src/Composer/PlainTextFormatter.cpp

   1 /* Copyright (C) 2012 Thomas Lübking <thomas.luebking@gmail.com>
   2    Copyright (C) 2006 - 2013 Jan Kundrát <jkt@flaska.net>
   3
   4    This file is part of the Trojita Qt IMAP e-mail client,
   5    http://trojita.flaska.net/
   6
   7    This program is free software; you can redistribute it and/or
   8    modify it under the terms of the GNU General Public License as
   9    published by the Free Software Foundation; either version 2 of
  10    the License or (at your option) version 3 or any later version
  11    accepted by the membership of KDE e.V. (or its successor approved
  12    by the membership of KDE e.V.), which shall act as a proxy
  13    defined in Section 14 of version 3 of the license.
  14
  15    This program is distributed in the hope that it will be useful,
  16    but WITHOUT ANY WARRANTY; without even the implied warranty of
  17    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18    GNU General Public License for more details.
  19
  20    You should have received a copy of the GNU General Public License
  21    along with this program.  If not, see <http://www.gnu.org/licenses/>.
  22 */
  23
  24 #include <QObject>
  25 #if QT_VERSION < QT_VERSION_CHECK(5, 0, 0)
  26 #include <QTextDocument>
  27 #endif
  28 #include "PlainTextFormatter.h"
  29
  30 #include <QDebug> // FIXME: remove me
  31
  32 namespace Composer {
  33 namespace Util {
  34
  35 /** @short Helper for plainTextToHtml for applying the HTML formatting
  36
  37 This funciton recognizes http and https links, e-mail addresses, *bold*, /italic/ and _underline_ text.
  38 */
  39 QString helperHtmlifySingleLine(QString line)
  40 {
  41     // Static regexps for the engine construction
  42     static const QRegExp linkRe("("
  43                               "https?://" // scheme prefix
  44                               "[;/?:@=&$\\-_.+!',0-9a-zA-Z%#~\\[\\]\\(\\)*]+" // allowed characters
  45                               "[/@=&$\\-_+'0-9a-zA-Z%#~]" // termination
  46                               ")");
  47     static const QRegExp mailRe("([a-zA-Z0-9\\.\\-_\\+]+@[a-zA-Z0-9\\.\\-_]+)");
  48     static QString intro("(^|[\\s\\(\\[\\{])");
  49     static QString extro("($|[\\s\\),;.\\]\\}])");
  50 #define TROJITA_RE_BOLD "\\*(\\S*)\\*"
  51 #define TROJITA_RE_ITALIC "/(\\S*)/"
  52 #define TROJITA_RE_UNDERLINE "_(\\S*)_"
  53     static const QRegExp boldRe(intro + TROJITA_RE_BOLD + extro);
  54     static const QRegExp italicRe(intro + TROJITA_RE_ITALIC + extro);
  55     static const QRegExp underlineRe(intro + TROJITA_RE_UNDERLINE + extro);
  56     static const QRegExp anyFormattingRe(intro + "(" TROJITA_RE_BOLD "|" TROJITA_RE_ITALIC "|" TROJITA_RE_UNDERLINE ")" + extro);
  57 #undef TROJITA_RE_BOLD
  58 #undef TROJITA_RE_ITALIC
  59 #undef TROJITA_RE_UNDERLINE
  60
  61     // RE instances to work on
  62     QRegExp link(linkRe), mail(mailRe), bold(boldRe), italic(italicRe), underline(underlineRe), anyFormatting(anyFormattingRe);
  63
  64     // Now prepare markup *bold*, /italic/ and _underline_ and also turn links into HTML.
  65     // This is a bit more involved because we want to apply the regular expressions in a certain order and also at the same
  66     // time prevent the lower-priority regexps from clobbering the output of the previous stages.
  67     int start = 0;
  68     while (start < line.size()) {
  69         qDebug() << "Main loop:" << start << line.size() << line;
  70         // Find the position of the first thing which matches
  71         int posLink = link.indexIn(line, start, QRegExp::CaretAtOffset);
  72         if (posLink == -1)
  73             posLink = line.size();
  74
  75         int posMail = mail.indexIn(line, start, QRegExp::CaretAtOffset);
  76         if (posMail == -1)
  77             posMail = line.size();
  78
  79         int posFormatting = anyFormatting.indexIn(line, start, QRegExp::CaretAtOffset);
  80         if (posFormatting == -1)
  81             posFormatting = line.size();
  82
  83         const int firstSpecial = qMin(qMin(posLink, posMail), posFormatting);
  84         if (firstSpecial == line.size()) {
  85             qDebug() << "nothing else";
  86             // No further matches for this line -> we're done
  87             break;
  88         }
  89         qDebug() << "some RE has matched";
  90
  91         if (firstSpecial == posLink) {
  92             QString replacement = QString::fromUtf8("<a href=\"%1\">%1</a>").arg(link.cap(1));
  93             line = line.left(firstSpecial) + replacement + line.mid(firstSpecial + link.matchedLength());
  94             start = firstSpecial + replacement.size();
  95         } else if (firstSpecial == posMail) {
  96             QString replacement = QString::fromUtf8("<a href=\"mailto:%1\">%1</a>").arg(mail.cap(1));
  97             line = line.left(firstSpecial) + replacement + line.mid(firstSpecial + mail.matchedLength());
  98             start = firstSpecial + replacement.size();
  99         } else if (firstSpecial == posFormatting) {
 100             // Careful here; the inner contents of the current match shall be formatted as well which is why we need recursion
 101             QChar elementName;
 102             QChar markupChar;
 103             const QRegExp *re = 0;
 104
 105             if (posFormatting == bold.indexIn(line, start, QRegExp::CaretAtOffset)) {
 106                 elementName = QLatin1Char('b');
 107                 markupChar = QLatin1Char('*');
 108                 re = &bold;
 109             } else if (posFormatting == italic.indexIn(line, start, QRegExp::CaretAtOffset)) {
 110                 elementName = QLatin1Char('i');
 111                 markupChar = QLatin1Char('/');
 112                 re = &italic;
 113             } else if (posFormatting == underline.indexIn(line, start, QRegExp::CaretAtOffset)) {
 114                 elementName = QLatin1Char('u');
 115                 markupChar = QLatin1Char('_');
 116                 re = &underline;
 117             }
 118             Q_ASSERT(re);
 119             qDebug() << "Got formatting";
 120             qDebug() << " old line:" << line;
 121             qDebug() << " at:" << line.mid(start);
 122             qDebug() << " prefix:" << line.left(firstSpecial);
 123             qDebug() << " suffix:" << line.mid(firstSpecial + re->matchedLength());
 124             QString replacement = QString::fromUtf8("%1<%2><span class=\"markup\">%3</span>%4<span class=\"markup\">%3</span></%2>%5")
 125                         .arg(re->cap(1), elementName, markupChar, helperHtmlifySingleLine(re->cap(2)), re->cap(3));
 126
 127             qDebug() << " replacement:" << replacement;
 128             line = line.left(firstSpecial) + replacement + line.mid(firstSpecial + re->matchedLength());
 129             start = firstSpecial + replacement.size();
 130             qDebug() << " chunk to be still processed:" << line.mid(start);
 131         } else {
 132             Q_ASSERT(false);
 133         }
 134     }
 135
 136     return line;
 137 }
 138
 139 QStringList plainTextToHtml(const QString &plaintext, const FlowedFormat flowed)
 140 {
 141
 142
 143     // Processing:
 144     // the plain text is split into lines
 145     // leading quotemarks are counted and stripped
 146     // next, the line is marked up (*bold*, /italic/, _underline_ and active link support)
 147     // if the last line ended with a space, the result is appended, otherwise canonical quotemarkes are
 148     // prepended and the line appended to the markup list (see http://tools.ietf.org/html/rfc3676)
 149     // whenever the quote level grows, a <blockquote> is opened and closed when it shrinks
 150
 151     int quoteLevel = 0;
 152     QStringList plain(plaintext.split('\n'));
 153     QStringList markup;
 154     // have we seen the signature separator and should we therefore explicitly close that block later?
 155     bool shallCloseSignature = false;
 156     for (int i = 0; i < plain.count(); ++i) {
 157         QString &line = plain[i];
 158
 159         // ignore empty lines
 160         if (line.isEmpty()) {
 161             markup << line;
 162             continue;
 163         }
 164         // determine quotelevel
 165         int cQuoteLevel = 0;
 166         if (line.at(0) == '>') {
 167             int j = 1;
 168             cQuoteLevel = 1;
 169             while (j < line.length() && (line.at(j) == '>' || line.at(j) == ' '))
 170                 cQuoteLevel += line.at(j++) == '>';
 171         }
 172         // strip quotemarks
 173         if (cQuoteLevel) {
 174             static QRegExp quotemarks("^[>\\s]*");
 175             line.remove(quotemarks);
 176         }
 177         // Escape the HTML entities
 178 #if QT_VERSION < QT_VERSION_CHECK(5, 0, 0)
 179         line = Qt::escape(line);
 180 #else
 181         line = line.toHtmlEscaped();
 182 #endif
 183
 184         line = helperHtmlifySingleLine(line);
 185
 186         // if this is a non floating new line, prepend canonical quotemarks
 187         if (cQuoteLevel && !(cQuoteLevel == quoteLevel && markup.last().endsWith(' '))) {
 188             QString quotemarks("<span class=\"quotemarks\">");
 189             for (int i = 0; i < cQuoteLevel; ++i)
 190                 quotemarks += "&gt;";
 191             quotemarks += " </span>";
 192             line.prepend(quotemarks);
 193         }
 194
 195         if (cQuoteLevel < quoteLevel) {
 196             // this line is ascending in the quoted depth
 197             Q_ASSERT(!markup.isEmpty());
 198             for (int i = 0; i < quoteLevel - cQuoteLevel; ++i) {
 199                 markup.last().append("</blockquote>");
 200             }
 201         } else if (cQuoteLevel > quoteLevel) {
 202             // even more nested quotations
 203             for (int i = 0; i < cQuoteLevel - quoteLevel; ++i) {
 204                 line.prepend("<blockquote>");
 205             }
 206         }
 207
 208         if (!shallCloseSignature && line == QLatin1String("-- ")) {
 209             // Only recognize the first signature separator
 210             shallCloseSignature = true;
 211             line.prepend(QLatin1String("<span class=\"signature\">"));
 212         }
 213
 214         // appaned or join the line
 215         if (markup.isEmpty()) {
 216             markup << line;
 217         } else if (flowed == FORMAT_FLOWED) {
 218             if ((quoteLevel == cQuoteLevel) && markup.last().endsWith(QLatin1Char(' ')) &&
 219                     markup.last() != QLatin1String("<span class=\"signature\">-- "))
 220                 markup.last().append(line);
 221             else
 222                 markup << line;
 223         } else {
 224             markup << line;
 225         }
 226
 227         quoteLevel = cQuoteLevel;
 228     }
 229
 230     // close any open elements
 231     QString closer;
 232     if (shallCloseSignature)
 233         closer = QLatin1String("</span>");
 234     // close open blockquotes
 235     // (bottom quoters, we're unfortunately -yet- not permittet to shoot them, so we need to deal with them ;-)
 236     while (quoteLevel > 0) {
 237         closer.append("</blockquote>");
 238         --quoteLevel;
 239     }
 240     if (!closer.isEmpty()) {
 241         Q_ASSERT(!markup.isEmpty());
 242         markup.last().append(closer);
 243     }
 244
 245     return markup;
 246 }
 247
 248 }
 249 }
 250
 251