REMINDER: Final review of VCS interfaces

[Andreas Pakulat]

On 23.05.07 14:01:05, Matthew Woehlke wrote:
> Andreas Pakulat wrote:
> > On 23.05.07 12:06:25, Matthew Woehlke wrote:
> >> Andreas Pakulat wrote:
> >>> On 23.05.07 04:06:12, dukju ahn wrote:
> > Also CVS does this only on local files too, I'm tempted to but it into
> > IBasicVC and take it local files only (remote stuff can be done in
> > IFooVcs), unless somebody knows a VCS that doesn't have annotate
> > (eclipse has this too in its classes, IIRC, which suggests that quite a
> > bunch have this feature).
> 
> Sounds OK to me. Want me to write it?

That would be wonderful, I doubt I have the time today or tomorrow.

> > About return value: cvs and svn can return revision, date and author,
> > does that fit into one of the classes used for change/log?
> 
> If you mean for change()/log(), these return VcsEvent, which has user, 
> date, revision and revision-per-file (also other stuff like message). Or 
> did you mean for annotate()?

Nevermind, I think we need a separate class for that, which encapsulates
the lines of a file and lets the user fetch date,author and revision for
each line.

> > Oh and a last one: does perforce support specifying a revision? CVS, svn
> > and git support that, clearcase seems to not support it (at least
> > there's no option for that). Does it make sense that the plugin ignores
> > the revision if it the system doesn't support that? Or is that too much
> > magic behind the scenes?
> 
> What function(s) are you asking about? If a VCS might ignore something 
> it should be documented (we already have some of that in log()).

I'm talking about annotate().

Andreas

-- 
Of course you have a purpose -- to find a purpose.