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 James <ja...@gmail.com> on 2005/06/18 16:23:01 UTC
Ruby test failures in r15098
Jani's Test Box wrote:
> cd /srv/svntest/svntest/obj-sh/subversion/bindings/swig/ruby; \
> /usr/bin/ruby -I /srv/svntest/svntest/svn/svn_trunk/subversion/bindings/swig/ruby \
> /srv/svntest/svntest/svn/svn_trunk/subversion/bindings/swig/ruby/test/run-test.rb
> Loaded suite test
> Started
> /svntest/obj-sh/subversion/bindings/swig/ruby/.ext/svn/ext/client.so: [BUG] Segmentation fault
> ruby 1.8.3 (2005-05-12) [x86_64-linux]
>
> make: *** [check-swig-rb] Error 134
> Complete log saved in /srv/svntest/svntest/logs/svn_trunk/LOG_swig-rb_check.failed
> FAIL: swig-rb::check
FYI, I get this same error whenever the test suite tries to use a port
that is already in use by my firewall. Now that Kouhei has switched
the port in r15077, the test suite works fine on my system. I don't
know what's causing Jani's problem here.
For more debugging information, see
http://subversion.tigris.org/servlets/ReadMsg?list=svn-breakage&msgNo=30609
Jani's Test Box wrote:
> /srv/svntest/svntest/svn/svn_trunk/ac-helpers/install-sh -c -d
> /usr/lib/ruby/site_ruby/1.8/x86_64-linux/svn/ext
> mkdir: cannot create directory `/usr/lib/ruby/site_ruby/1.8/x86_64-linux/svn': Permission denied
> mkdir: cannot create directory `/usr/lib/ruby/site_ruby/1.8/x86_64-linux/svn/ext': No such > file or directory
This problem is caused by the fact that the test suite does not have
access to the Ruby installation dir, so we cannot test whether the
installation process works. Jani, can you install Ruby in a local dir
where the test suite will have access?
--
David James -- http://www.cs.toronto.edu/~james
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Ruby test failures in r15098
Posted by David James <ja...@gmail.com>.
On 6/20/05, Jani Averbach <ja...@jaa.iki.fi> wrote:
> If you have time, could you compile your apr and subversion with
> "--enable-pool-debug=yes", and see if the Ruby test suite still passes?
Works for me.
Cheers,
David
--
David James -- http://www.cs.toronto.edu/~james
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Ruby test failures in r15098
Posted by Jani Averbach <ja...@jaa.iki.fi>.
On 2005-06-18 12:23-0400, David James wrote:
Hi David,
> FYI, I get this same error whenever the test suite tries to use a port
> that is already in use by my firewall.
The ports should be free for use.
> Now that Kouhei has switched the port in r15077, the test suite
> works fine on my system. I don't know what's causing Jani's problem
> here.
If you have time, could you compile your apr and subversion with
"--enable-pool-debug=yes", and see if the Ruby test suite still passes?
BR, Jani
--
Jani Averbach
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Ruby test failures in r15098
Posted by David James <ja...@gmail.com>.
On 6/18/05, Jani Averbach <ja...@jaa.iki.fi> wrote:
> Yep. I am running test suite without installing it first, isn't there
> a bug in ruby's make system, when it tries to install something even
> when it has been called without "install"?
>
> The case is similar with perl unit tests which have to be called with only
> following "swig-pl" "" "check-swig-pl" targets.
>
> I am running the ruby test suite with following targets at the moment:
> "swig-rb" "" "check-swig-rb".
Ah, ok -- it must be a bug in the Ruby bindings then.
> > Jani, can you install Ruby in a local dir where the test suite will
> > have access?
>
> I am sorry, but I won't do that. The system is running under chroot
> jail, so in theory I could do what ever with it, but to get reliable
> test results I think it is better if things get installed under
> recyclable locations which could be erased before a new test round.
Good point!
Cheers,
David
--
David James -- http://www.cs.toronto.edu/~james
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Ruby test failures in r15098
Posted by Kouhei Sutou <ko...@cozmixng.org>.
Hi,
In <20...@jaa.iki.fi>
"Re: Ruby test failures in r15098" on Sat, 18 Jun 2005 16:14:44 -0600,
Jani Averbach <ja...@jaa.iki.fi> wrote:
> > Jani's Test Box wrote:
> > > /srv/svntest/svntest/svn/svn_trunk/ac-helpers/install-sh -c -d
> > > /usr/lib/ruby/site_ruby/1.8/x86_64-linux/svn/ext
> > > mkdir: cannot create directory `/usr/lib/ruby/site_ruby/1.8/x86_64-linux/svn': Permission denied
> > > mkdir: cannot create directory `/usr/lib/ruby/site_ruby/1.8/x86_64-linux/svn/ext': No such > file or directory
> > This problem is caused by the fact that the test suite does not have
> > access to the Ruby installation dir, so we cannot test whether the
> > installation process works.
>
> Yep. I am running test suite without installing it first, isn't there
> a bug in ruby's make system, when it tries to install something even
> when it has been called without "install"?
>
> The case is similar with perl unit tests which have to be called with only
> following "swig-pl" "" "check-swig-pl" targets.
>
> I am running the ruby test suite with following targets at the moment:
> "swig-rb" "" "check-swig-rb".
Does "--with-swig-ruby-sitedir" option of configure script
help you?
--
kou
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: Ruby test failures in r15098
Posted by Jani Averbach <ja...@jaa.iki.fi>.
On 2005-06-18 12:23-0400, David James wrote:
> Jani's Test Box wrote:
...
> FYI, I get this same error whenever the test suite tries to use a port
> that is already in use by my firewall. Now that Kouhei has switched
> the port in r15077, the test suite works fine on my system. I don't
> know what's causing Jani's problem here.
I will give it a look at some point.
>
> For more debugging information, see
> http://subversion.tigris.org/servlets/ReadMsg?list=svn-breakage&msgNo=30609
>
> Jani's Test Box wrote:
> > /srv/svntest/svntest/svn/svn_trunk/ac-helpers/install-sh -c -d
> > /usr/lib/ruby/site_ruby/1.8/x86_64-linux/svn/ext
> > mkdir: cannot create directory `/usr/lib/ruby/site_ruby/1.8/x86_64-linux/svn': Permission denied
> > mkdir: cannot create directory `/usr/lib/ruby/site_ruby/1.8/x86_64-linux/svn/ext': No such > file or directory
> This problem is caused by the fact that the test suite does not have
> access to the Ruby installation dir, so we cannot test whether the
> installation process works.
Yep. I am running test suite without installing it first, isn't there
a bug in ruby's make system, when it tries to install something even
when it has been called without "install"?
The case is similar with perl unit tests which have to be called with only
following "swig-pl" "" "check-swig-pl" targets.
I am running the ruby test suite with following targets at the moment:
"swig-rb" "" "check-swig-rb".
> Jani, can you install Ruby in a local dir where the test suite will
> have access?
I am sorry, but I won't do that. The system is running under chroot
jail, so in theory I could do what ever with it, but to get reliable
test results I think it is better if things get installed under
recyclable locations which could be erased before a new test round.
BR, Jani
--
Jani Averbach
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org