QSyntaxHighlighter with parser

[Sikarjan]

Hi,

as suggested by this forum I tried to redo the QSyntaxHighlighter example with a text parser instead of the regular expression. The syntax highlighting shall be for PHP. A PHP file can contain multiple other languages. So far I added the highlighting for PHP and HTML. Before I move on to add support for CSS and JavaScript I wanted to ask for feedback. Am I on the right track here or am I doing this completely wrong? This is my first parser.
Passing the key words to the highlighter needs to be improved, I know that. Here I need to come up with a good idea how I can do that for different files. If I use the highlighter with a css file, I only need and want the css highlighting.
I also have a second question. Besides the language specific syntax highlighting I would like to have language specific auto completion. How can I pass the current block state to the completer?

Here is the modified code. I already included the matching parenthesis.
highlighter.h

#ifndef HIGHLIGHTER_H
#define HIGHLIGHTER_H

#include <QSyntaxHighlighter>
#include <QVector>
#include <QStringListModel>
#include <QRegularExpression>

class QString;
class QTextDocument;

struct ParenthesisInfo {
    char character;
    int position;
};

class TextBlockData : public QTextBlockUserData {
public:
    TextBlockData();

    QVector<ParenthesisInfo *> parentheses();
    void insert(ParenthesisInfo *info);

private:
    QVector<ParenthesisInfo *> m_parentheses;
};

class Highlighter : public QSyntaxHighlighter {
    Q_OBJECT

public:
    Highlighter(QTextDocument *document, QStringListModel *keyWords, QStringListModel *funcNames);

protected:
    void highlightBlock(const QString &text);

private:
    QMap<QString, QTextCharFormat> phpKeyWordsRules;
    QMap<QString, QTextCharFormat> htmlKeyWordsRules;

    QTextCharFormat currentTagFormat;

    QTextCharFormat phpKeywordFormat;
    QTextCharFormat phpTagFormat;
    QTextCharFormat phpCommentFormat;
    QTextCharFormat phpQuoteFormat;
    QTextCharFormat phpFunctionFormat;
    QTextCharFormat phpVariableFormat;
    QTextCharFormat phpNumberFormat;
    QTextCharFormat phpBoolFormat;

    QTextCharFormat htmlTagFormat;
    QTextCharFormat htmlGenTagFormat;
    QTextCharFormat htmlTableFormat;
    QTextCharFormat htmlFormFormat;
    QTextCharFormat htmlQuoteFormat;
    QTextCharFormat htmlCommentFormat;

    QRegularExpressionMatchIterator parser;
    QString currentLine;

    void closeTag(QString tag, int returnState, int startPos = 0);
};

#endif

highlighter.cpp

#include <QtGui>

#include "highlighter.h"

TextBlockData::TextBlockData()
{
    // Nothing to do
}

QVector<ParenthesisInfo *> TextBlockData::parentheses()
{
    return m_parentheses;
}


void TextBlockData::insert(ParenthesisInfo *info)
{
    int i = 0;
    while (i < m_parentheses.size() &&
        info->position > m_parentheses.at(i)->position)
        ++i;

    m_parentheses.insert(i, info);
}

Highlighter::Highlighter(QTextDocument *document, QStringListModel *keyWords, QStringListModel *funcNames): QSyntaxHighlighter(document){
// PHP formatting
    phpTagFormat.setForeground(Qt::red);
    phpVariableFormat.setForeground(Qt::blue);
    phpCommentFormat.setForeground(Qt::gray);
    phpQuoteFormat.setForeground(Qt::darkGreen);
    phpBoolFormat.setForeground(QColor(0,100,0));
    phpNumberFormat.setForeground(Qt::darkRed);

    phpKeywordFormat.setForeground(QColor(128, 128, 128));
    phpKeywordFormat.setFontWeight(QFont::Normal);
    phpKeywordFormat.setFontItalic(true);
    int rows = keyWords->rowCount();
    for (int j = 0; j < rows; j++) {
       phpKeyWordsRules.insert(keyWords->data(keyWords->index(j), Qt::DisplayRole).toString(), phpKeywordFormat);
    }
    phpKeyWordsRules.insert("true", phpBoolFormat);
    phpKeyWordsRules.insert("false", phpBoolFormat);

    phpFunctionFormat.setForeground(Qt::darkBlue);
    rows = funcNames->rowCount();
    for (int k = 0; k < rows; k++) {
        phpKeyWordsRules.insert(funcNames->data(funcNames->index(k), Qt::DisplayRole).toString(), phpFunctionFormat);
    }

// HTML formatting
    htmlGenTagFormat.setForeground(Qt::blue);
    htmlTableFormat.setForeground(Qt::cyan);
    htmlFormFormat.setForeground(QColor(255, 153, 0));
    htmlQuoteFormat.setForeground(Qt::darkGreen);
    htmlCommentFormat.setForeground(Qt::gray);

    QStringList htmlList;
    htmlList << "DOCTYPE" << "a" << "abbr" << "address" << "area" << "article" << "aside" << "audio" << "b" << "base" << "bdi" << "bdo" << "blockquote" << "body" << "br" << "button" << "canvas" << "caption" << "cite" << "code" << "col" << "colgroup" << "command" << "datalist" << "dd" << "del" << "details" << "dfn" << "dir" << "div" << "dl" << "dt" << "em" << "embed" << "fieldset" << "figcaption" << "figure" << "footer" << "h" << "head" << "header" << "hgroup" << "hr" << "html" << "i" << "iframe" << "img" << "ins" << "kbd" << "keygen" << "label" << "legend" << "li" << "link" << "map" << "mark" << "menu" << "meta" << "meter" << "nav" << "noscript" << "object" << "ol" << "optgroup" << "option" << "output" << "p" << "param" << "pre" << "progress" << "q" << "rp" << "rt" << "ruby" << "s" << "samp" << "script" << "section" << "small" << "source" << "span" << "strong" << "style" << "sub" << "summary" << "sup" << "textarea" << "tfoot" << "time" << "title" << "track" << "u" << "ul" << "var" << "video" << "wbr";
    foreach (QString tag, htmlList) {
        htmlKeyWordsRules.insert(tag, htmlGenTagFormat);
    }

    QStringList htmlTableList;
    htmlTableList << "table" << "tr" << "th" << "td" << "tbody" << "thead";
    foreach (QString tag, htmlTableList) {
        htmlKeyWordsRules.insert(tag, htmlTableFormat);
    }

    QStringList htmlFormList;
    htmlFormList << "form" << "input" << "select";
    foreach (QString tag, htmlFormList) {
        htmlKeyWordsRules.insert(tag, htmlFormFormat);
    }
}

void Highlighter::highlightBlock(const QString &text){
    currentLine = text;
    // highlight matching parenthesis
    TextBlockData *data = new TextBlockData;

    int leftPos = text.indexOf('(');
    while (leftPos != -1) {
        ParenthesisInfo *info = new ParenthesisInfo;
        info->character = '(';
        info->position = leftPos;

        data->insert(info);
        leftPos = text.indexOf('(', leftPos + 1);
    }

    int rightPos = text.indexOf(')');
    while (rightPos != -1) {
        ParenthesisInfo *info = new ParenthesisInfo;
        info->character = ')';
        info->position = rightPos;

        data->insert(info);

        rightPos = text.indexOf(')', rightPos +1);
    }

    setCurrentBlockUserData(data);

//- Text parser -
    // check if there is something to parse
    if(text.simplified().trimmed().isEmpty()){
        setCurrentBlockState(previousBlockState());
        return;
    }
    // Split line by coding language identifiers, here <? and ?>, then by comment tags, then by (key)words and then by any other char that is not escaped.
    QRegularExpression splitter = QRegularExpression("(<\\?)|(\\?>)|(\\/\\*)|(\\*\\/)|(\\w+)|((?<!\\\\)([^\\s\\\\]))");
    parser = splitter.globalMatch(text);

//    qDebug() <<"Starting with" << currentBlockState() << "last state was" << previousBlockState() << "looking at" << text;
    // check if the last block can tell us what language we should start in

    if(previousBlockState() != -1){
        setCurrentBlockState(previousBlockState());
    }else{
        // Assuming it should be html then
        setCurrentBlockState(10);
    }

    // Close any open comments/strings if necessary
    if(currentBlockState() == 1){
        currentTagFormat = phpQuoteFormat;
        closeTag("\"", 0);
    }else if(currentBlockState() == 2){
        currentTagFormat = phpQuoteFormat;
        closeTag("'", 0);
    }else if(currentBlockState() == 3){
        currentTagFormat = phpCommentFormat;
        closeTag("*/", 0);
    }

    while(parser.hasNext()){
        QRegularExpressionMatch word = parser.next();

        if(word.captured() == "<?"){
            // from now on this is PHP
            setCurrentBlockState(0);
            int start = word.capturedStart();
            if(parser.hasNext() && parser.peekNext().captured() == "php"){
                word = parser.next();
            }else if(parser.hasNext() && parser.peekNext().captured() == "="){
                word = parser.next();
            }

            setFormat(start, word.capturedEnd(), phpTagFormat);
            if(parser.hasNext())
                word = parser.next();
        }

        // Start highlighting according to coding language
        if(currentBlockState() < 10){
            // *** PHP highlighting ***
            int startExp = 0;

           if(word.captured() == "$"){
               // we found a variable
               startExp = word.capturedStart();

               word = parser.next();
               if(word.captured() == "$")
                   word = parser.next();
               setFormat(startExp,word.capturedEnd()-startExp, phpVariableFormat);
           }else if(phpKeyWordsRules.contains(word.captured().toLower())){
               setFormat(word.capturedStart(), word.capturedLength(), phpKeyWordsRules[word.captured()]);
           }else if(word.captured() == "/" && parser.hasNext() && parser.peekNext().captured() == "/"){
               setFormat(word.capturedStart(), text.length()-word.capturedStart(), phpCommentFormat);
           }else if(word.captured() == "?>"){
               setCurrentBlockState(10); // muss geändert werden falls zurück zu css oder js
               setFormat(word.capturedStart(), word.capturedLength(), phpTagFormat);
           }else if(word.captured().contains(QRegExp("\\d+"))){
               startExp = word.capturedStart();
               if(parser.hasNext() && parser.peekNext().captured() == ".")
                   word = parser.next();
               setFormat(startExp, word.capturedEnd()-startExp, phpNumberFormat);
           }else if(word.captured() == "\""){
               currentTagFormat = phpQuoteFormat;
               setCurrentBlockState(1);
               closeTag("\"", 0, word.capturedStart());
           }else if(word.captured() == "'"){
               currentTagFormat = phpQuoteFormat;
               setCurrentBlockState(2);
               closeTag("'", 0, word.capturedStart());
           }else if(word.captured() == "/*"){
               currentTagFormat = phpCommentFormat;
               setCurrentBlockState(3);
               closeTag("*/", 0, word.capturedStart());
           }
        }else{
            // HTML highlighting
            if(word.captured() == "<"){
                int startExp = word.capturedStart();
                if(parser.hasNext() && parser.peekNext().captured().toLower().contains(QRegExp("[^a-z]")))
                    word = parser.next();

                if(parser.hasNext()){
                    QString htmlTag = parser.peekNext().captured().toLower();
                    if(htmlKeyWordsRules.contains(htmlTag))
                        htmlTagFormat = htmlKeyWordsRules[htmlTag];

                    setFormat(startExp, word.capturedEnd()-startExp, htmlTagFormat);
                }

                bool tagAtEnd = false;
                while(parser.hasNext() && !tagAtEnd){
                    word = parser.next();
                    if(word.captured() == "\""){
                        currentTagFormat = htmlQuoteFormat;
                        closeTag("\"", 10, word.capturedStart());
                    }else{
                        setFormat(word.capturedStart(), word.capturedLength(), htmlTagFormat);
                        if(word.captured() == ">")
                            tagAtEnd = true;
                    }
                }
            }
        }
    }
//    qDebug() << "exeting line" << text << "with status" << currentBlockState();
}

void Highlighter::closeTag(QString tag, int returnState, int startPos){
    int length = 0;

    while(parser.hasNext()){
        QRegularExpressionMatch word = parser.next();
        if(word.captured() == tag){
            setCurrentBlockState(returnState);
            length = word.capturedEnd() - startPos;
            break;
        }
    }

    if(length == 0){
        // No end was found, multiline continues on next line
        length = currentLine.length();
    }

    setFormat(startPos, length, currentTagFormat);
}