Reporting detailed XML errors when using QDomDocument

I'm currently modifying some old code that is using QDomDocument to parse XML file content. Example code below:

QFile file(file_.filePath());   

if (file.open(QIODevice::ReadOnly)) 
{
    QTextStream in(&file);
    file.close();

    QString errorMsg;
    int errorLine;
    int errorColumn;

    if (!doc.setContent(in.readAll(), &errorMsg, &errorLine, &errorColumn))
    {
        qWarning("Invalid XML in file %s. Error = %s, Line = %d, Column = %d", qPrintable(file_name), qPrintable(errorMsg), errorLine, errorColumn);
    }
}

Unfortunately the error reporting is extremely limited. Removing a closing tag half way down the file simply reports the following error:

Invalid XML in file config.xml. Error = unexpected end of file, Line = 1, Column = 1

Which is fairly useless.

Any suggestions of how to get more descriptive errors out of Qt's XML parsers? An accurate line number would be a good start.

P.S. I'm using Qt version 4.7.4.

Solution

QDomDocument::setContent should give you the right information to know where the problem is.

For example, with this snippet of code:

#include <QtXml>
#include <QtCore>

int main()
{
    QFile file(":/myxml_error.xml");

    qDebug() << "File path:" << QFileInfo(file).absoluteFilePath();
    qDebug() << "File exists:" << file.exists();

    file.open(QFile::ReadOnly|QFile::Text);

    qDebug() << "File open:" << file.isOpen();

    QDomDocument dom;
    QString error;

    int line, column;

    if(dom.setContent(&file, &error, &line, &column)){
        qDebug() << dom.toString(4);
    } else {
        qDebug() << "Error:" << error << "in line " << line << "column" << column;
    }
    return 0;
}

And this xml file:

<?xml version="1.0" encoding="UTF-8"?>
<note>
    <to>Tove</to>
    <from>Jani</Ffrom>
    <heading>Reminder</heading>
    <body>Don't forget me this weekend!</body>
</note>

We see the following error:

Error: "tag mismatch" in line 4 column 19

I've updated this answer after reading some comments from @kh25.

The problem could be due to the file encoding. In this case, we could try removing carriage returns and new lines. dos2unix can help if we're working on Linux.
Another option would be to use an online tool to check the XML.

But in this particular case, it seems the problem is related to closing the file -- file.close() -- before calling in.readAll(). In this case, QDomDocument::setContent is reading an empty string and we see the error unexpected end of file.

Another way of getting the error is calling QDomDocument::setContent when the file has reached the end of the stream.

So, if we call for example QTextStream::readAll() twice, we get the same error. I.e:

QTextStream in(&file);
in.readAll();

if(dom.setContent(in.readAll(), &error, &line, &column)) {
    qDebug() << "Content: " << dom.toString(4);
} else {        
    qDebug() << "Error:" << error << "in line " << line << "column" << column;
}

@MrEricSir commented about using QXMLStreamReader. If you need an example, I have a little project in GitHub where I use that class.