monotone-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Monotone-devel] Approval revisited...


From: Richard Levitte - VMS Whacker
Subject: [Monotone-devel] Approval revisited...
Date: Fri, 10 Feb 2006 17:25:36 +0100 (CET)

Hi,

I'm taking a look at the current revision approval possibilities, and
there are things I don't quite understand.  Also, it looks like this
hasn't been looked at for ages.

First of all, we probably need to rewrite the example for the
get_revision_cert_trust, as it currently uses the "ancestor" cert,
which hasn't existed since versoin 0.15 (when we switched to
changesets).

Second, the approve command adds a branch cert, and I would like for
someone to explain the rationale behind that.  I assume that the
branch name is meant to express a little more exactly what is being
approved, could that be correct?  It seems to me that setting an
explicit branch might even be a bit dangerous.  If I decided to
approve monotone with a special branch, for example
net.venge.monotone.approved.linux, it would mean that we end up with a
branch that seems to have multiple heads or something like that.  It
would be a very disconnect graph at the very least.

It would be interesting to know how many are actually using these two
features.  Does anyone use the approve command?  Does anyone use the
get_revision_cert_trust hook?  If there was a change in those, would
anyone get hurt?

I've a proposal:

1. change the approve command to add a "approved" cert rather than a
   "branch" one.  Other than that, it works exactly the same way as
   before.
2. rewrite the example for the get_revision_cert_trust to look at a
   "approved" cert.

Comments?

Cheers,
Richard

-----
Please consider sponsoring my work on free software.
See http://www.free.lp.se/sponsoring.html for details.

-- 
Richard Levitte         address@hidden
OpenSSL Project         http://www.openssl.org/~levitte/




reply via email to

[Prev in Thread] Current Thread [Next in Thread]