You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@subversion.apache.org by Peter Samuelson <pe...@p12n.org> on 2007/12/13 20:54:40 UTC

ruby failure (was: Subversion 1.4.6 tarballs up for testing/signing)

[Hyrum K. Wright]
> After completely screwing up the dependencies yesterday, I've
> rerolled the 1.4.6 tarball with (hopefully) the correct dependencies.

I get two failures in the ruby bindings, using ruby 1.8.6.111.
However, I did rebuild the swig stuff with swig 1.3.31; I'll see if I
can reproduce it with the swig 1.3.25 stuff in the tarball, and so on.
-- 
Peter Samuelson | org-tld!p12n!peter | http://p12n.org/

$ make check-swig-rb
cd {SRC}/BUILD/subversion/bindings/swig/ruby; \
	  /usr/bin/ruby1.8 -I {SRC}/subversion/bindings/swig/ruby \
	    {SRC}/subversion/bindings/swig/ruby/test/run-test.rb \
	    --verbose=normal
Loaded suite .
Started
................FF....................................................................................................................
Finished in 352.43096 seconds.

  1) Failure:
test_diff_summarize(SvnClientTest) [{SRC}/subversion/bindings/swig/ruby/test/test_client.rb:763]:
<["hello.txt"]> expected but was
<[""]>.

  2) Failure:
test_diff_summarize_peg(SvnClientTest) [{SRC}/subversion/bindings/swig/ruby/test/test_client.rb:815]:
<["after.txt"]> expected but was
<[""]>.

134 tests, 4554 assertions, 2 failures, 0 errors

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

Re: ruby failure

Posted by Joe Swatosh <jo...@gmail.com>.

Hi All,

On Dec 13, 2007 6:04 PM, Hyrum K. Wright <hy...@mail.utexas.edu> wrote:
> Peter Samuelson wrote:
> > [Peter Samuelson]
> >> I get two failures in the ruby bindings, using ruby 1.8.6.111.
> >
> > Fixed in trunk r26071; please backport this.  I don't know if it's
> > worth rerolling the 1.4.6 tarball for.
>
> Hmm, I don't know either.
>
> Joe, how do you feel about this?  (And, if needed, could you nominate it
> for backport?)
>

I'm not sure.  I think if its important kou should nominate it.  I can't run the
ruby bindings from 1.4.x so I'd have no way of validating the backport.

--
Joe

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

Re: ruby failure

Posted by "Hyrum K. Wright" <hy...@mail.utexas.edu>.

Peter Samuelson wrote:
> [Peter Samuelson]
>> I get two failures in the ruby bindings, using ruby 1.8.6.111.
> 
> Fixed in trunk r26071; please backport this.  I don't know if it's
> worth rerolling the 1.4.6 tarball for.

Hmm, I don't know either.

Joe, how do you feel about this?  (And, if needed, could you nominate it
for backport?)

-Hyrum

Re: ruby failure (was: Subversion 1.4.6 tarballs up for testing/signing)

Posted by David Glasser <gl...@davidglasser.net>.

On Dec 14, 2007 11:36 AM, David Glasser <gl...@davidglasser.net> wrote:
> On Dec 14, 2007 10:13 AM, David Glasser <gl...@davidglasser.net> wrote:
> > On Dec 14, 2007 10:11 AM, David Glasser <gl...@davidglasser.net> wrote:
> > > On Dec 13, 2007 5:23 PM, Peter Samuelson <pe...@p12n.org> wrote:
> > > >
> > > > [Peter Samuelson]
> > > > > I get two failures in the ruby bindings, using ruby 1.8.6.111.
> > > >
> > > > Fixed in trunk r26071; please backport this.  I don't know if it's
> > > > worth rerolling the 1.4.6 tarball for.
> > >
> > > Yes, we need to re-roll; we should not release with failing tests.
> >
> > I'm not really qualified to say whether or not r26071 should be backported.
> >
> > Are we sure that the API change in svn_client_diff_summarize2 that
> > they're describing isn't a bug?
>
> OK.  So r25654 made an incompatible API change, but it was a bugfix.
>
> r26071 adjusted the ruby tests to deal with this change (note, though,
> that Joe could have also changed the expectation to expect [""]
> instead of [file]).
>
> So I think backporting r26071 is fine.  I haven't actually tested it
> myself, because I don't usually build the Ruby bindings.  I'll test it
> now and then nominate it if that works.

Nominated.  Just needs a +0 from any (partial or full) committer.

--dave

-- 
David Glasser | glasser@davidglasser.net | http://www.davidglasser.net/

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

Re: ruby failure (was: Subversion 1.4.6 tarballs up for testing/signing)

Posted by David Glasser <gl...@davidglasser.net>.

On Dec 14, 2007 10:13 AM, David Glasser <gl...@davidglasser.net> wrote:
> On Dec 14, 2007 10:11 AM, David Glasser <gl...@davidglasser.net> wrote:
> > On Dec 13, 2007 5:23 PM, Peter Samuelson <pe...@p12n.org> wrote:
> > >
> > > [Peter Samuelson]
> > > > I get two failures in the ruby bindings, using ruby 1.8.6.111.
> > >
> > > Fixed in trunk r26071; please backport this.  I don't know if it's
> > > worth rerolling the 1.4.6 tarball for.
> >
> > Yes, we need to re-roll; we should not release with failing tests.
>
> I'm not really qualified to say whether or not r26071 should be backported.
>
> Are we sure that the API change in svn_client_diff_summarize2 that
> they're describing isn't a bug?

OK.  So r25654 made an incompatible API change, but it was a bugfix.

r26071 adjusted the ruby tests to deal with this change (note, though,
that Joe could have also changed the expectation to expect [""]
instead of [file]).

So I think backporting r26071 is fine.  I haven't actually tested it
myself, because I don't usually build the Ruby bindings.  I'll test it
now and then nominate it if that works.

--dave


-- 
David Glasser | glasser@davidglasser.net | http://www.davidglasser.net/

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

Re: ruby failure (was: Subversion 1.4.6 tarballs up for testing/signing)

Posted by David Glasser <gl...@davidglasser.net>.

On Dec 14, 2007 10:11 AM, David Glasser <gl...@davidglasser.net> wrote:
> On Dec 13, 2007 5:23 PM, Peter Samuelson <pe...@p12n.org> wrote:
> >
> > [Peter Samuelson]
> > > I get two failures in the ruby bindings, using ruby 1.8.6.111.
> >
> > Fixed in trunk r26071; please backport this.  I don't know if it's
> > worth rerolling the 1.4.6 tarball for.
>
> Yes, we need to re-roll; we should not release with failing tests.

I'm not really qualified to say whether or not r26071 should be backported.

Are we sure that the API change in svn_client_diff_summarize2 that
they're describing isn't a bug?

--dave

-- 
David Glasser | glasser@davidglasser.net | http://www.davidglasser.net/

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

Re: ruby failure (was: Subversion 1.4.6 tarballs up for testing/signing)

Posted by David Glasser <gl...@davidglasser.net>.

On Dec 13, 2007 5:23 PM, Peter Samuelson <pe...@p12n.org> wrote:
>
> [Peter Samuelson]
> > I get two failures in the ruby bindings, using ruby 1.8.6.111.
>
> Fixed in trunk r26071; please backport this.  I don't know if it's
> worth rerolling the 1.4.6 tarball for.

Yes, we need to re-roll; we should not release with failing tests.

--dave


-- 
David Glasser | glasser@davidglasser.net | http://www.davidglasser.net/

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

Re: ruby failure (was: Subversion 1.4.6 tarballs up for testing/signing)

Posted by Peter Samuelson <pe...@p12n.org>.

[Peter Samuelson]
> I get two failures in the ruby bindings, using ruby 1.8.6.111.

Fixed in trunk r26071; please backport this.  I don't know if it's
worth rerolling the 1.4.6 tarball for.
-- 
Peter Samuelson | org-tld!p12n!peter | http://p12n.org/

Re: ruby failure (was: Subversion 1.4.6 tarballs up for testing/signing)

Posted by Peter Samuelson <pe...@p12n.org>.

[Peter Samuelson]
> I get two failures in the ruby bindings, using ruby 1.8.6.111.
> However, I did rebuild the swig stuff with swig 1.3.31

I've reproduced the failure using the swig stuff shipped in the
tarball, and also confirmed that this is a regression: 1.4.4 passes in
the same environment.

I've put my 1.4.6 build logs on http://p12n.org/tmp/svn146log/.
Build environment is Debian unstable:

  gcc 4.2.2
  ruby 1.8.6 patch 111
  apr 1.2.11
  apr-util 1.2.7

Not configured with apache, neon, etc.
-- 
Peter Samuelson | org-tld!p12n!peter | http://p12n.org/