Re: [Monotone-devel] Re: Re: Re: How will policy branches work?

[Markus Schiltknecht]

Hi,

Zack Weinberg wrote:
> On Wed, Feb 6, 2008 at 10:42 AM, Boris <address@hidden> wrote:
> >  The problem is the co-worker shouldn't be allowed to check out the files
> >  at all.

That's what I'm referring to with the "partial checkout" thing. It's 
currently not possible with monotone. And it would certainly require 
changes to netsync, if you want to prevent the files from going over the 
wire. It's a different kind of partial pull: not restricted to parts of 
all revisions, but restricted to some files within the revisions.

That's not very likely to get implemented anytime soon in monotone. If 
at all.

Instead, better create completely different branches for these modules. 
Possibly with propagating from the lesser critical branch to the the 
activation components branch, but certainly not the other way around.

> What is the rationale for this requirement?  My knee-jerk reaction is
> that this is ultimately impossible to enforce (untrusted dev A can go
> over to trusted dev B and ask to be shown), counterproductive (people
> code better when they know what the code on the other side of the
> fence is like), and not something we *ought* to support.  But maybe I
> would think differently if I knew the situation you're in.

While I agree, I don't think this is convincing managers. After all, 
there are reasons for proprietary software to still exist. And I'd 
rather like monotone to be appealing to business as well.

> Now, not filling up disk space *by default* with code that dev A has
> no intention of looking at is a different story, and as I said
> upthread, we can already do that (netsync branch patterns).

Agreed.

Markus