dolfin team mailing list archive

[Johan Hake <johan.hake@xxxxxxxxx>]

On Thursday October 7 2010 14:28:25 Kristian Ølgaard wrote:
> On 7 October 2010 22:44, Johan Hake <johan.hake@xxxxxxxxx> wrote:
> > On Thursday October 7 2010 13:06:22 Kristian Ølgaard wrote:
> >> This bug should be fixed now. It was introduced when commenting the
> >> line:
> >> 
> >> //%feature("autodoc", "1");
> >> 
> >> in dolfin/swig/dolfin.i
> >> 
> >> I did this to avoid the Swig generated docstring signatures which make
> >> Sphinx unhappy, formatting issues, when building the programmer's
> >> reference in fenics-docs.
> >> 
> >> For some reason the class extensions in the *_post.i files where not
> >> added to the cpp.py module by Swig if the 'autodoc' line was absent.
> > 
> >> However, by adding lines like:
> > This does not make sense. I tested on my computer and the _pre and _post
> > files get included.
> 
> What I meant was that the code in the extensions are not added to the
> code for the classes in cpp.py.
> The line:
> Progress.__set = new_instancemethod(_cpp.Progress___set,None,Progress)
> 
> is there alright.
> 
> If the docstrings are switched on, I both get the code in the class
> and the new_instancemethod line.

Ok.

> > The trouble with the buildbot is that it is not happy with extending
> > classes using double underscore attributes. This is not a problem when
> > the method is added to the Python proxy in cpp.py, as it is when an
> > automatic docstring is added. But when these are turned off it gets
> > added using:
> > 
> >  Progress.__set = new_instancemethod(_cpp.Progress___set,None,Progress)
> 
> as I point out above, this line is always there, at least when I build
> on my machine.
> 
> > it seems like the Python interpreter get confused and cannot find the
> > "private" attribute of __set in this case.
> > 
> > Changing __set to _set in log_post.i fixed the problem.
> 
> OK.
> 
> >> %feature("docstring") dolfin::Array::array "Missing docstring";
> >> 
> >> before the extensions the problem disappeared.
> > 
> > I guess I can revert these changes, or do you need them for other stuff?
> 
> We need them if we want the docstrings :) For the private functions it
> doesn't matter, but some functions are public so we should provide
> docstrings for those.

Nice! Now I see what you have done ;) You are able to include a stub to C++ 
extended classes.

But do you need docstrings on protected methods? Is this needed to Sphinx does 
not complain?

Johan

> Kristian
> 
> > Johan
> > 
> >> Kristian
> >> 
> >> On 7 October 2010 01:51,  <buildbot@xxxxxxxxxx> wrote:
> >> > The Buildbot has detected a new failure of dolfin-jaunty-amd64 on
> >> > FEniCS Buildbot. Full details are available at:
> >> >  http://fenicsproject.org:8080/builders/dolfin-jaunty-amd64/builds/977
> >> > 
> >> > Buildbot URL: http://fenicsproject.org:8080/
> >> > 
> >> > Buildslave for this Build: jaunty-amd64
> >> > 
> >> > Build Reason:
> >> > Build Source Stamp: HEAD
> >> > Blamelist: Kristian B. =C3=98lgaard
> >> > 
> >> > BUILD FAILED: failed dolfin check
> >> > 
> >> > sincerely,
> >> >  -The Buildbot
> >> > 
> >> > 
> >> > _______________________________________________
> >> > Mailing list: https://launchpad.net/~dolfin
> >> > Post to     : dolfin@xxxxxxxxxxxxxxxxxxx
> >> > Unsubscribe : https://launchpad.net/~dolfin
> >> > More help   : https://help.launchpad.net/ListHelp
> >> 
> >> _______________________________________________
> >> Mailing list: https://launchpad.net/~dolfin
> >> Post to     : dolfin@xxxxxxxxxxxxxxxxxxx
> >> Unsubscribe : https://launchpad.net/~dolfin
> >> More help   : https://help.launchpad.net/ListHelp