You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Magnus Torfason <zu...@gmail.com> on 2009/03/05 18:03:06 UTC
Minor regression in svnsync 1.6rc3 (from 1.5.6)
I just tried syncing my repository to a fresh 1.6rc3 build and got the
following message:
...
Committed revision 75.
Copied properties for revision 75.
Committed revision 76.
Copied properties for revision 76.
svnsync: Cannot accept non-LF line endings in 'svn:log' property
Turns out that these very old log entries had something strange going on
in the line endings in the log messages. The workaround was to go back
and change them (I did it manually using TortoiseSVN). Trivial really in
my case.
The only reason I mention this is that I retested this on a 1.5.6 build
(both are running on Linux 2.6.* kernels, although not on the same
machine), and that worked without a hitch.
My guess is that a test was added in 1.6 (A Good Thing), but perhaps the
consequences for svnsync were unforeseen? I could imagine scenarios
where the required changing of log messages would be nontrivial.
Best,
Magnus
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1273107
Re: Minor regression in svnsync 1.6rc3 (from 1.5.6)
Posted by Magnus Torfason <zu...@gmail.com>.
On 3/5/2009 1:31 PM, C. Michael Pilato wrote:
> It's a bugfix with a loud bit of motivational messaging for cleanup. :-)
On 3/5/2009 1:34 PM, Hyrum K. Wright wrote:
> This is a feature, not a bug. :)
'Nuff said :-)
Best,
Magnus
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1273212
Re: Minor regression in svnsync 1.6rc3 (from 1.5.6)
Posted by "C. Michael Pilato" <cm...@collab.net>.
Magnus Torfason wrote:
> I just tried syncing my repository to a fresh 1.6rc3 build and got the
> following message:
>
> ...
> Committed revision 75.
> Copied properties for revision 75.
> Committed revision 76.
> Copied properties for revision 76.
> svnsync: Cannot accept non-LF line endings in 'svn:log' property
>
> Turns out that these very old log entries had something strange going on
> in the line endings in the log messages. The workaround was to go back
> and change them (I did it manually using TortoiseSVN). Trivial really in
> my case.
>
> The only reason I mention this is that I retested this on a 1.5.6 build
> (both are running on Linux 2.6.* kernels, although not on the same
> machine), and that worked without a hitch.
>
> My guess is that a test was added in 1.6 (A Good Thing), but perhaps the
> consequences for svnsync were unforeseen? I could imagine scenarios
> where the required changing of log messages would be nontrivial.
I've survive such a required change. I automated it with a loop around
svnsync that, upon error, would use 'svn pget' and 'svn pset' to fix the
messed-up log message, and then resume the synsync loop.
I wouldn't call this a regression really. It's a bugfix with a loud bit of
motivational messaging for cleanup. :-)
--
C. Michael Pilato <cm...@collab.net>
CollabNet <> www.collab.net <> Distributed Development On Demand
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1273181
Re: Minor regression in svnsync 1.6rc3 (from 1.5.6)
Posted by "Hyrum K. Wright" <hy...@mail.utexas.edu>.
On Mar 5, 2009, at 12:03 PM, Magnus Torfason wrote:
> I just tried syncing my repository to a fresh 1.6rc3 build and got the
> following message:
>
> ...
> Committed revision 75.
> Copied properties for revision 75.
> Committed revision 76.
> Copied properties for revision 76.
> svnsync: Cannot accept non-LF line endings in 'svn:log' property
>
> Turns out that these very old log entries had something strange
> going on
> in the line endings in the log messages. The workaround was to go back
> and change them (I did it manually using TortoiseSVN). Trivial
> really in
> my case.
>
> The only reason I mention this is that I retested this on a 1.5.6
> build
> (both are running on Linux 2.6.* kernels, although not on the same
> machine), and that worked without a hitch.
>
> My guess is that a test was added in 1.6 (A Good Thing), but perhaps
> the
> consequences for svnsync were unforeseen? I could imagine scenarios
> where the required changing of log messages would be nontrivial.
This is a feature, not a bug. :)
We actually saw this on our own repository several months ago, and
have since fixed the log messages to be compliant. I suspect we'll
see this type of complaint a lot on user@ once 1.6 is released. I
don't recall what the official stance is on this, but we'd probably
ought to mention it on the Release Notes.
-Hyrum
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1273190