Re: add/remove against a read-only repository.

[Sergei Organov]

"Jim.Hyslop" <address@hidden> writes:
> Sergei Organov wrote:
> > CVS (1.11.1p1) doesn't seem to allow add/remove to be made against a
> > read-only repository. This behavior makes it inconvenient to work with
> > such repositories (I mean public CVS repositories of open-source
> > projects). In particular, locally rm'ed files re-appear as soon as update
> > is invoked, and contents of locally added files don't appear in the 'cvs
> > diff'.
> > 
> > At first glance it seems that adding/removing files is a local activity
> > similar to other changes and thus shouldn't require write access to the
> > repository until the time the changes are committed. Is the idea
> > fundamentally broken, just difficult to implement, or what?
>
> Sounds to me like CVS is working perfectly well. What behaviour were you
> expecting?

I'm expecting to add/remove files in my working copy and issue "cvs add"
and/or "cvs remove" to let CVS know the changes has been made *locally*. Then,
only if and when I attempt "cvs commit" CVS should complain the repository
is read-only.

> 
> Let's take a step back here. 
> 
> Why do you want to modify someone else's repository?

I don't, -- that's the issue. I wish to modify my local copy and still be able
to do "cvs update" so that removed files don't re-appear and to do "cvs diff"
to generate a correct patch against repository.

> Instead of checking out the source code, you could export it, then import it
> into a local repository.

That's exactly what I wish to avoid. While I'm working on the patch it's *way
simpler* to periodically do "cvs update" against the master read-only
repository than to re-import from it to the local repository. Just take a look
at the FAQ for all the tricks that are required to correctly re-import complex
repository to see the point.

> There, you can do what you want.

I know, but there are a lot of cases when it's just very inconvenient.

-- 
Sergei.