RE: Re: Re: ISVNClientAdapter.diff - how are the paths calculated?

Robert Munteanu <[email protected]> Thu, 8 May 2014 14:43:04 -0700 (PDT)
Newsgroups gmane.comp.version-control.subversion.subclipse.user
Message-ID <2064986686.4073.1399585384027.JavaMail.httpd@localhost>
> On Sun, May 4, 2014 at 3:26 AM, Robert Munteanu
> <robert dot munteanu at gmail dot com>wrote:
> 
> > Thanks for the reply.
> >
> > createPatch has the same problem - I only looked into more low-level
> > methods when createMethod started failing.
> >
> > I tried debugging this myself, but I quickly got lost in SVNKit internals
> > when doing so.
> >
> >
> Have you tried the Subclipse Create Patch option to compare results?  It
> ultimately calls ISVNClientAdapter.createPatch().  It sets the relativeTo
> path based on what the user selected in the UI.  See the
> GenerateDiffFileOperation class in Subclipse UI.

Just tried it and it behaves the same. Not sure if it qualifies as a bug or not ( SVNKit 1.7.9.9659 ).

Anyway, right now the simplest way out is to create a diff file-by-file and then rewrite the patch headers, since I already know the files and the repo location.

Thanks,

Robert

> 
> Subclipse is written to SVNClientAdapter which is ultimately written to
> JavaHL.  We use SVNKit based on its implementation of the JavaHL interface
> and rely on them to make sure it behaves the same as JavaHL on all methods.
>  If SVNKit behaves different, I would assume that would show up when
> creating a patch via Subclipse as well.
> 
> 
> 
> -- 
> Thanks
> 
> Mark Phippard
> http://markphip.blogspot.com/

------------------------------------------------------
http://subclipse.tigris.org/ds/viewMessage.do?dsForumId=1047&dsMessageId=3077885

To unsubscribe from this discussion, e-mail: [[email protected]].