Quering the web from a QDialog

published at 13.03.2025 16:36 by Jens Weller
Save to Instapaper Pocket

I've recently made a Meeting C++ news page public, which lists all the links I post on the social media of Meeting C++. That way you don't rely on an algorithm to know whats the newest blog posts and news in the C++ world.

Some of this is weekly posted to the Meeting C++ blogroll, which originates in the RSS Reader of Meeting C++. But the newspage now holds the data in the database and makes it also searchable to you. I usually don't add manually links to the blogroll, so if a blog does not have an RSS Feed or it breaks, its not included. Also it might be simply not worth the effort to include a page into the RSS Feeds of Meeting C++, as not all interesting C++ articles show up on pages focused on C++ articles. Thats why I've added a small form for myself to share things with the newspage when I post such a link to my network. This blog post is now about how to get this into the blogroll.

I've created an endpoint that returns the title + link for all things manually posted to the news page in the last week as JSON. So the task is now to download this JSON data and transform it into an html snipped which is than added to the blogroll. This turned out to be a bit easier then I had thought. The selection for the links shared in the blogroll is done via a dialog, and I was not sure if that is a modal dialog (or a dialog from Qt) or a custom dialog. It turned out to be custom dialog named SelectTitles (naming is hard), and with this I could add the simple code to query the web to this dialog.

I do have a wrapper class named HttpDownloader for downloading via http(s), wrapping the mechanics of QNetworkAccessManager. I've documented this class earlier on my blog, when I've used it to download the images of speakers and books for the door signs at Meeting C++. For this usage I've decided that its easiest to add a QByteArray member and an owned member of HttpDownloader to the SelectTitles class. Then there is just plugging signals into slots and assigning the rawdata from the web to the QByteArray.

If this code would have used not a custom dialog but a Qt provided one, I could not add this ability to the dialog class. Then one would have to create local instance of HttpDownloader and overwrite the signals with lambdas to obtain the data from the web. The dialog runs long enough to ensure the webquery goes through and is stored when its returning from its modal state. An alternative would be to create a future for downloading the data and then simply call wait() once everything else is done. Which would also allow for better error handling in case the query does not go through.

Which leaves the task of converting the rawdata to a QJsonDocument and creating the actual content for the blogroll:

QString newslinks;
QByteArray json = select.getPostingjson();
if(!json.isEmpty())
{
    QJsonDocument doc = QJsonDocument::fromJson(json);
    if(doc.isArray())
    {
        QJsonArray jarr = doc.array();
        for(const QJsonValue jval:jarr)
        {
            QJsonObject jobj = jval.toObject();
            if(jobj.contains("title") && jobj["title"].isString() && jobj.contains("link") && jobj["link"].isString())
                newslinks += htmllist.arg(jobj["link"].toString()).arg(jobj["title"].toString());
        }
    }
    else qDebug() << "JSON Document is not an array!";
}

The Meeting C++ blogroll exists since February 2015, and since then the code creating it has been pretty stable. Its a set of text templates which then are fed with the needed data, as you can see with htmllist, which has the value of "<li><a href=\"%1\" rel=\"nofollow\">%2</a></li>", newslinks contains then - once the code has run - the list of postings from the website as raw html. Which is what is needed, as this is then posted to the OS clipboard and used as input for sharing the blogroll with its subscribers and adding it as a post to the static website.

Join the Meeting C++ patreon community!
This and other posts on Meeting C++ are enabled by my supporters on patreon!