Great I will try
> Remove calls to #inset:. If your asthetic sense cries out, replacing it
> with #borderWidth: might help a little, but it works like that. Got the
> ModulesBrowser to work for me. Well, to load. Caused another walkback
> pretty quick.
> 
> Another something that helps - it seems the sources are filled with lfs
> and crlfs in all the wrong places. (Could this be a packaging problem?
> Stef, does your code have this problem?)

This was in june 29 code. So I will change that.

> After changing all those around to Squeaks preffered style, I got my
> menus back and the tools worked some more.

Could you send me the files you have so that I can do another version?

I will not be able to work today on it. Could you change the files I
provided, include your ideas, fixes and upload at SqueakFoundation so that I
can play with it. This would be great.



> I can get into all the packages in the package tree, but some of them
> are empty. For example, the Refactory. And I can't find it in the
> unmanaged code either. Very mysterious. BTW, this is after building the
> modules, removing the RB tests, and rebuilding the modules again. Any
> ideas?


No this is me that mess up the definition of the modules because when I
started I could not filled RB ;)

Look in the class ModuleBuilder and change the definitions of the packages
You will see that I commented out a lot of modules because I just wanted to
try. In the list you may renamed classes so I added a method
checkIfClassExist that produces a trace in the transcript to help the
definition

I'm not sure but in ModuleBuilder it seems to have functionaility to
reproduce the module specification.



> Anyway, I can look at code in the ModulesBrowser. Two problems now - the
> submenus cause walkbacks. File in the attached file, that should fix it
> (though it's a hack - probably the client code in the tool should be
> fixed instead. For now...). And there's a silly text morph below the
> code pane. What is that?
> 
>> The zipped file contains everything to build Ginus from a Squeak3.0-4332
>> image. You should load first the RB paroject and remove the test of RB else
>> this is toolong ;)
> Damn, didn't understand what you meant until after I'd made that
> mistake... :-)
> 
>> I tested the zipped file and recreated everything well so normally you
>> should not have problem. However, I will try to updload my images to the
>> SqueakFoundation Wiki but my bandwidth is small ;). I tired but failed
>> several times ;(
> Instead of including the RB project, upload only your code. People can
> load the RB by doing -
> 
> Project fromUrl: 
> 'http://squeak.heeg.de:8080/TFODC69UQ0MFRIHMG40DRML1CU7VYZ9-3/Squeak%20R
> B.pr.gz'

ok I will do a light version

>> Things that have to be done:
>> - fix the tests 
> Using SUnit 3.0 from SourceForge brings the error down to 26! it's an
> easy step forward! ;-)

Ok 


> 
>> - complement the semanticl model (or find a way to remove it I was
>> wondering if we could not have a globalVariableDefinition and
>> ClassDefinition in palin Smalltalk but causally connected).
> !? remove it? I heard you say that before. Why? being able to work on
> unloaded code sounds nice. What's the bad side?
I would like to discuss this topic. Roel told me that this is good, joseph
too, but I'm not sure. That's it.

>> - fix the UI so that we can use the tools
> Not that hard...
> 
>> - work on the versioning
> What exactly is missing? where do you see it?

We need a CVS repository or anythign that can play this roles

 
>> Stef
>> 
>> So I'm waiting for your feedback.
> I like what I see for now. Still need to test what those repository
> related menu items do...