On Wed, Mar 09, Lars Segerlund wrote:
I have a theoretical question, as follows.
What would be needed for Dillo to support javascript ? Of course it would need a javascript interpreter, but there are also issues about variables and how to handle stuff like mouseover and such in the rest of the code.
The script interpreter could be a dpi; to handle variables, the dpi should provide the possibility to handle different states (something like a key, which dillo creates for each page, and which helps the dpi to distinguish different states, which have independant sets of variables etc.). Then there are events ('on...' attributes of the HTML tags). There is a document tree already in the CSS prototype (see <http://www.dillo.org/CSS.html>), which provides some signals based on document *elements* (not on dillo widgets anymore), this could be extended. Somehow, these signals need to be connected to the script engine dpi. Also, there needs to be another interface, so that the dpi can access properties of dillo (i.e. HTML DOM). For the XML subset (XML DOM), the document tree in the CSS prototype must be extended a bit. (The planned functionality is not that much less than the functionality needed for a full XML DOM implementation). The rest of HTML DOM is related to the DilloHtmlLB structure.
Is there any possibility that the dpi interface would suffice for a task like this ? ( what I mean here is if there is a possibility to somhow use it for this functionality, perhaps 'filtering' the pages through it or something ? ).
Basicly I'm curious, and I would like to have a look at if it's doable.
It should be doable, but there are some changes in dillo needed. Sebastian