINode* aNode, int32_t aOffset,
|
||||
const nsINode* aRoot) {
|
||||
MOZ_ASSERT(aNode, "Null starting node?");
|
||||
NS_ASSERTION(!IsSpellCheckingTextNode(aNode),
|
||||
"FindNextTextNode should start with a non-text node");
|
||||
MOZ_ASSERT(!IsSpellCheckingTextNode(aNode),
|
||||
"FindNextTextNode should start with a non-text node");
|
||||
|
||||
nsINode* checkNode;
|
||||
// Need to start at the aOffset'th child
|
||||
|
@ -217,7 +217,7 @@ nsresult mozInlineSpellWordUtil::SetPositionAndEnd(nsINode* aPositionNode,
|
|||
MOZ_ASSERT(aPositionNode, "Null begin node?");
|
||||
MOZ_ASSERT(aEndNode, "Null end node?");
|
||||
|
||||
NS_ASSERTION(mRootNode, "Not initialized");
|
||||
MOZ_ASSERT(mRootNode, "Not initialized");
|
||||
|
||||
// Find a appropriate root if we are dealing with contenteditable nodes which
|
||||
// are in the shadow DOM.
|
||||
|
@ -453,8 +453,8 @@ struct MOZ_STACK_CLASS WordSplitState {
|
|||
template <class T>
|
||||
CharClass WordSplitState<T>::ClassifyCharacter(int32_t aIndex,
|
||||
bool aRecurse) const {
|
||||
NS_ASSERTION(aIndex >= 0 && aIndex <= int32_t(mDOMWordText.Length()),
|
||||
"Index out of range");
|
||||
MOZ_ASSERT(aIndex >= 0 && aIndex <= int32_t(mDOMWordText.Length()),
|
||||
"Index out of range");
|
||||
if (aIndex == int32_t(mDOMWordText.Length())) return CHAR_CLASS_SEPARATOR;
|
||||
|
||||
// this will classify the character, we want to treat "ignorable" characters
|
||||
|
@ -534,9 +534,9 @@ CharClass WordSplitState<T>::ClassifyCharacter(int32_t aIndex,
|
|||
// WordSplitState::Advance
|
||||
template <class T>
|
||||
void WordSplitState<T>::Advance() {
|
||||
NS_ASSERTION(mDOMWordOffset >= 0, "Negative word index");
|
||||
NS_ASSERTION(mDOMWordOffset < (int32_t)mDOMWordText.Length(),
|
||||
"Length beyond end");
|
||||
MOZ_ASSERT(mDOMWordOffset >= 0, "Negative word index");
|
||||
MOZ_ASSERT(mDOMWordOffset < (int32_t)mDOMWordText.Length(),
|
||||
"Length beyond end");
|
||||
|
||||
mDOMWordOffset++;
|
||||
if (mDOMWordOffset >= (int32_t)mDOMWordText.Length())
|
||||
|
@ -682,7 +682,7 @@ static inline bool IsBRElement(nsINode* aNode) {
|
|||
static Maybe<int32_t> FindOffsetOfLastDOMWordSeparatorSequence(
|
||||
nsIContent* aContent, int32_t aBeforeOffset) {
|
||||
const nsTextFragment* textFragment = aContent->GetText();
|
||||
NS_ASSERTION(textFragment, "Where is our text?");
|
||||
MOZ_ASSERT(textFragment, "Where is our text?");
|
||||
int32_t end = std::min(aBeforeOffset, int32_t(textFragment->GetLength()));
|
||||
|
||||
if (textFragment->Is2b()) {
|
||||
|
@ -853,9 +853,9 @@ void mozInlineSpellWordUtil::SoftText::AdjustBeginAndBuildText(
|
|||
bool exit = false;
|
||||
if (IsSpellCheckingTextNode(node)) {
|
||||
nsIContent* content = static_cast<nsIContent*>(node);
|
||||
NS_ASSERTION(content, "Where is our content?");
|
||||
MOZ_ASSERT(content, "Where is our content?");
|
||||
const nsTextFragment* textFragment = content->GetText();
|
||||
NS_ASSERTION(textFragment, "Where is our text?");
|
||||
MOZ_ASSERT(textFragment, "Where is our text?");
|
||||
int32_t lastOffsetInNode = textFragment->GetLength();
|
||||
|
||||
if (seenSoftEnd) {
|
||||
|
@ -1006,8 +1006,8 @@ bool FindLastNongreaterOffset(const nsTArray<T>& aContainer,
|
|||
|
||||
NodeOffset mozInlineSpellWordUtil::MapSoftTextOffsetToDOMPosition(
|
||||
int32_t aSoftTextOffset, DOMMapHint aHint) const {
|
||||
NS_ASSERTION(mSoftText.mIsValid,
|
||||
"Soft text must be valid if we're to map out of it");
|
||||
MOZ_ASSERT(mSoftText.mIsValid,
|
||||
"Soft text must be valid if we're to map out of it");
|
||||
if (!mSoftText.mIsValid) return NodeOffset(nullptr, -1);
|
||||
|
||||
// Find the last mapping, if any, such that mSoftTextOffset <= aSoftTextOffset
|
||||
|
@ -1065,8 +1065,8 @@ int32_t mozInlineSpellWordUtil::FindRealWordContaining(
|
|||
aSoftTextOffset, hint.get(), static_cast<int32_t>(aSearchForward)));
|
||||
}
|
||||
|
||||
NS_ASSERTION(mSoftText.mIsValid,
|
||||
"Soft text must be valid if we're to map out of it");
|
||||
MOZ_ASSERT(mSoftText.mIsValid,
|
||||
"Soft text must be valid if we're to map out of it");
|
||||
if (!mSoftText.mIsValid) return -1;
|
||||
|
||||
// Find the last word, if any, such that mSoftTextOffset <= aSoftTextOffset
|
||||
|
|
Загрузка…
Ссылка в новой задаче