You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by "Hyrum K. Wright" <hy...@mail.utexas.edu> on 2007/09/22 14:53:17 UTC

Re: svn commit: r26752 - in trunk/subversion/bindings/javahl: native src/org/tigris/subversion/javahl

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

hwright@tigris.org wrote:
> Author: hwright
> Date: Sat Sep 22 07:48:39 2007
> New Revision: 26752
> 
> Log:
> Extend depth parameter from svn_client_propset3() to JavaHL.  Followup to
> r26745.
> 
> (Note: This adds another JavaHL test failure to the one already present.)

This test failure was present even if I just use the
SVN_DEPTH_TO_RECURSE macro in SVNClient::propertySet() instead of
pushing the depth parameter all the way through to the Java API.  I
decided that getting JavaHL to build was of higher priority than
immediately diagnosing the failing test.

The two currently failing tests are:

testOODStatus(org.tigris.subversion.javahl.BasicTests)junit.framework.AssertionFailedError:
commit item for
file:///home/hwright/dev/svn-apache/subversion/bindings/javahl/test-work/repositories/basic_test4/iota
not expected

and

testMergeInfoRetrieval(org.tigris.subversion.javahl.BasicTests)junit.framework.AssertionFailedError:
Missing merge info for source '/A/mu' on
'/home/hwright/dev/svn-apache/subversion/bindings/javahl/test-work/working_copies/basic_test32/branches/A/mu'

The second one has been there for at least the past couple of days.  If
someone familiar with JavaHL and the tests would take a look at it, I'd
appreciate it.

- -Hyrum
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFG9SxdCwOubk4kUXwRAqJUAKCUz3HvzWZPrS7XnJ6l4xSR7sOTpgCfbequ
y+piNqhwlFEzSBWareyEcG8=
=8hf+
-----END PGP SIGNATURE-----

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: Failing testMergeInfoRetrieval test (was Re: svn commit: r26752 ...)

Posted by Daniel Rall <dl...@collab.net>.
On Mon, 24 Sep 2007, Daniel L. Rall wrote:

> On Sat, 22 Sep 2007, Hyrum K. Wright wrote:
> ...
> > testMergeInfoRetrieval(org.tigris.subversion.javahl.BasicTests)junit.framework.AssertionFailedError:
> > Missing merge info for source '/A/mu' on
> > '/home/hwright/dev/svn-apache/subversion/bindings/javahl/test-work/working_copies/basic_test32/branches/A/mu'
> > 
> > The second one has been there for at least the past couple of days.  If
> > someone familiar with JavaHL and the tests would take a look at it, I'd
> > appreciate it.
> 
> I'm taking a look at this failure now...

I subsequently commited the fix for this failure (injected in r26666)
in r26774.

Failing testMergeInfoRetrieval test (was Re: svn commit: r26752 ...)

Posted by "Daniel L. Rall" <dl...@finemaltcoding.com>.
On Sat, 22 Sep 2007, Hyrum K. Wright wrote:
...
> testMergeInfoRetrieval(org.tigris.subversion.javahl.BasicTests)junit.framework.AssertionFailedError:
> Missing merge info for source '/A/mu' on
> '/home/hwright/dev/svn-apache/subversion/bindings/javahl/test-work/working_copies/basic_test32/branches/A/mu'
> 
> The second one has been there for at least the past couple of days.  If
> someone familiar with JavaHL and the tests would take a look at it, I'd
> appreciate it.

I'm taking a look at this failure now...
-- 

Daniel Rall

[sparse directories] Non-recursive mode and 'info', 'propset', etc. APIs (was Re: svn commit: r26752 ...)

Posted by "Daniel L. Rall" <dl...@collab.net>.
On Sat, 22 Sep 2007, Hyrum K. Wright wrote:

> hwright@tigris.org wrote:
...
> > Log:
> > Extend depth parameter from svn_client_propset3() to JavaHL.  Followup to
> > r26745.
> > 
> > (Note: This adds another JavaHL test failure to the one already present.)
> 
> This test failure was present even if I just use the
> SVN_DEPTH_TO_RECURSE macro in SVNClient::propertySet() instead of
> pushing the depth parameter all the way through to the Java API.  I
> decided that getting JavaHL to build was of higher priority than
> immediately diagnosing the failing test.
... 
> testOODStatus(org.tigris.subversion.javahl.BasicTests)junit.framework.AssertionFailedError:
> commit item for
> file:///home/hwright/dev/svn-apache/subversion/bindings/javahl/test-work/repositories/basic_test4/iota
> not expected
 
After discussing this with Hyrum, I fixed this problem today in r26766.

This was caused by use of Depth.fromRecurse(), Java code with the same
behavior as that of the SVN_DEPTH_FROM_RECURSE() macro.  Both the 'propset'
and 'info' APIs were making use of this type of conversion, which maps
non-recursive modes to svn_depth_files, breaking backwards compatibility.
APIs of this type should instead map non-recursive to svn_depth_empty.

Are there more cases of this problem?