Thanks for keeping this on the boil. The Forrestdoc link you gave me shows the Javasrc at a much greater level of integration than was visible when I last looked. My problem originally was that only line-number links were provided. In the current forrestdoc version, everything is linked in to comprehensive cross-reference web. Nice.

The appearance of the the source text is not optimal, but there is scope in Javasrc to correct this, I believe. There are probably two levels of tuning. Firstly, syntax highlighting, and secondly. link presentation. The syntax highlighting variations (if any) would be Javasrc configuration options. The link presentation would be available through the CSS used to present the document. Most things of interest, and, AFAICT, all links are htmlized with a relevant "class" attribute, so it is feasible to select the link colour or colours for various classes to allow appropriate highlighting.

Keywords, although emphasised in the text, are not given a class attribute, which is unfortunate, but may be tunable.

The source html I used was generated by htmlize.el, an emacs and XEmacs module by Hrvoje Niksic. In order to use it, you will need to have an emacs or xemacs binary available to the build process. Forrestdoc looks the far better option.


Clay Leeds wrote:

On Jun 29, 2004, at 7:04 PM, Peter B. West wrote:


FYI, Java 1.4 javadoc tool supports a -linksource argument, which generates html of source files. However, the process seems to have pretty much the same restrictions as the Maven JXR - the only references are to line numbers, which is just about the most useless form imaginable.

It might be of interest to someone at a later stage to look at extending the standard doclet to utilise Javasrc to perform that generation.


I've been hoping to get movement on the forrest front, then include the java doc stuff once I have xml-fop up & running. As that's not moving along very fast, perhaps I can get a bit more information on what we need to do with the java doc front.

Here's something I 'dug' up[1]. Looks like it's along the lines of what we're looking for, although I like what you've got on this page[2] better. Your version seems more accessible. However, you indicate:

The problem is that there was no clean way to automatically generate the htmlized source. It's that supplementary facility that I'm looking for.

What is the procedure used to generate the htmlized source? Could it be automated using gump or ant or something? BTW, what's the tool called you use to generate this?

Web Maestro Clay


[2] overview.html

-- Peter B. West <>

Reply via email to