[cvsnt] Re: Question about cvs remove.
mangled at to.avoid.spam
Tue Aug 17 10:23:12 BST 2004
"Tony Hoyle" <tmh at nodomain.org> wrote in message
news:cfsi8c$p9f$1 at paris.nodomain.org...
> > Is there anything that could go wrong with this? I do not see a
> > with a 'disconnected remove' which gets all of its checks done at commit
> > time.
> In theory, but it's just not done that way. Remove does some server
> side stuff so it needs a connection (it does version checks based on the
> repository, checks whether the tag is sticky or not, and then notifies
> the client of its decision).
And if it does require access to the repository, then why does this still
work if I trick it by using an invalid repository but with the same name?
> The sticky tag issue probably isn't solvable without a server
> connection, so it makes sense that it is done this way actually.
But could those checks be delayed until commit if the user so chooses?
More information about the cvsnt