Re: [Monotone-devel] Re: version 1.14 fails "make check"

[Richard Levitte - VMS Whacker]

In message <address@hidden> on 16 Aug 2004 13:47:11 +0200, Peter Simons 
<address@hidden> said:

simons> Richard Levitte writes:
simons> 
simons>  > You need the following:
simons> 
simons>  >  1. the branch name
simons>  >  2. the key that signed the branch cert
simons>  >  3. the list of <branch,key> trust tupples.
simons> 
simons>  > You get 1 and 2 from the branch cert.
simons> 
simons> What if the certificate -- for which I have to decide
simons> whether to trust it or not -- is not a BRANCH certificate? I
simons> would like to accept a, say, APPROVE signed by Graydon in
simons> the net.venge.monotone branch, but not in the
simons> simons.super.duper.private.stuff branch. How do I implement
simons> that?

Ah, you don't today.  That would require get_manifest_cert_trust to
have a different set of arguments.  Basically, "signers", "name" and
"val" would have to be replaced with a table of triples <signers,
name, val>.

Maybe this is something for Graydon to think about...

simons> What about ANCESTOR, for instance? Won't trusted ANCESTOR
simons> certificates be used for 3-way merges, regardless of the
simons> branch they are in? In that case, trusting (or not trusting)
simons> BRANCH won't be good enough.
simons> 
simons> Or do I misunderstand something?

Probably not, but understand that you previously only talked about
branches.  As of this email from you, we're getting into a whole
different level of difficulty.  Or did *I* misunderstand something?

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

-- 
Richard Levitte                         address@hidden
                                        http://richard.levitte.org/