Thread overview
idea -- inline docs ala Javadoc
Sep 10, 2002
chris jones
Sep 10, 2002
Pavel Minayev
Sep 10, 2002
chris jones
Sep 10, 2002
Burton Radons
Sep 12, 2002
Walter
Sep 11, 2002
andy
Sep 11, 2002
Juarez Rudsatz
September 10, 2002
One of the things i liked about Java was the idea of Javadoc. So imo some basic specs for inline comments that could be easily interpreted into referance docs would be a good adition to D. I have never used Javadoc but i think it parsed the source file and made a basic html page from the source, and you put any extra info for Javadoc in coments in the source code. As alot people do put coments in their code, if there was a spec for these coments so they could be parsed by 'DDocGen' then i think they would use it. As they would be comments then the compiler will ignore them, no extra work for Walter :) And i am thinking a sophisticated Doc generater could scan multiple files and generated a prety complex referance all nicely crosslinked and such. Also it ocurs to me that it could be alot easier to integrate the docs from diferant projects/librarys if the directory structure follows the namespace structure. Its a long time since i used java so i cant remember what the syntax of the coments were but i dont think it needs to be very complicated., but it would need to be part of the D spec,

what do you think?

chris


September 10, 2002
chris jones wrote:

> One of the things i liked about Java was the idea of Javadoc. So imo some

It was discussed already. See "Code documentation" threads. Right now, it's the syntax that's being discussed, there is no working doc generator yet.

September 10, 2002
Great :)

"Pavel Minayev" <evilone@omen.ru> wrote in message news:alljd6$gr7$3@digitaldaemon.com...
> chris jones wrote:
>
> > One of the things i liked about Java was the idea of Javadoc. So imo
some
>
> It was discussed already. See "Code documentation" threads. Right now, it's the syntax that's being discussed, there is no working doc generator yet.
>


September 10, 2002
Pavel Minayev wrote:

> chris jones wrote:
> 
>> One of the things i liked about Java was the idea of Javadoc. So imo some
> 
> 
> It was discussed already. See "Code documentation" threads. Right now, it's the syntax that's being discussed, there is no working doc generator yet.

The working copy of my port produces some docs, but it's been delayed. The nice thing is that the fastest way to compile a large project is just to send all the files through dli (it's more than fast enough - the 400kb Phobos takes just 2.3 seconds, and could probably be brought under one if it did assembly itself), so project-wide documentation can be gathered rather easily.

Plus, of course, the compiler knows the language better than any code generator could manage.

September 11, 2002
Pavel Minayev wrote:
> chris jones wrote:
> 
>> One of the things i liked about Java was the idea of Javadoc. So imo some
> 
> 
> It was discussed already. See "Code documentation" threads. Right now, it's the syntax that's being discussed, there is no working doc generator yet.
> 

One thing is it might be nice to have an enhanced version of the html deal that can grab other defined tags inside of code and (assume XML basically) and use that as doco.  Then others who chose could create stylized version through XSLT.  Just a thought...

September 11, 2002
andy <acoliver@apache.org> wrote in news:3D7E96DE.1000509@apache.org:

> One thing is it might be nice to have an enhanced version of the html deal that can grab other defined tags inside of code and (assume XML basically) and use that as doco.  Then others who chose could create stylized version through XSLT.  Just a thought...

Have you seen the xml docbook format ?

A compiler could extract the documentation from comments in source code and generate a xml docbook. Than you can use your favorite docbook transformer ( XSLT, Jade, ... ) for generating the final version.

September 12, 2002
"Burton Radons" <loth@users.sourceforge.net> wrote in message news:3D7E826C.9000504@users.sourceforge.net...
> The nice thing is that the fastest way to compile a large project is just to send all the files through dli (it's more than fast enough - the 400kb Phobos takes just 2.3 seconds, and could probably be brought under one if it did assembly itself), so project-wide documentation can be gathered rather easily.

<g> D is designed for fast compiling.