[buckminster-dev] Re: Problems getting started with Headless Buckminster

["Mark Phippard" <markp@xxxxxxxxxxxxxxx>]

Thomas Hallgren wrote:

> Hi Mark,
> 
> Mark Phippard wrote:
> > It looks like the dependency exists so that our BaseResourceStorage
> > class can implement the
> > org.eclipse.compare.IEncodedStreamContentAccessor interface.  The
> > annotation for this change says it was so that we could properly set
> > the encoding on files when doing compares from the Synch view.
> > 
> >  
> I made an experiment and removed that interface from the class in
> question together with the dependency it introduces to
> org.eclipse.compare. After that, the core plugin runs fine without
> any UI components.
> 
> I would like to ask you to move the encoding support into your UI
> component. Nothing in the core component calls on any functionality
> in the org.eclipse.compare component so from the looks of it, such a
> move should be possible. Perhaps a wrapper for the
> BaseResourceStorage could be added to the UI component that
> implements the interface?
> 
> I'm not sure what to do with our broken Buckminster headless svn
> support at this point. One idea is to create a special version of the
> org.tigris.subversion.subclipse.core where the dependency is removed
> and then bundle that with our headless svn feature. Not ideal since I
> then take on the burden of maintaining it. Another is to rewrite our
> plugin to go directly to the client libs. Lot's of work that produces
> code that also must be maintained. A third alternative, a patch from
> Subclipse that removes the problem altogether, would of course make
> life a lot simpler :-)

I have no problem with making the right change, I just do not know what
it is and do not personally have the time to investigate.  This change
was introduced back in 2005 by one of our committers from Japan.
Obviously the encoding of files is more sensitive for them, as well as
easier to see problems.  I would not take the fact that you did not
observe any problems as "gold" that there are not any problems.  The
interface in question just exposes the same method as IEncodedStorage
but I assume that the compare plugin must be checking to see if the
class is implementing this interface or not.

Mark