Target for 1.0

Well, I just decided that the code completion provided for now is not good enough for a 1.0 version.

That means that for the 1.0 version to be released, not only should it be stable, but code completion should not be based on a python shell (even though it would need to get some info from the shell, as getting completions from a dll or other modules existant only in the c level would be needed).

So, the only way is to use a mixed approach... analyzing source files and compiled modules...

Target is:
- 0.8.1: completion for imports without shell (definition of pythonpath on project properties) - I might take out pylint, as it is still not stable... that's not because of support, because Sylvain Thénault from logilab has been very helpful on its problems, it is just that I can't find time to do it, and a really good code completion is my highest priority right now...
- 0.8.2: full code completion for source files
- 0.9.0: mixed code completion using source and compiled extensions

1.0 will be the "stable version" with the above features
this probably will give enough time for Grig to do the pyunit integration too...