[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: r39592

From: Joe Swatosh <joe.swatosh_at_gmail.com>
Date: Sun, 18 Oct 2009 08:24:37 -0700

On Sat, Oct 17, 2009 at 8:39 AM, Bert Huijben <bert_at_qqmail.nl> wrote:
>
>
>> -----Original Message-----

>>
>
>        Hi,
>
> Thanks for looking into this.
>
> I'll verify the 1.6 behavior and I will change the depth unknown behavior to
> infinity after confirming your findings.
>

Great (and already done!)

>
> BTW. There was some on-irc talk about the svn_wc_ensure_adm() behavior on no
> longer allowing NULL uuid and repository root.
>
> The idea was (if I remember correctly) to disallow uuid and repository empty
> on a completely standalone working copies (like the one in the ruby test),
> but to try copying the data from its parent directory if that is a working
> copy. (And probably write an erratum to explain the behavior change).
>
> I don't think the copying from parent part is implemented yet, but the WC-NG
> data model doesn't allow working copies without repository/uuid. So it would
> be very hard to support the old behavior.. Which would give you a mostly
> unusable working copy anyway.
>

Thanks for the update, Bert. I don't have any skin in this game: I can make
the Ruby bindings tests pass as things stand now, if that is the
descision. But
I'm glad to hear it is getting attention because (from my my naive
POV) it would
seem to break the backward compatibility guarantees. Not that breaking them
can't be done, but that it be done thoughtfully.

Thanks again for the quick response.

--
Joe
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2408672
Received on 2009-10-18 17:24:56 CEST

This is an archived mail posted to the Subversion Dev mailing list.