Glenn McCorkle wrote: > On Sat, 23 Mar 2002 00:07:12 -0400, Clarence Verge wrote: > > >>On Sat, 23 Mar 2002 13:28:43 +1100, John Tate wrote: > > >>>Caldera Web Spyder isnt bloated and it has Java Script >> > >>Maybe we should get Web Spyder and reverse engineer it then. >>It's Arachne. <g>
Document Object Model used in Arachne isn't suitable for JavaScript. I don't know how they did it in WebSpyder... I already introduced some changes and used them also for CSS, but it is not easy. So far, I am able to parse any tag-embeded source code (<TEXTAREA>, <STYLE> and <JAVASCRIPT> into Arachne's text editor structure (you can call it text editor class, if you are into OOP), which means we can read it line by line (but on the other hand, for interpreting JavaScript, simple character buffer would work just enough...of course, if there was no 640 KB limit for ALL character buffers in use ;-) . But this is one thing. We can have ib_editor structur with all JavaScript used in HTML page, if you really care about that. Now we need detail - JavaScript interpreter - and binding of all JavaScript objects to some real objects in Arachne. I almost blow my mind when I tried to imagine this... it is simply lot of work. Naming of objects using JS-namespace should be probably introduced since the very begining of HTML-parser design, ie. for every introuced object (text, input field, image, whatever), JS attributes should have been descibred. Doing this in reverse order would drive anyone insane...
