[MonoDevelop] Ideas for GSOC

[Michael Hutchinson]

Hi Krzysztof,

Both the projects you describe would be very useful, but I would
definitely prefer Stetic. It's been in maintenance mode for quite a
while, as it works pretty well as it is, and we have limited
resources, and it's difficult to work on without major refactoring. It
has a number of active users, and it would be great to breath some new
life into it.

The lightweight project model, codebehind generation, and 1
widget-per-file are all closely linked, and essentially all require a
major refactoring of stetic, keeping all the GUI-level code, but
building a new core backed directly onto the MD project model and code
DOM. A plan that's been proposed in the past was to fork stetic into a
new addin, so that backwards compatibility would not be a concern at
all, as both could exist simultaneously (though a migration mechanism
would be needed). Mike Kestner did some initial work on the
1-widget-per-file model in a hack week, which might be useful as a
starting point. Don't underestimate the size of this part, as it
probably requires discarding a third of the stetic code, and mostly
rewriting another third.

Rendering custom widgets and loading custom designers would both
require moving stetic outside the MD process (so it can be unloaded)
and embedding in MD via a GTK Plug/Socket. There's some unfinished
support for this already but I'm not sure why it was ever enabled, and
it might not survive the refactoring. FWIW back when I was working on
ASP.NET support in SoC 2006 I made the ASP.NET designer (also GTK#
widgets) run out-of-process in a week or so, so I don't expect it's
too hard.

Beyond that, a custom designer framework sounds like a good open-ended
task to work on when the core parts are done, if you have time.

- Michael

On Tue, Mar 23, 2010 at 12:37 PM, krzysztof marecki
<marecki.krzysztof at gmail.com> wrote:
> Hello everyone,
>
> I am seasoned C# developer with couple years experience, recently
> helping develop Pinta.
>
> I would like to participate in GSOC. Before I start investigate
> implementation details
>
> needed for GSOC application process, i need some guidance, which idea is
> more
>
> worthy to pursuit. My goal is to improve MonoDevelop usability and
> productivity
>
> needed for writing native Gtk applications.
>
> I have two proposals :
>
> I was thinking about working on Stetic. On projects page there are couple
> tasks which can be
>
> combined into one (1-widget per file, lightweight project model,
> improvements for generating
>
> partial files). Besides them I would try to implement some enhancements
> needed for using
>
> custom widgets. Maybe support for custom designers and editing non simple
> widget properties.
>
> Currently, there is lack of reporting tool for Gtk. I've played a bit with
> http://www.fyireporting.com/
>
>  which is open source C# implementation of Microsoft RDL. It is using
> Apache2 license
>
>  and has good separation between platform independent core and various
>
>  viewers (PDF, Html etc.)  Its engine is well written and capable of
> rendering many reports
>
>  designed in  Sql Business Intelligence Studio.
>
>  Under linux it is using windows forms for its viewer and designer tool. I
> would like to
>
>  implement them natively using Gtk with cairo. Ideally, designer could be
> implemented as
>
>  Monodevelop addin.
>
> I am myself leaning towards working on Stetic because it could be
> beneficiary for more people,
>
> but don't know if that kind of project is feasible for GSOC.
>
> Thanks for your comments and suggestions.
>
> Krzysztof Marecki
>
> _______________________________________________
> Monodevelop-list mailing list
> Monodevelop-list at lists.ximian.com
> http://lists.ximian.com/mailman/listinfo/monodevelop-list
>
>



-- 
Michael Hutchinson
http://mjhutchinson.com