|
How AntDoc worksThis section discusses how AntDoc works. It mainly focuses on AntDoc HTML documentation generation; a few words are said about
AntDocGUI.
AntDocIt is not a surprise if I tell you that most of the job concerning the generation is performed thanks to eXtensible Stylesheet Language Transformation (XSLT).
Before the actual transformations are invoked, some Java code is needed in order to prepare the transformations: the filtering, the interpretation of the Ant properties are not handled within the XSLT. As a matter of fact, the stylesheet takes an XML file as an input, and this file is generated on the fly by the task, by analysing the Ant scripts, and by creating a new Document Object Model (DOM). The previous filtering is performed through Regexp V1.2, as a pre-treatment, so that you may use regular expressions when expresing the kind of filtering that you wish to apply on the targets. Moreover, the generated HTML pages use some very minor JavaScript in order to perform some call on a lightweight applet; on its turns, the applet behaves on a client and connects to the AntDocGUI (the server part) in order to ask for Ant target execution. This applet code base is the AntDoc jar file itself, and this file is signed in order to be allowed to send information on a socket. Because of the Java sandbox model, a server part has been splitted from the client part. In order to natively run both under Internet Explorer and Netscape Communicator, two files are provided in the release. A Java ARchive File (a file with extension AntDocGUIThe GUI part itself has been developed with the Java Swing API: the GUI also uses a socket, in order to receive target execution demand from the AntDoc HTML generated pages. Real explanation coming up...And Ant?Of course, this site is administered by a single Ant script that enables to re-build everything from scratch: the distribution resouces and this web site. Some tests are also automatically performed. |
Current date: 2025.08.11 19:42:35 |
Number of hit:
|