On 8 août 08, at 12:18, malaka ekanayake wrote:

Hi devs,

I have complied the content assistance for xwiki API.Now the proposals are readout from a seperate resource file.There are seperate file for each context type ( i.e seperate files for $xwiki ,$doc .. ) .At the moment the API information is stored in the files in the following format

xwiki.copyDocument(java.lang.String docname, java.lang.String sourceWiki, java.lang.String targetWiki, java.lang.String wikilanguage) is transtatet to a single line entry in the xwiki.txt resourse file as bellow

copyDocument String docname String sourceWiki String targetWiki String wikilanguage

I will review your code asap and I will give you more detailed feedback. In this representation have you taken into account the fact that some identifiers semantically represent pages, spaces, etc.? You need this information for providing further assistance, as I suggested you in a previous email. For example whenever a docname is requested as a parameter you can propose completions showing all the available pages. In the previous example you gave, are you assuming that all the parameters that represent a page are called "docname"?

Now i'm looking at velocity variable proposal and error notification.In order to provide content assistance and error marking, there is a requirement to parse and build the an object model of document which contains both XWiki and Velocity. AFAIK, XDOM only supports XWiki. Is there an object model which can r present these two languages.

For velocity code, there is a possibility to use the parser shipped with Apache Velocity. But then we need to develop our own object model.

Any help would be greatly appreciated.

I think that in order to complete your GSoC task on the velocity subsystem, you can directly use the Velocity parser and its object model in order to analyze the document, find errors and report them as markers in the document. WDYT? -Fabio