qtextboundaryfinder.cpp

Go to the documentation of this file.

// Copyright (C) 2022 The Qt Company Ltd.
// SPDX-License-Identifier: LicenseRef-Qt-Commercial OR LGPL-3.0-only OR GPL-2.0-only OR GPL-3.0-only
#include <QtCore/qtextboundaryfinder.h>
#include <QtCore/qvarlengtharray.h>
 
#include <private/qunicodetools_p.h>
 
QT_BEGIN_NAMESPACE
 
static void init(QTextBoundaryFinder::BoundaryType type, QStringView str, QCharAttributes *attributes)
{
    QUnicodeTools::ScriptItemArray scriptItems;
    QUnicodeTools::initScripts(str, &scriptItems);
 
    QUnicodeTools::CharAttributeOptions options;
    switch (type) {
    case QTextBoundaryFinder::Grapheme: options |= QUnicodeTools::GraphemeBreaks; break;
    case QTextBoundaryFinder::Word: options |= QUnicodeTools::WordBreaks; break;
    case QTextBoundaryFinder::Sentence: options |= QUnicodeTools::SentenceBreaks; break;
    case QTextBoundaryFinder::Line: options |= QUnicodeTools::LineBreaks; break;
    default: break;
    }
    QUnicodeTools::initCharAttributes(str, scriptItems.data(), scriptItems.size(), attributes, options);
}
 
QTextBoundaryFinder::QTextBoundaryFinder()
    : freeBuffer(true)
{
}
 
QTextBoundaryFinder::QTextBoundaryFinder(const QTextBoundaryFinder &other)
    : t(other.t)
    , s(other.s)
    , sv(other.sv)
    , pos(other.pos)
    , freeBuffer(true)
{
    if (other.attributes) {
        Q_ASSERT(sv.size() > 0);
        attributes = (QCharAttributes *) malloc((sv.size() + 1) * sizeof(QCharAttributes));
        Q_CHECK_PTR(attributes);
        memcpy(attributes, other.attributes, (sv.size() + 1) * sizeof(QCharAttributes));
    }
}
 
QTextBoundaryFinder &QTextBoundaryFinder::operator=(const QTextBoundaryFinder &other)
{
    if (&other == this)
        return *this;
 
    if (other.attributes) {
        Q_ASSERT(other.sv.size() > 0);
        size_t newCapacity = (size_t(other.sv.size()) + 1) * sizeof(QCharAttributes);
        QCharAttributes *newD = (QCharAttributes *) realloc(freeBuffer ? attributes : nullptr, newCapacity);
        Q_CHECK_PTR(newD);
        freeBuffer = true;
        attributes = newD;
    }
 
    t = other.t;
    s = other.s;
    sv = other.sv;
    pos = other.pos;
 
    if (other.attributes) {
        memcpy(attributes, other.attributes, (sv.size() + 1) * sizeof(QCharAttributes));
    } else {
        if (freeBuffer)
            free(attributes);
        attributes = nullptr;
    }
 
    return *this;
}
 
QTextBoundaryFinder::~QTextBoundaryFinder()
{
    Q_UNUSED(unused);
    if (freeBuffer)
        free(attributes);
}
 
QTextBoundaryFinder::QTextBoundaryFinder(BoundaryType type, const QString &string)
    : t(type)
    , s(string)
    , sv(s)
    , freeBuffer(true)
{
    if (sv.size() > 0) {
        attributes = (QCharAttributes *) malloc((sv.size() + 1) * sizeof(QCharAttributes));
        Q_CHECK_PTR(attributes);
        init(t, sv, attributes);
    }
}
 
QTextBoundaryFinder::QTextBoundaryFinder(BoundaryType type, QStringView string, unsigned char *buffer, qsizetype bufferSize)
    : t(type)
    , sv(string)
    , freeBuffer(true)
{
    if (!sv.isEmpty()) {
        if (buffer && bufferSize / int(sizeof(QCharAttributes)) >= sv.size() + 1) {
            attributes = reinterpret_cast<QCharAttributes *>(buffer);
            freeBuffer = false;
        } else {
            attributes = (QCharAttributes *) malloc((sv.size() + 1) * sizeof(QCharAttributes));
            Q_CHECK_PTR(attributes);
        }
        init(t, sv, attributes);
    }
}
 
void QTextBoundaryFinder::toStart()
{
    pos = 0;
}
 
void QTextBoundaryFinder::toEnd()
{
    pos = sv.size();
}
 
qsizetype QTextBoundaryFinder::position() const
{
    return pos;
}
 
void QTextBoundaryFinder::setPosition(qsizetype position)
{
    pos = qBound(0, position, sv.size());
}
 
QString QTextBoundaryFinder::string() const
{
    if (sv.data() == s.unicode() && sv.size() == s.size())
        return s;
    return sv.toString();
}
 
 
qsizetype QTextBoundaryFinder::toNextBoundary()
{
    if (!attributes || pos < 0 || pos >= sv.size()) {
        pos = -1;
        return pos;
    }
 
    ++pos;
    switch(t) {
    case Grapheme:
        while (pos < sv.size() && !attributes[pos].graphemeBoundary)
            ++pos;
        break;
    case Word:
        while (pos < sv.size() && !attributes[pos].wordBreak)
            ++pos;
        break;
    case Sentence:
        while (pos < sv.size() && !attributes[pos].sentenceBoundary)
            ++pos;
        break;
    case Line:
        while (pos < sv.size() && !attributes[pos].lineBreak)
            ++pos;
        break;
    }
 
    return pos;
}
 
qsizetype QTextBoundaryFinder::toPreviousBoundary()
{
    if (!attributes || pos <= 0 || pos > sv.size()) {
        pos = -1;
        return pos;
    }
 
    --pos;
    switch(t) {
    case Grapheme:
        while (pos > 0 && !attributes[pos].graphemeBoundary)
            --pos;
        break;
    case Word:
        while (pos > 0 && !attributes[pos].wordBreak)
            --pos;
        break;
    case Sentence:
        while (pos > 0 && !attributes[pos].sentenceBoundary)
            --pos;
        break;
    case Line:
        while (pos > 0 && !attributes[pos].lineBreak)
            --pos;
        break;
    }
 
    return pos;
}
 
bool QTextBoundaryFinder::isAtBoundary() const
{
    if (!attributes || pos < 0 || pos > sv.size())
        return false;
 
    switch(t) {
    case Grapheme:
        return attributes[pos].graphemeBoundary;
    case Word:
        return attributes[pos].wordBreak;
    case Sentence:
        return attributes[pos].sentenceBoundary;
    case Line:
        // ### TR#14 LB2 prohibits break at sot
        return attributes[pos].lineBreak || pos == 0;
    }
    return false;
}
 
QTextBoundaryFinder::BoundaryReasons QTextBoundaryFinder::boundaryReasons() const
{
    BoundaryReasons reasons = NotAtBoundary;
    if (!attributes || pos < 0 || pos > sv.size())
        return reasons;
 
    const QCharAttributes attr = attributes[pos];
    switch (t) {
    case Grapheme:
        if (attr.graphemeBoundary) {
            reasons |= BreakOpportunity | StartOfItem | EndOfItem;
            if (pos == 0)
                reasons &= (~EndOfItem);
            else if (pos == sv.size())
                reasons &= (~StartOfItem);
        }
        break;
    case Word:
        if (attr.wordBreak) {
            reasons |= BreakOpportunity;
            if (attr.wordStart)
                reasons |= StartOfItem;
            if (attr.wordEnd)
                reasons |= EndOfItem;
        }
        break;
    case Sentence:
        if (attr.sentenceBoundary) {
            reasons |= BreakOpportunity | StartOfItem | EndOfItem;
            if (pos == 0)
                reasons &= (~EndOfItem);
            else if (pos == sv.size())
                reasons &= (~StartOfItem);
        }
        break;
    case Line:
        // ### TR#14 LB2 prohibits break at sot
        if (attr.lineBreak || pos == 0) {
            reasons |= BreakOpportunity;
            if (attr.mandatoryBreak || pos == 0) {
                reasons |= MandatoryBreak | StartOfItem | EndOfItem;
                if (pos == 0)
                    reasons &= (~EndOfItem);
                else if (pos == sv.size())
                    reasons &= (~StartOfItem);
            } else if (pos > 0 && sv[pos - 1].unicode() == QChar::SoftHyphen) {
                reasons |= SoftHyphen;
            }
        }
        break;
    default:
        break;
    }
 
    return reasons;
}
 
QT_END_NAMESPACE