headers
Dustin J. Mitchell | 1 Jan 2010 02:54
Favicon
Gravatar

First release candidate for 0.7.12

Happy New Year, everyone.  We have a candidate for a new release!
This isn't on pypi, because it's probably not ready for prime-time
just yet, but is available for download.

Signed tag:

  http://github.com/djmitche/buildbot/commit/08fc7d873d5f5961ae30f7cc632e8ca4c1dd06b8

Tarball/Zipball:

  http://github.com/djmitche/buildbot/tarball/buildbot-0.7.12rc1
  http://github.com/djmitche/buildbot/zipball/buildbot-0.7.12rc1

Please download and test it out.  I'd like to hear about:
 * bugs (file a bug in the 0.7.12 milestone)
 * changes omitted from NEWS - I skimmed the ChangeLog, but it *is*
New Years' Eve, and I probably missed a lot..
 * test failures (I disabled the wonky lock tests, so hopefully
nothing will fail)

I won't do another release until I've heard back from a number of
people, and fixed anything that was broken.

Thanks!

Dustin

--

-- 
Open Source Storage Engineer
http://www.zmanda.com
(Continue reading)

Permalink | Reply | 9 comments |
headers
Dustin J. Mitchell | 1 Jan 2010 02:55
Favicon
Gravatar

'master' branch now open for 0.8.0 merges

The remaining work for 0.7.12 will be done on the "release" branch,
and merged back to master as necessary.

So have at it, folks!  Get your mildly-backward-incompatible changes
in now, before it's too late!

Dustin

--

-- 
Open Source Storage Engineer
http://www.zmanda.com

------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev
Permalink | Reply | 3 comments |
headers
Charles Lepple | 1 Jan 2010 03:44
Picon

Re: First release candidate for 0.7.12

On Thu, Dec 31, 2009 at 8:54 PM, Dustin J. Mitchell <dustin@...> wrote:
>  * test failures (I disabled the wonky lock tests, so hopefully
> nothing will fail)

You asked for it... :-)

I'm seeing two issues on Mac OS X, using Twisted 8.2.0 and Python 2.6.

What bugs me is that I'm seeing "Buildbot needs to be installed" in
test_limitlogs.py (the first failure). Do we have a test procedure
that takes this into account? Is there a way to do this from the
install directory that makes sure we are testing the new code, not
what is installed?

Here's the tail end of the log. Let me know if you need more details.

===============================================================================
[FAIL]: buildbot.test.test_limitlogs.MasterLogs.testLog

Traceback (most recent call last):
  File "/sw.unstable/src/fink.build/buildbot-py26-0.7.12-0/djmitche-buildbot-1320e1c/buildbot/test/test_limitlogs.py",
line 91, in _do_tests
    self.failUnlessEqual(lf.listLogs(), [1,2])
twisted.trial.unittest.FailTest: not equal:
a = [1]
b = [1, 2]

===============================================================================
[ERROR]: buildbot.test.test_mergerequests.MergeRequestsTest.testNoMerges
(Continue reading)

Permalink | Reply | 6 comments |
headers
Dustin J. Mitchell | 1 Jan 2010 19:32
Favicon
Gravatar

Re: First release candidate for 0.7.12

On Thu, Dec 31, 2009 at 8:44 PM, Charles Lepple <clepple@...> wrote:
> What bugs me is that I'm seeing "Buildbot needs to be installed" in
> test_limitlogs.py (the first failure). Do we have a test procedure
> that takes this into account? Is there a way to do this from the
> install directory that makes sure we are testing the new code, not
> what is installed?

Thanks!

I don't know if that test could be configured to run buildbot from the
build directory, but as written you're right to note that it could
accidentally test an old buildbot.  The test is really checking how
well the new format of buildbot.tac works with twistd, and that the
'buildbot' executable creates that file correctly.

If we assume that tests are run in a virtualenv, with 'python setup.py
develop' in effect, then the test is fine.  But I don't see another
way to run this test effectively.

Dustin

--

-- 
Open Source Storage Engineer
http://www.zmanda.com

------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
(Continue reading)

Permalink | Reply | 5 comments |
headers
Charles Lepple | 2 Jan 2010 02:04
Picon

Re: First release candidate for 0.7.12

On Fri, Jan 1, 2010 at 1:32 PM, Dustin J. Mitchell <dustin@...> wrote:
> On Thu, Dec 31, 2009 at 8:44 PM, Charles Lepple <clepple@...> wrote:
>> What bugs me is that I'm seeing "Buildbot needs to be installed" in
>> test_limitlogs.py (the first failure). Do we have a test procedure
>> that takes this into account? Is there a way to do this from the
>> install directory that makes sure we are testing the new code, not
>> what is installed?
>
> Thanks!
>
> I don't know if that test could be configured to run buildbot from the
> build directory, but as written you're right to note that it could
> accidentally test an old buildbot.  The test is really checking how
> well the new format of buildbot.tac works with twistd, and that the
> 'buildbot' executable creates that file correctly.
>
> If we assume that tests are run in a virtualenv, with 'python setup.py
> develop' in effect, then the test is fine.  But I don't see another
> way to run this test effectively.

Hmm, I will have to look into virtualenv again.

Thanks,

--

-- 
- Charles Lepple

------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
(Continue reading)

Permalink | Reply | 0 comments |
headers
Nicolas Alvarez | 2 Jan 2010 04:39
Picon
Gravatar

Re: problem about test buildbot on ubuntu-8.04

Guo Hongruan wrote:
> Hi guys,
>    I tried to test buildbot on ubuntu-8.04 and found the following
>    warning.
> Does it mean the installation of buildbot on my machine failed.
> 
> By the way, do I have to run the test using root account?
> 
> camelguo <at> guohr-laptop-ubuntu:~$ PYTHONPATH=. trial buildbot.test
> Traceback (most recent call last):
>    File
> "/usr/lib/python2.5/site-packages/Twisted-9.0.0-py2.5-linux-
i686.egg/twisted/python/usage.py",
> line 241, in parseOptions
>      self.postOptions()
>    File
> "/usr/lib/python2.5/site-packages/Twisted-9.0.0-py2.5-linux-
i686.egg/twisted/scripts/trial.py",
> line 295, in postOptions
>      self['reporter'] = self._loadReporterByName(self['reporter'])
>    File
> "/usr/lib/python2.5/site-packages/Twisted-9.0.0-py2.5-linux-
i686.egg/twisted/scripts/trial.py",
> line 281, in _loadReporterByName
>      for p in plugin.getPlugins(itrial.IReporter):
>    File
> "/usr/lib/python2.5/site-packages/Twisted-9.0.0-py2.5-linux-
i686.egg/twisted/plugin.py",
> line 200, in getPlugins
>      allDropins = getCache(package)
(Continue reading)

Permalink | Reply | 1 comment |
headers
Guo Hongruan | 2 Jan 2010 04:44
Picon
Gravatar

Re: problem about test buildbot on ubuntu-8.04

When running as root, there are no such errors. But the test lock3 failed.

在 Sat, 02 Jan 2010 11:39:44 +0800，Nicolas Alvarez  
<nicolas.alvarez <at> gmail.com> 写道:

> Guo Hongruan wrote:
>> Hi guys,
>>    I tried to test buildbot on ubuntu-8.04 and found the following
>>    warning.
>> Does it mean the installation of buildbot on my machine failed.
>>
>> By the way, do I have to run the test using root account?
>>
>> camelguo <at> guohr-laptop-ubuntu:~$ PYTHONPATH=. trial buildbot.test
>> Traceback (most recent call last):
>>    File
>> "/usr/lib/python2.5/site-packages/Twisted-9.0.0-py2.5-linux-
> i686.egg/twisted/python/usage.py",
>> line 241, in parseOptions
>>      self.postOptions()
>>    File
>> "/usr/lib/python2.5/site-packages/Twisted-9.0.0-py2.5-linux-
> i686.egg/twisted/scripts/trial.py",
>> line 295, in postOptions
>>      self['reporter'] = self._loadReporterByName(self['reporter'])
>>    File
>> "/usr/lib/python2.5/site-packages/Twisted-9.0.0-py2.5-linux-
> i686.egg/twisted/scripts/trial.py",
>> line 281, in _loadReporterByName
>>      for p in plugin.getPlugins(itrial.IReporter):
(Continue reading)

Permalink | Reply | 0 comments |
headers
Marcus Lindblom | 2 Jan 2010 17:34
Picon
Gravatar

Re: First release candidate for 0.7.12

On 2010-01-01 19:32, Dustin J. Mitchell wrote:
> On Thu, Dec 31, 2009 at 8:44 PM, Charles Lepple<clepple@...>  wrote:
>> What bugs me is that I'm seeing "Buildbot needs to be installed" in
>> test_limitlogs.py (the first failure). Do we have a test procedure
>> that takes this into account? Is there a way to do this from the
>> install directory that makes sure we are testing the new code, not
>> what is installed?
>
> Thanks!
>
> I don't know if that test could be configured to run buildbot from the
> build directory, but as written you're right to note that it could
> accidentally test an old buildbot.  The test is really checking how
> well the new format of buildbot.tac works with twistd, and that the
> 'buildbot' executable creates that file correctly.
>
> If we assume that tests are run in a virtualenv, with 'python setup.py
> develop' in effect, then the test is fine.  But I don't see another
> way to run this test effectively.

I meddled with that last week (on the jinja branch, regrettably) when I 
was banging away at making more tests work properly (esp. on windows):

http://github.com/marcusl/buildbot/commit/7ef927159109651e4ff3555f5be1bfeb993619e4

Since 'buildbot' isn't directly runnable on windows (which doesn't find 
it) I changed it to lookup 'python' instead, and add the tested buildbot 
to PYTHONPATH.

I dunno if it helps the above case, but it made the test pass for me.
(Continue reading)

Permalink | Reply | 3 comments |
headers
Marcus Lindblom | 2 Jan 2010 18:01
Picon
Gravatar

Re: 'master' branch now open for 0.8.0 merges

On 2010-01-01 02:55, Dustin J. Mitchell wrote:
> The remaining work for 0.7.12 will be done on the "release" branch,
> and merged back to master as necessary.
>
> So have at it, folks!  Get your mildly-backward-incompatible changes
> in now, before it's too late!

So, time to merge jinja-branch into master? :)

Cheers,
/Marcus

------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev
Permalink | Reply | 2 comments |
headers
Martin v. Löwis | 2 Jan 2010 19:51
Picon
Gravatar

Slave not recognizing process end

We have a slave where the test command is configured with
a timeout of 1800s. The command completed before
that, yet the slave still believed it had to kill it.

So we see the following pieces in the log:

2010-01-01 15:31:43-0500 [-] command timed out: 1800 seconds without output
2010-01-01 15:31:43-0500 [-] self.process has no pid
2010-01-01 15:31:43-0500 [-] trying process.signalProcess('KILL')
2010-01-01 15:31:43-0500 [-] Unhandled Error
[...]
"/usr/lib/python2.6/site-packages/twisted/internet/process.py", line
312, in signalProcess
	    if os.WIFEXITED(status):
	twisted.internet.error.ProcessExitedAlready:

As a consequence, the slave now still believes that the process
is running, and any attempts to cancel the process from the master
all lead to the same exception (ProcessExitedAlready).

Restarting the slave would work around the problem, however, since this
happens often, I would like to fix it for good.

Where should I look?

Regards,
Martin

------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
(Continue reading)

Permalink | Reply | 12 comments |

Gmane