[swig#] Let's start!
Status: Beta
Brought to you by:
mdupont
|
From: Peter G. <pg...@gm...> - 2003-01-31 21:50:10
|
As suggested by Mike I added the following tasks to the task
list :
1. Check in the current source code into cvs
2. Apply the patchs
3. test with compiler mono C#
4. test with compiler pnet / cscc
5. test with compiler Rotor
6. test with compiler DotNET Csc
7. Test with the lib redland
8. Test with the other swig libs (name some here)
9. Create debian package
10. Create web page
11. Create announcement.
However, personally I feel a little uncomfortable with that
list (but lets see as a starter). Do we really need all those
things (some of them I even don't know what it is)?
I'm a little bit concerned that if we don't concentrate on
just one goal, we will not get it done in a timely manner.
PROJECT GOALS:
Before actually starting we should discuss and define the
project goals.
My personal and primary interest is to see a C# module that
produces wrapper code to access an existing API (unmanaged
code) from C#. To achieve this, all I need is a working
csharp.cxx (or whatever you want to call it) module.
Anything else I need is already there. Whenever possible I
will never touch the other existing code. However, to achive
my goal I need at least
1. access to the latest SWIG release and patches,
2. a practical environment that supports a team
to work on the project (the C# module) and
3. a motivated team.
JAVA APPROACH:
The approach to take the Java module as a "template" makes
sense to me since both Java and C# have a lot in common.
This will make it much easier to maintain it. Additions to
the Java module can easily be adopted.
In addition, from what I have seen, Neil's clr.cxx module
provides much more functionality than Feng Songyan's
csharp.cxx module.
One of the most important things I'm missing is support of
callback routines.
TIME FRAME:
Time is limited and important to me. I would like to get it
done within a short time frame. I need it soon. If we can
get it done within one month or so it is o.k. to me.
I think there is a lot that can be done within that time. I
don't expect a perfect and complete implementation within that
time frame, but something that is almost complete and
documented.
Since I have to do something anyway I am willing to spend some
time on this project to get my other work finished faster. I
could implement callback support within a few days.
TESTING:
I'm working on a relatively large project that could provide
a ideal test bed. Except for API's written in C++ there is
nothing I can imagine that is not touched by this project
(the API has been written in native C).
I assume that if it's working for my API it will work for any
other API written in ANSI C and following standard calling
conventions too.
For legal reasons I cannot provide the API, but I could build
some examples working the same way.
LEGAL STUFF:
As I said in earlier messages, as long as I don't violate
anybodys rights I don't care much about the legal things.
HOSTING:
I'm glad we have got a "home". My thanks goes to Mike!
However, if we could get CVS access at swig.org and a
place to put our things I have nothing against it.
MOTIVATION:
First, my personal primary motivation to work on this
project is to get my (primary) job done.
Second, for me it's a new experience. It's the first time
I'm trying to do something with others I never met or have
seen before.
FURTHER PROCEEDINGS:
Please let me know the current status and how you think we
should proceed.
What can you provide at the moment and what are your interests.
What are you interested in doing and how much time can you
spend.
Who cares about the CVS and the patches to be applied (I
already assigned this task to Neil).
Who writes the documentation (sorry my English) and other
things that will be published.
Give me your input so I can setup the task list as acurate
and detailled as possible. Don't forget this will be an
ongoing process.
Always let me know if you need any kind of help or assistance.
Feel free to ask, I will ask too.
Awaiting feedback!
Peter
|