You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by David Summers <da...@summersoft.fay.ar.us> on 2003/03/17 20:15:22 UTC
RA_SVN TEST FAILURE: log_tests.py 3: test 'svn log' on an empty
repository
Is anyone else seeing this? I've been seeing it for the last couple of
days.
This run is on Rev 5363.
CMD: svnadmin "create" "repositories/log_tests-3" "--bdb-txn-nosync" <TIME
= 0.250673>
CMD: svn "log" "--username" "jrandom" "--password" "rayjandom"
"svn://localhost/repositories/log_tests-3" <TIME = 0.070540>
svn: Network connection closed unexpectedly
svn: Connection closed unexpectedly
WARNING: Test driver returned a status code
FAIL: log_tests.py 3: test 'svn log' on an empty repository
--
David Wayne Summers "Linux: Because reboots are for hardware upgrades!"
david@summersoft.fay.ar.us PGP Key: http://summersoft.fay.ar.us/~david/pgp.txt
PGP Key fingerprint = C0 E0 4F 50 DD A9 B6 2B 60 A1 31 7E D2 28 6D A8
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: RA_SVN TEST FAILURE: log_tests.py 3: test 'svn log' on an empty repository
Posted by Greg Stein <gs...@lyra.org>.
On Tue, Mar 18, 2003 at 12:38:51AM -0500, Greg Hudson wrote:
> On Mon, 2003-03-17 at 22:48, Karl Fogel wrote:
> > It would be nice to set up ra_svn at svn.collab.net, to serve the
> > Subversion repository. The only thing lacking is some sort of simple
> > auth system (like the http basic auth we use with ra_dav)... Or am I
> > behind the times, and ra_svn is capable of this now?
>
> If Collabnet want to set up ssh accounts for developers, we can do
> tunneled ra_svn.
Unfortunately, no... I don't think that's really doable in our current
environment.
Cheers,
-g
--
Greg Stein, http://www.lyra.org/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: RA_SVN TEST FAILURE: log_tests.py 3: test 'svn log' on an empty
repository
Posted by Greg Hudson <gh...@MIT.EDU>.
On Mon, 2003-03-17 at 22:48, Karl Fogel wrote:
> It would be nice to set up ra_svn at svn.collab.net, to serve the
> Subversion repository. The only thing lacking is some sort of simple
> auth system (like the http basic auth we use with ra_dav)... Or am I
> behind the times, and ra_svn is capable of this now?
If Collabnet want to set up ssh accounts for developers, we can do
tunneled ra_svn.
If not, best to wait for Cyrus SASL integration. If we implement plain
or md5 auth ourselves, it won't interoperate with the Cyrus SASL user
database files in the future.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: RA_SVN TEST FAILURE: log_tests.py 3: test 'svn log' on an empty repository
Posted by Karl Fogel <kf...@newton.ch.collab.net>.
Greg Hudson <gh...@MIT.EDU> writes:
> Hm, that's not exactly how I envisioned fixing the error reporting issue
> for "svn log" (I had figured each log message would go in a command
> response), but I can't see any reason why my way would be better. So
> the patch looks fine.
>
> It's a lot easier if we don't worry about ra_svn compatibility at this
> stage. (Since we're not eating the ra_svn dog food, it's less of a pain
> to deal with incompatibilities, I think. I haven't seen any user
> complaints about them.)
It would be nice to set up ra_svn at svn.collab.net, to serve the
Subversion repository. The only thing lacking is some sort of simple
auth system (like the http basic auth we use with ra_dav)... Or am I
behind the times, and ra_svn is capable of this now?
I can't think of any reason why ra_svn and ra_dav wouldn't play nice
with each other.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: RA_SVN TEST FAILURE: log_tests.py 3: test 'svn log' on an empty
repository
Posted by Greg Hudson <gh...@MIT.EDU>.
Hm, that's not exactly how I envisioned fixing the error reporting issue
for "svn log" (I had figured each log message would go in a command
response), but I can't see any reason why my way would be better. So
the patch looks fine.
It's a lot easier if we don't worry about ra_svn compatibility at this
stage. (Since we're not eating the ra_svn dog food, it's less of a pain
to deal with incompatibilities, I think. I haven't seen any user
complaints about them.) If we were fixing this after beta or so, we'd
use a capability to indicate that the client and server both understand
the new style of log response, and transmit/read the new style only if
the other side supports it.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: RA_SVN TEST FAILURE: log_tests.py 3: test 'svn log' on an empty repository
Posted by Karl Fogel <kf...@newton.ch.collab.net>.
Matt Kraai <kr...@alumni.cmu.edu> writes:
> ra_svn isn't passing the error code from svn_repos_get_logs to the
> client. The following patch fixes it but it is not compatible
> with older servers and clients. Does the Compatibility section of
> HACKING refer to ra_svn as well as ra_dav?
Haven't looked over the patch yet, but to answer your question
immediately: I had originally our compatibility concerns should apply
to all RA layers, but at one point Greg Hudson posted saying he
personally didn't plan to worry about ra_svn compatibility issues much
at this stage. I don't know if that's still true; Greg?
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: RA_SVN TEST FAILURE: log_tests.py 3: test 'svn log' on an empty repository
Posted by Matt Kraai <kr...@alumni.cmu.edu>.
On Mon, Mar 17, 2003 at 03:14:18PM -0600, Karl Fogel wrote:
> David Summers <da...@summersoft.fay.ar.us> writes:
> > Is anyone else seeing this? I've been seeing it for the last couple of
> > days.
> >
> > This run is on Rev 5363.
>
> Yes, I'm seeing this too now, only over ra_svn.
>
> I need to finish up issue #630 right now, so I won't get a chance to
> look at this until that's done. Hope someone else (hint hint) will be
> seized by a desire to have a whack at it before then...
ra_svn isn't passing the error code from svn_repos_get_logs to the
client. The following patch fixes it but it is not compatible
with older servers and clients. Does the Compatibility section of
HACKING refer to ra_svn as well as ra_dav?
* subversion/libsvn_ra_svn/client.c
(svn_ra_log): Process response after log message data.
* subversion/svnserve/serve.c
(log_cmd): Send response after log message data.
Index: subversion/libsvn_ra_svn/client.c
===================================================================
--- subversion/libsvn_ra_svn/client.c (revision 5365)
+++ subversion/libsvn_ra_svn/client.c (working copy)
@@ -789,9 +789,6 @@
SVN_ERR(svn_ra_svn_end_list(conn, pool));
SVN_ERR(svn_ra_svn_end_list(conn, pool));
- /* Read the response. */
- SVN_ERR(svn_ra_svn_read_cmd_response(conn, pool, ""));
-
/* Read the log messages. */
subpool = svn_pool_create(pool);
while (1)
@@ -832,6 +829,10 @@
apr_pool_clear(subpool);
}
apr_pool_destroy(subpool);
+
+ /* Read the response. */
+ SVN_ERR(svn_ra_svn_read_cmd_response(conn, pool, ""));
+
return SVN_NO_ERROR;
}
Index: subversion/svnserve/serve.c
===================================================================
--- subversion/svnserve/serve.c (revision 5365)
+++ subversion/svnserve/serve.c (working copy)
@@ -761,6 +761,7 @@
static svn_error_t *log_cmd(svn_ra_svn_conn_t *conn, apr_pool_t *pool,
apr_array_header_t *params, void *baton)
{
+ svn_error_t *err;
server_baton_t *b = baton;
svn_revnum_t start_rev, end_rev;
const char *full_path;
@@ -785,17 +786,17 @@
*((const char **) apr_array_push(full_paths)) = full_path;
}
- /* Write an empty command-reponse, telling the client logs are coming. */
- SVN_ERR(svn_ra_svn_write_cmd_response(conn, pool, ""));
-
/* Get logs. (Can't report errors back to the client at this point.) */
lb.fs_path = b->fs_path;
lb.conn = conn;
- SVN_ERR(svn_repos_get_logs(b->repos, full_paths, start_rev, end_rev,
- changed_paths, strict_node, log_receiver, &lb,
- pool));
+ err = svn_repos_get_logs(b->repos, full_paths, start_rev, end_rev,
+ changed_paths, strict_node, log_receiver, &lb,
+ pool);
- return svn_ra_svn_write_word(conn, pool, "done");
+ SVN_ERR(svn_ra_svn_write_word(conn, pool, "done"));
+ SVN_CMD_ERR(err);
+ SVN_ERR(svn_ra_svn_write_cmd_response(conn, pool, ""));
+ return SVN_NO_ERROR;
}
static svn_error_t *check_path(svn_ra_svn_conn_t *conn, apr_pool_t *pool,
Matt
--
Oink!
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: RA_SVN TEST FAILURE: log_tests.py 3: test 'svn log' on an empty repository
Posted by Karl Fogel <kf...@newton.ch.collab.net>.
David Summers <da...@summersoft.fay.ar.us> writes:
> Is anyone else seeing this? I've been seeing it for the last couple of
> days.
>
> This run is on Rev 5363.
Yes, I'm seeing this too now, only over ra_svn.
I need to finish up issue #630 right now, so I won't get a chance to
look at this until that's done. Hope someone else (hint hint) will be
seized by a desire to have a whack at it before then...
-K
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org