All good:
% svnsync sync file://`pwd`/lala http://localhost/repos/viewcvs
Committed revision 1.
Copied properties for revision 1.
subversion/libsvn_ra_dav/util.c:948: (apr_err=200014)
svnsync: REPORT request failed on 'http://localhost/repos/viewcvs'
subversion/libsvn_repos/commit.c:607: (apr_err=200014)
svnsync: Checksum mismatch for resulting fulltext
(trunk/INSTALL):
expected checksum: bla bla
actual checksum: 084720a16a018c9c8c91a449f6e5f4e9
David Glasser wrote:
> On 11/19/06, glasser@tigris.org <glasser@tigris.org> wrote:
>> In the DAV implementations of svn_ra_replay, make sure to pass the
>> checksum argument to close_file through instead of ignoring. (Note
>> that this is basically untested, because the svn_repos_replay2 editor
>> driver doesn't actually calculate checksums; I would like to fix that,
>> and this is consistent with ra_svn and ra_local at least.)
>
> So I don't have a good apache setup for testiing. Can somebody do me
> a favor and apply the following patch and run svnsync_tests under
> ra_dav and/or ra_serf, and confirm that they fail with a checksum
> error? (That is, confirm that the checksum argument to close_file
> actually does make it through to the client now?) And that it passes
> without these checks?
>
> Thanks!
> dave
>
> === subversion/libsvn_repos/replay.c
> ==================================================================
> --- subversion/libsvn_repos/replay.c (revision 108736)
> +++ subversion/libsvn_repos/replay.c (local)
> @@ -615,7 +615,8 @@
> unsigned char digest[APR_MD5_DIGESTSIZE];
> SVN_ERR(svn_fs_file_md5_checksum(digest, root, path, pool));
> SVN_ERR(editor->close_file(file_baton,
> - svn_md5_digest_to_cstring(digest, pool),
> +/* svn_md5_digest_to_cstring(digest,
> pool), */
> + "bla bla",
> pool));
> }
>
>
>
--
C. Michael Pilato <cmpilato@collab.net>
CollabNet <> www.collab.net <> Distributed Development On Demand
Received on Mon Nov 20 15:38:53 2006