Parsing And Layout Of The Browser. Vxworks-based Embedded Module

As the increasing demand of people for network resources and the rapid development of Internet technology, large number of embedded devices which can provide Internet access, pour out into a variety of electronic products markets. As we know, the Embedded Browser plays an important role in exploring and developing information aspects of the embedded devices, so Embedded Browser becomes a hot research area with High-performance, diversification and customization development trend. Nowdays, there are many types of Embedded Browser products at home or abroad, and even any of them are very excellent and full-featured. But most of the Embedded Browsers are develped based on Linux environment, and their parse module and layout module are coupled with GUI module, the layer between modules is uncertainty, so it is not good for possible Cross-platform migration or new support for GUI development.VxWorks operating system is an embedded real-time operating system which is designed and developed by Wind River Company, with its high-performance core, good sustainable development capabilities as well as user-friendly development environment that is widely used in communications, military, aerospace and any other sophisticated real-time requirements. Unfortunately, there are very few Embedded Browser applications in this real-time operating system. As VxWoks supports common network API features and standard BSD4.4, this article attempts to design the parse module and layout module independently from display modules in the framework of Embedded Browser, and the result of the implementation of the parse and layout module is a DOM tree data structure list which will be displayed by display module in the next step. After decoupling with the display module, we will probably be easily to implement the parse and layout module cross-platform migration, supports multi-purpose GUI implementation.Use VxWorks' multi-task concurrency mechanism between the web page getting module and parse module. Through shared data structure integrate with binary semphore to achieve getting web page and parse_layout operation at the same time, and for getting specified length of html string each time, it not only can reduce the memory pressure of the system, but also can improve the efficiency of the entire browser. The parse process is divided into three stages: lexical analysis, syntax analysis and semantic analysis. Based on lexical ananlysis's algorithm and the categories of HTML tags, we design a DFA(Deterministic Finite Automaton) to discribe the process of lexical analysis. In HTML tag's matching part, we introduced the idea of bitmap and conbined with Hash function, it can more rapidly and more efficently achieve the goal than previous parse algorithm. In layout module, we divided the HTML tags into image type, text type and table type based on their different characteristics of the layout process, use different layout methods to deal with different html tags. Finally, the result will run upon IE browser, after comparation of the display result, we analyze the lack of the parse module, layout module and summarize the next step to be done.Although the design did not yet support CSS and JavsScript,CSS and Javascript content is separated and stored in the process of parse module,we retains the interface of their interpretation. As embedded systems have inherent weak CPU, low power consumption, limited memory capacity and so on, the functional requirements for the embedded browser will not as the requirements demanded in PC, exclusively, it usually need functional cut depending on different requirements, so the parse module and layout module designed in this article have applicability for any embedded browser which have similar customizable demand.