The API is composed of some classes (UserGroup, Namespace, General, Page...) and a MediaWiki class to access data.
Library architecture at the begining
The library is composed of a unique class: MediaWiki.
Synchronous or asynchronous?
Two choices are possible to develop the request management in the library: synchronous and asynchronous.
Synchronous way will lead us to use normal function call. But, since we expects the end of the function, the interface can freeze especially if there is a lot of latency. Moreover, this way will allow to not use signals.
Asynchronous way don’t block the interface because the application doesn’t wait the result of each request to execute the program. This method will lead us to use signals to access to the result.
On the one hand, use synchronous way will lead to use a QEventloop to avoid the freezing and a QNetworkAccessManager in synchronous mode. Otherwise, this way can involve the use of different threads in the application.
On the other hand, use asynchronous way will lead us to use signals and a QNetworkAccessManager in asynchronous mode.
To conclude, we choose the asynchronous one. Moreover, this method is the one used in the library architecture at the begining. But the asynchronous way give the problem of access to the result.
Get the request’s result
We can use two methods to get the result. First, we can develop accessors. This way can cause risks of a call before the signal. This method use a synchronous call. A solution can be to specify this trouble in the documentation of the API.
The second method is to use a signal having the result as parameter. This way respect the asynchronous semantic but the result is pass by value. The pass by value will be a trouble because it is something heavy. A possible problem will be that the copy is useless. A solution could be using pass by pointer or smart pointer.
Develop Unique class vs multiple classes?
- No memory managing
- The request calls are normal : wiki->allpagesRequest()
- Can handle only 1 request at a time.
- Can handle some requests at a time.
- The memory management
Technical contacts requirements
First, the technical contacts want a asynchronous interface: ”Even if you use a QEventLoop,you'll often need different threads in the application, and those can be a bitch. You can have a look at the current MediaWiki class to how those signals / slots should look like.”
Secondly, to call a request, the code should look like:
MediaWiki * mw = new MediaWiki(“mon_url”); mw->allpagesRequest(); connect(mw, SIGNAL(allpagesResult(QList<MediaWiki::Page>)), SLOT(allpagesProcess(QList<MediaWiki::Page>));
The MediaWiki class allows developers to call asynchronous request to access the MediaWiki data. For this, developers call a request like allpagesRequest() who send the request with the QNetworkAccessManager. To process the result, developers connect the signal allpagesResult() with his own slot. Because the class MediaWiki needs to emit signals, MediaWiki inherit QObject.
Why don’t separate the code
To follow the first development of the MediaWiki class and with the agreement of the technical contacts, we don’t separate MediaWiki and its requests.
Who manages the memory?
There is one class, so the library can manage the memory.
It’s the Qt class for HTTP requests and it is asynchronous. By the way, we can use this class in a synchronous way using QEventLoop.
At this time, we don’t use KDE library because Qt provides the necessary.
To avoid name conflicts with others libraries, we propose to define a namespace like silk::. mediawiki:: was a possibility but there will be a redundancy (mediawiki::MediaWiki).
For ensure the binary compatibility, we’ll use only a pointer to MediaWikiPrivate, which will contain attributes. In this case, modify the attributes doesn’t fail a class which uses MediaWiki.