why doing RPackage is a pain

RPackage is nice simple fast and robust now the integration within categories crap and packageInfo and PackageOrganizer and MCWorkingCopies....
are a pain.

Examples:

MCWorkingCopy allManagers size 177
PackageOrganizer default packages size 422

Stef :(

On Oct 16, 2011, at 9:21 PM, Stéphane Ducasse wrote:

> RPackage is nice simple fast and robust now the integration within categories crap and packageInfo and PackageOrganizer and MCWorkingCopies....
> are a pain.
>
> Examples:
>
> MCWorkingCopy allManagers size 177
> PackageOrganizer default packages size 422
>
> Stef :(


I imagine that I could only look at MCWorkingCopies.
Now I have not why wrong time to time I get a new package in Monticello-Browser that is not a package but just
the left most part of a package name

I get
        RPackage ()
        RPackage-Core (StephaneDucasse.107)

but I never mentioned that RPackage was a package (since it is not).

:(

Stef

setUp

       "FIXME: Unregister Monticellomocks if it got created in another test
       (for example MCMethodDefinitionTest may create it implicitly).
       This avoids a nasty failure of MCChangeNotificationTest due to
       some inconsistency about the package info registered with it.
       If Monticellomocks was created earlier it will contain a 'regular'
       PackageInfo instance but the test requires it to be an MCMockPackageInfo"
       MCWorkingCopy registry
               removeKey: (MCPackage new name: 'Monticellomocks')
               ifAbsent:[].
       PackageOrganizer default unregisterPackageNamed: 'Monticellomocks'.

       workingCopy := MCWorkingCopy forPackage: self mockPackage.




PackageOrganizer>>unregisterPackageNamed: aString
        self unregisterPackage: (self packageNamed: aString ifAbsent: [^ self])

so this can explain why PackageOrganizer contains 'Monticellomocks' while it should not :(

Stef