Lettore 0.4.0: rifatto senza WebView (Text rich text QML)

Il WebView di Sailfish non supporta lo scroll programmatico affidabile
(window.scrollY stale, contentHeight/scrollPosition non scrivibili): 0.3.x
non ha mai cambiato capitolo in modo affidabile. Ora i capitoli sono rich
text Qt in un Text QML dentro un SilicaFlickable: contentY/contentHeight
nativi, paginazione a tap deterministico, fine capitolo esatta. HTML->rich
text in C++ (tag base, entità, fallback). Bookmark v2 invariato (offset
caratteri stimato da frazione). Niente immagini inline (limite rich text).
This commit is contained in:
2026-08-21 11:32:43 +02:00
parent a5088c3986
commit 60f53ab4e4
5 changed files with 414 additions and 280 deletions
+239
View File
@@ -193,6 +193,234 @@ int plainTextLength(const QString &html)
return count;
}
// ---------------------------------------------------------------------------
// HTML -> rich text Qt (per Text QML, niente WebView): tag di base convertiti,
// blocchi raw (script/style/head/...) saltati, immagini ignorate, i tag non
// supportati diventano trasparenti (resta il contenuto).
// ---------------------------------------------------------------------------
// Entità HTML nominate NON standard XML (QXmlStreamReader le rifiuta senza DTD):
// decodificate prima del parse. Quelle XML standard (amp lt gt quot apos) e le
// numeriche (&#NNN; &#xHH;) le gestisce il parser.
QString decodeHtmlEntities(const QString &s)
{
static const QHash<QString, QChar> entities = [] {
QHash<QString, QChar> h;
h.insert(QStringLiteral("nbsp"), QChar(0x00A0));
h.insert(QStringLiteral("mdash"), QChar(0x2014));
h.insert(QStringLiteral("ndash"), QChar(0x2013));
h.insert(QStringLiteral("rsquo"), QChar(0x2019));
h.insert(QStringLiteral("lsquo"), QChar(0x2018));
h.insert(QStringLiteral("rdquo"), QChar(0x201D));
h.insert(QStringLiteral("ldquo"), QChar(0x201C));
h.insert(QStringLiteral("sbquo"), QChar(0x201A));
h.insert(QStringLiteral("bdquo"), QChar(0x201E));
h.insert(QStringLiteral("hellip"), QChar(0x2026));
h.insert(QStringLiteral("dagger"), QChar(0x2020));
h.insert(QStringLiteral("Dagger"), QChar(0x2021));
h.insert(QStringLiteral("bull"), QChar(0x2022));
h.insert(QStringLiteral("middot"), QChar(0x00B7));
h.insert(QStringLiteral("eacute"), QChar(0x00E9));
h.insert(QStringLiteral("egrave"), QChar(0x00E8));
h.insert(QStringLiteral("agrave"), QChar(0x00E0));
h.insert(QStringLiteral("ugrave"), QChar(0x00F9));
h.insert(QStringLiteral("igrave"), QChar(0x00EC));
h.insert(QStringLiteral("ograve"), QChar(0x00F2));
h.insert(QStringLiteral("aacute"), QChar(0x00E1));
h.insert(QStringLiteral("uacute"), QChar(0x00FA));
h.insert(QStringLiteral("iacute"), QChar(0x00ED));
h.insert(QStringLiteral("oacute"), QChar(0x00F3));
h.insert(QStringLiteral("auml"), QChar(0x00E4));
h.insert(QStringLiteral("euml"), QChar(0x00EB));
h.insert(QStringLiteral("iuml"), QChar(0x00EF));
h.insert(QStringLiteral("ouml"), QChar(0x00F6));
h.insert(QStringLiteral("uuml"), QChar(0x00FC));
h.insert(QStringLiteral("szlig"), QChar(0x00DF));
h.insert(QStringLiteral("deg"), QChar(0x00B0));
h.insert(QStringLiteral("plusmn"), QChar(0x00B1));
h.insert(QStringLiteral("times"), QChar(0x00D7));
h.insert(QStringLiteral("divide"), QChar(0x00F7));
h.insert(QStringLiteral("copy"), QChar(0x00A9));
h.insert(QStringLiteral("reg"), QChar(0x00AE));
h.insert(QStringLiteral("trade"), QChar(0x2122));
h.insert(QStringLiteral("euro"), QChar(0x20AC));
h.insert(QStringLiteral("pound"), QChar(0x00A3));
h.insert(QStringLiteral("cent"), QChar(0x00A2));
h.insert(QStringLiteral("yen"), QChar(0x00A5));
h.insert(QStringLiteral("para"), QChar(0x00B6));
h.insert(QStringLiteral("sect"), QChar(0x00A7));
h.insert(QStringLiteral("laquo"), QChar(0x00AB));
h.insert(QStringLiteral("raquo"), QChar(0x00BB));
h.insert(QStringLiteral("lsaquo"), QChar(0x2039));
h.insert(QStringLiteral("rsaquo"), QChar(0x203A));
h.insert(QStringLiteral("prime"), QChar(0x2032));
h.insert(QStringLiteral("Prime"), QChar(0x2033));
h.insert(QStringLiteral("permil"), QChar(0x2030));
return h;
}();
QString out;
out.reserve(s.size());
int i = 0;
const int n = s.size();
while (i < n) {
const QChar c = s.at(i);
if (c == QLatin1Char('&')) {
const int semi = s.indexOf(QLatin1Char(';'), i + 1);
if (semi > i && semi - i <= 12) {
const QString ent = s.mid(i + 1, semi - i - 1);
const QChar ch = entities.value(ent);
if (!ch.isNull()) {
out += ch;
i = semi + 1;
continue;
}
}
}
out += c;
++i;
}
return out;
}
QString htmlToRichText(const QString &html)
{
const int bodyStart = html.indexOf(QStringLiteral("<body"), 0, Qt::CaseInsensitive);
const int bodyEnd = html.indexOf(QStringLiteral("</body"), 0, Qt::CaseInsensitive);
QString scope;
if (bodyStart >= 0) {
// includi anche la chiusura </body>: il frammento deve essere XML valido
const int end = (bodyEnd >= bodyStart) ? bodyEnd + 7 : html.size();
scope = html.mid(bodyStart, end - bodyStart);
} else {
scope = html;
}
scope = decodeHtmlEntities(scope);
QString out;
QStringList stack;
QXmlStreamReader r(scope);
while (!r.atEnd() && !r.hasError()) {
const QXmlStreamReader::TokenType t = r.readNext();
if (t == QXmlStreamReader::StartElement) {
const QString name = r.name().toString().toLower();
if (name == QLatin1String("script") || name == QLatin1String("style")
|| name == QLatin1String("noscript") || name == QLatin1String("head")
|| name == QLatin1String("title")) {
r.skipCurrentElement();
continue;
}
if (name == QLatin1String("br")) {
out += QStringLiteral("<br>");
stack.append(QString());
continue;
}
if (name == QLatin1String("img") || name == QLatin1String("hr")
|| name == QLatin1String("input")) {
stack.append(QString());
continue;
}
QString tag;
if (name == QLatin1String("b") || name == QLatin1String("strong"))
tag = QStringLiteral("b");
else if (name == QLatin1String("i") || name == QLatin1String("em"))
tag = QStringLiteral("i");
else if (name == QLatin1String("u"))
tag = QStringLiteral("u");
else if (name == QLatin1String("s") || name == QLatin1String("strike")
|| name == QLatin1String("del"))
tag = QStringLiteral("s");
else if (name == QLatin1String("ul"))
tag = QStringLiteral("ul");
else if (name == QLatin1String("ol"))
tag = QStringLiteral("ol");
else if (name == QLatin1String("li"))
tag = QStringLiteral("li");
else if (name == QLatin1String("table"))
tag = QStringLiteral("table border=\"1\" cellspacing=\"0\" cellpadding=\"4\"");
else if (name == QLatin1String("tr"))
tag = QStringLiteral("tr");
else if (name == QLatin1String("td") || name == QLatin1String("th"))
tag = QStringLiteral("td");
else if (name == QLatin1String("p") || name == QLatin1String("blockquote"))
tag = QStringLiteral("p");
else if (name == QLatin1String("h1") || name == QLatin1String("h2")
|| name == QLatin1String("h3") || name == QLatin1String("h4")
|| name == QLatin1String("h5") || name == QLatin1String("h6"))
tag = name;
// div/span/a/section/... -> trasparenti (tag vuoto)
if (!tag.isEmpty()) {
out += QLatin1Char('<') + tag + QLatin1Char('>');
stack.append(tag);
} else {
stack.append(QString());
}
} else if (t == QXmlStreamReader::EndElement) {
if (!stack.isEmpty()) {
const QString tag = stack.takeLast();
if (!tag.isEmpty()) {
const int sp = tag.indexOf(QLatin1Char(' '));
out += QStringLiteral("</") + (sp > 0 ? tag.left(sp) : tag) + QLatin1Char('>');
}
}
} else if (t == QXmlStreamReader::Characters) {
out += r.text();
}
}
// HTML non ben formato (parse fallito): fallback grezzo — testo con <br>
// sui newline e paragrafi sui doppi newline
if (r.hasError()) {
out.clear();
int i = 0;
const int n = scope.size();
while (i < n) {
const QChar c = scope.at(i);
if (c == QLatin1Char('<')) {
const int close = scope.indexOf(QLatin1Char('>'), i + 1);
i = (close < 0) ? n : close + 1;
continue;
}
if (c == QLatin1Char('&')) {
const int semi = scope.indexOf(QLatin1Char(';'), i + 1);
if (semi > i && semi - i <= 10) {
const QString ent = scope.mid(i + 1, semi - i - 1);
if (ent == QLatin1String("amp")) out += QLatin1Char('&');
else if (ent == QLatin1String("lt")) out += QLatin1Char('<');
else if (ent == QLatin1String("gt")) out += QLatin1Char('>');
else if (ent == QLatin1String("nbsp")) out += QLatin1Char(' ');
else if (ent.startsWith(QLatin1String("#"))) {
bool ok = false;
const int cp = ent.mid(1).toInt(&ok);
if (ok && cp > 0 && cp < 0x10FFFF)
out += QChar(cp);
} else {
out += c;
out += ent;
out += QLatin1Char(';');
}
i = semi + 1;
continue;
}
}
if (c == QLatin1Char('\n')) {
// doppio newline = paragrafo, singolo = a capo
if (i + 1 < n && scope.at(i + 1) == QLatin1Char('\n')) {
out += QStringLiteral("</p><p>");
++i;
} else {
out += QStringLiteral("<br>");
}
++i;
continue;
}
out += c;
++i;
}
}
return out;
}
} // namespace
// ---------------------------------------------------------------------------
@@ -524,6 +752,17 @@ QString EpubBook::chapterBaseUrl(int index) const
.toString();
}
QString EpubBook::chapterRichText(int index) const
{
if (index < 0 || index >= m_chapters.size())
return QString();
QFile f(m_chapters.at(index).path);
if (!f.open(QIODevice::ReadOnly))
return QString();
const QString html = QString::fromUtf8(f.readAll());
return htmlToRichText(html);
}
void EpubBook::computeTextLengths()
{
m_textLengths.clear();