Hi Anteru,
Firstly, a big thanks for answering in such detail, it is nice to have a
good understanding of your intentions and motivations :)
I think that yours is a bit different of a use case than ours and in a
sense I can see that you may need a different project for that kind of
stuff.
We are using breathe for a mixed language project, but to be able to
illustrate uses with examples etc. rather than providing full
documentation, at least at the moment.
For doing 'fully automatic' stuff we just use Doxygen - it can generate
neat stuff if you (and especially your customers) need it. I think Sphinx
is not meant for such things anyway, I mean it would be nice to be able to
do that, why not, but good Sphinx docs can't be automatic, they need the
human attention. Though I do remember forcing breathe to generate docs for
an entire class, so doing a whole project should not be so terrible if you
had to.
What is for me a downside of your solution is the mongodb - one more
service you need to run, one more thing that can break :)
Still, congratulations on an ambitious project and for contributing it to
the community - and under BSD, nice!
Best,
Michael
On Saturday, 1 September 2012 21:15:07 UTC+2, Anteru wrote:
>
> Hi,
>
> we did try Breathe actually; the main problem we had with it was that we
> couldn't hack the output easily and that there was no "automodule"
> support in it to generate documentation for all files in a project. Our
> goal was to make a minimal bridge which is easy to hack, so anything
> that is missing can be added easily, and to abstract away from Doxygen
> as quickly as possible.
>
> There's still some stuff which can be implemented better in Robin (for
> instance, we currently need two passes over the generated .rst files)
> but our feeling is that hacking Robin to get it working will be easier
> than hacking Breathe. If you take a look at Robin, it's cleanly
> separated into one pass which converts from Doxygen to MongoDB, and a
> second pass which works on well-structured data. We did actually a split
> on the development side, with one author writing each of the parts :)
>
> That's however our personal view, and so far, we do generate the
> documentation for a large C++ project with it and we're happy. If
> there's anything we can do to make the code more accessible, feel free
> to drop us a line.
>
> Cheers
>
> Am 28.08.2012 22:51, schrieb Michael Gielda:
> > Hi,
> >
> > My thoughts exactly, Michael, I use your Breathe extension and most of
> > the stuff just works... gotta test if the new thing provides more
> > functionality but perhaps the authors can point us in the right
> direction?
> > I mean Breathe is not ideal but perhaps it would be wiser to fill in the
> > blanks than write a new framework from scratch. Still, if it is there,
> > perhaps they can benefit mutually from the 'competition'?
> >
> > Best,
> > Michael (it's a fairly popular name ;) )
> >
> > On Sunday, 26 August 2012 00:18:01 UTC+2, mpj wrote:
> >
> > Hey,
> >
> > I wrote Breathe, if you'd be up for providing a short explanation of
> > why you started your project and what advantages you feel your
> > approach has then I'd love to put it in the Breathe readme so that
> > people can see alternatives.
> >
> > Cheers,
> > Michael
> >
> > On Sat, Jun 16, 2012 at 3:59 AM, Anteru
> > > wrote:
> > > Hi,
> > >
> > > We're happy to announce robin, a new Doxygen/C++ to Sphinx bridge.
> > Robin
> > > provides an easy-to-use, easy-to-hack integration of Doxygen
> > > documentation into Sphinx. Robin is licensed under the BSD and can
> be
> > > found at Bitbucket: https://bitbucket.org/reima/robin
> > <https://bitbucket.org/reima/robin>
> > >
> > > Features
> > >
> > >
> > > * Robust extraction of Doxygen XML data via an easy-to-hack parser
> > > * Intermediate data is stored in a database (mongodb) for simple
> > > extraction and processing
> > > * Directive-driven output; each directive provides callbacks and
> > hooks
> > > which allows for deep customization
> > > * Automated generation of driver ReST documents: Similar to
> > automodule;
> > > however, robin generates actual ReST documents which can be
> inspected
> > >
> > > Prerequisites
> > > =