d-programming-language.com

[Jacob Carlborg]

On 2011-12-11 05:14, Adam D. Ruppe wrote:
> Andrei Alexandrescu Wrote:
>> Anything automated rox.
>
> cool. I updated the program today to handle the new site. Take a
> look at the output:
>
> http://arsdnet.net/d-web-site/std_stdio.html
>
> Biggest improvement imo is anchors work better:
> http://arsdnet.net/d-web-site/std_stdio.html#File.writeln
>
> is distinct from #writeln.
>
>
> You might remember this from last winter when you introduced
> the cheatsheet idea.
>
> I figured I could automate the process by examining the html, and
> this is the result. (Improved a little tonight from how it was last time.)
>
>
> What it does is determine the fully qualified names from the html structure,
> then pulls out the first paragraph for each name.
>
> Using that, it fixes the link and generates the table, categorizing the
> found names as struct, class, function, enum, or unknown, with some
> support for subcategorizing if we add tags to the ddoc (I can write a macro
> to help that along)
>
>
> Anyway, with that in hand, it modifies the html to add these tables on
> top of each section.
>
>
> First, there's the module scope names, sorted alphabetically.
>
> Then, if you click struct File, you'll see that under it's description, right
> before it's members, there's a sub-table of member quick references.
>
> They are not categorized by class/struct since I think that's overkill in
> there. But, they are sorted by name and link right to the inside.
>
> If a file has more nesting, it should just work like this all the way through
> recursively.
>
>
>
> This little program runs on the command line and modifies a given html
> file, just writing out another static file you can upload to the server.
>
>
>
> Anyway it's pretty much complete here. Let me know what you think.

Wouldn't it be better if ddoc generate the correct HTML from the beginning.

-- 
/Jacob Carlborg