Mailing-List: contact dev-help@subversion.apache.org; run by ezmlm
Precedence: bulk
Received-SPF: neutral (athena.apache.org: local policy)
References: <1417537985.48137.YahooMailNeo@web87704.mail.ir2.yahoo.com>  
Message-ID: <1417715744.62760.YahooMailNeo@web87701.mail.ir2.yahoo.com>
Date: Thu, 4 Dec 2014 17:55:44 +0000
From: Julian Foad <julianfoad@btopenworld.com>
Reply-To: Julian Foad <julianfoad@btopenworld.com>
Subject: Re: Issue #4476 - Mergeinfo containing r0 makes svnsync and svnadmin
 dump fail
To: "dev@subversion.apache.org" <dev@subversion.apache.org>
In-Reply-To: <1417537985.48137.YahooMailNeo@web87704.mail.ir2.yahoo.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable

Julian Foad wrote:=0A> http://subversion.tigris.org/issues/show_bug.cgi?id=
=3D4476=0A[...]=0A> PROPOSAL=0A> =0A> In the presence of a mergeinfo proper=
ty that refers to r0, at the libsvn_repos =0A> API:=0A> =0A> 'dump'=0A> =A0=
 shall dump the property verbatim (with nowarning). (Done.)=0A> =0A> 'load'=
, with 'validate properties' OFF and not adjusting =0A> revision numbers or=
 paths,=0A> =A0 shall load theproperty verbatim, and warn.=0A> =0A> 'load',=
 with 'validate properties' OFF and adjusting revision =0A> numbers or path=
s,=0A> =A0 shall load the property verbatim (unadjusted), and warn.=0A> =0A=
> 'load', with 'validate properties' ON,=0A> =A0 shall fail.=0A> =0A> And I=
 would say the same rules should apply to any property that fails the =0A> =
validation rules.=0A> =0A> At the application layer:=0A> =0A> 'svnadmin dum=
p' and 'svnadmin load'=0A> =A0 should behave the same as the repos layer.=
=0A> =0A> Does that sound good as a fix that I can do now and back-port to =
1.8.x and =0A> 1.7.x?=0A=0AI have done this, and proposed it for backport t=
o 1.7 and 1.8.=0A=0A(My fix adds a new enumeration constant to an enumerati=
on type, for the warning. For back-porting, I renamed this constant to a do=
uble-underscore name to indicate it's not for public use in the 1.7/1.8 API=
s.)=0A=0A> I'll come to 'svnsync' later, but basically my current thought i=
s it =0A> should 'correct' the mergeinfo by removing the r0 reference.=0A=
=0AI have written a test for this, and hacked up code to do this by textual=
 substitution. It isn't quite right -- for example it would go wrong if a p=
ath contained the character sequence ":0", among other issues. I ought to r=
ewrite it in the form of a proper mergeinfo parser but one that is more len=
ient than the regular parser.=0A=0AI also noticed that loading from a dump =
file also modifies mergeinfo in other ways such as removing references to r=
1. I don't think is a good idea: I think there should be a way to load exac=
tly what is in the dump. But that's a different, though related, issue.=0A=
=0A- Julian