From thewoolleyman at gmail.com  Fri May  1 00:03:32 2009
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Thu, 30 Apr 2009 21:03:32 -0700
Subject: [Cruisecontrolrb-users] Cancelling a build
In-Reply-To: <87zldxnccx.fsf@pond.riseup.net>
References: <0562C1EF-159B-439F-B448-7A7C911B8ADC@mattwynne.net> 
	<e3af97510904011204l3d78c0ffub19b62ad7e73b24a@mail.gmail.com> 
	<04C197D2-78DF-4F52-B812-65742276B748@mattwynne.net>
	<e3af97510904021430g59a6149ag8ea14e8f4a9ee625@mail.gmail.com> 
	<877i13ofgt.fsf@pond.riseup.net>
	<e3af97510904291834v2231bf2dm1704a02d15a6a6ec@mail.gmail.com> 
	<87zldxnccx.fsf@pond.riseup.net>
Message-ID: <e3af97510904302103j102501bbga491bee175843de0@mail.gmail.com>

On Thu, Apr 30, 2009 at 12:57 PM, Micah Anderson <micah at riseup.net> wrote:
> Actually it seems like when I kill the builders, the next time a build
> starts it too goes on for infinity.
>
> How can I track this down? I tried running every rake task that is setup
> for the build for this project, they all return like normal.

Add logging at every step of your build, and try to narrow down what
is hanging.  Once it stops getting to a log step, dig deeper and add
more logging.

Also check your processes.  Learn the 'ps' command, and how to view
process groups and display parent/child process ids.  Find all child
processes of your builder/build (I think the daemon may give you some
hints here, that's what it does).  It may just be "ruby", in which
case you'll still need to keep logging debug statements to find out
what is going on, but you may find something that give you a hint.

Good luck,
-- Chad

From zippitydodah at hotmail.com  Wed May  6 15:47:06 2009
From: zippitydodah at hotmail.com (Mark Knave)
Date: Wed, 6 May 2009 19:47:06 +0000
Subject: [Cruisecontrolrb-users] Notifier Plugin Install fails:
	load_missing_constant
Message-ID: <BAY130-W50E028C063077CAF1BF91DE660@phx.gbl>


Running 1.3 on Ubuntu.  CruiseControl.rb is installed in /home/cruise.

I have a simple notifier (Gnome popup).  As I understand the documentation, I am supposed to copy the plugin class file into /home/cruise/.cruise/builder_plugin/installed/; the builder_plugin directory exists, but installed directory did not.  I created the directory, copied the file and restarted cruise.  I got a load_missing_constant error.  I removed my plugin class file, restarted cruise, and got the same error.

Can you point me to what I'm doing wrong?


Quote from documentation I am referring to:  

"All installed plugins are found in the [cruise]/builder_plugins/installed/ directory.
To install a plugin, all you need to do is move it into this directory and restart the cruise."

Error message

./script/../config/../vendor/rails/activesupport/lib/active_support/dependencies.rb:249:in `load_missing_constant': Expected script/../config/../app/models/project.rb to define Project (LoadError)


Thank you!

_________________________________________________________________
Hotmail? has a new way to see what's up with your friends.
http://windowslive.com/Tutorial/Hotmail/WhatsNew?ocid=TXT_TAGLM_WL_HM_Tutorial_WhatsNew1_052009

From thewoolleyman at gmail.com  Mon May 11 19:48:15 2009
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Mon, 11 May 2009 16:48:15 -0700
Subject: [Cruisecontrolrb-users] Sometimes 'git fetch origin' hangs
	builds
In-Reply-To: <e3af97510902270932j142f3b0ew80ade5edd8445d3b@mail.gmail.com>
References: <e3af97510902270932j142f3b0ew80ade5edd8445d3b@mail.gmail.com>
Message-ID: <e3af97510905111648q24860a9et4f21944b8a607f2b@mail.gmail.com>

On Fri, Feb 27, 2009 at 10:32 AM, Chad Woolley <thewoolleyman at gmail.com> wrote:
> I opened a ticket:
>
> http://cruisecontrolrb.lighthouseapp.com/projects/9150-cruise-control-rb/tickets/229-sometimes-git-hangs

We (Josh Susser and I) just committed a change which hopefully fixes
this error (we were unable to reproduce manually).  Let me know if
anyone still encounters builds hanging in git projects.

Thanks,
-- Chad

From thewoolleyman at gmail.com  Mon May 11 21:09:13 2009
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Mon, 11 May 2009 18:09:13 -0700
Subject: [Cruisecontrolrb-users] Using GIT w/CC.RB
In-Reply-To: <C8C8E80C-AC0A-41F5-859C-86353E8957C6@mattwynne.net>
References: <D8B340B6-99D8-481F-BE64-945DA5B5F8E9@gmail.com> 
	<3945c4270810091344j56fdebabv8e09bd14083c76fa@mail.gmail.com> 
	<251B90A2-5918-4164-8E37-1BFA39B7983F@gmail.com>
	<3945c4270810091505g7ac88ad6ye7386c4517e5a9c@mail.gmail.com> 
	<648eacf20810091512s47f9fc84jcbb32b5abd257b53@mail.gmail.com> 
	<C8C8E80C-AC0A-41F5-859C-86353E8957C6@mattwynne.net>
Message-ID: <e3af97510905111809l13219708j59fb3975ea59245c@mail.gmail.com>

On Fri, Oct 10, 2008 at 1:00 AM, Matt Wynne <matt at mattwynne.net> wrote:
> I guess my only negative feedback is that some support for submodules
> (automatically doing git submodule init / update after pulling) would be
> nice, but we're happy doing that manually as and when required.

Just checked this in.  Try it out in master branch:

https://cruisecontrolrb.lighthouseapp.com/projects/9150/tickets/234-no-git-submodule-support

From matt at mattwynne.net  Tue May 12 04:17:09 2009
From: matt at mattwynne.net (Matt Wynne)
Date: Tue, 12 May 2009 09:17:09 +0100
Subject: [Cruisecontrolrb-users] Using GIT w/CC.RB
In-Reply-To: <e3af97510905111809l13219708j59fb3975ea59245c@mail.gmail.com>
References: <D8B340B6-99D8-481F-BE64-945DA5B5F8E9@gmail.com>
	<3945c4270810091344j56fdebabv8e09bd14083c76fa@mail.gmail.com>
	<251B90A2-5918-4164-8E37-1BFA39B7983F@gmail.com>
	<3945c4270810091505g7ac88ad6ye7386c4517e5a9c@mail.gmail.com>
	<648eacf20810091512s47f9fc84jcbb32b5abd257b53@mail.gmail.com>
	<C8C8E80C-AC0A-41F5-859C-86353E8957C6@mattwynne.net>
	<e3af97510905111809l13219708j59fb3975ea59245c@mail.gmail.com>
Message-ID: <58168F24-57D1-46F6-ADF5-9F20AD05DCE1@mattwynne.net>


On 12 May 2009, at 02:09, Chad Woolley wrote:

> On Fri, Oct 10, 2008 at 1:00 AM, Matt Wynne <matt at mattwynne.net>  
> wrote:
>> I guess my only negative feedback is that some support for submodules
>> (automatically doing git submodule init / update after pulling)  
>> would be
>> nice, but we're happy doing that manually as and when required.
>
> Just checked this in.  Try it out in master branch:
>
> https://cruisecontrolrb.lighthouseapp.com/projects/9150/tickets/234-no-git-submodule-support

Thanks for your efforts Chad, I'm afraid I have defected :)

http://blog.mattwynne.net/2009/04/21/goodbye-cruisecontrolrb-hello-hudson/

Matt Wynne
http://blog.mattwynne.net
http://www.songkick.com


From thewoolleyman at gmail.com  Fri May 22 18:30:55 2009
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Fri, 22 May 2009 15:30:55 -0700
Subject: [Cruisecontrolrb-users] Campfire notifier plugin?
Message-ID: <e3af97510905221530wc228ed6ob6e3ca6af286799@mail.gmail.com>

Anyone have one?

From Bil.Kleb at NASA.gov  Fri May 22 19:06:09 2009
From: Bil.Kleb at NASA.gov (Bil Kleb)
Date: Fri, 22 May 2009 19:06:09 -0400
Subject: [Cruisecontrolrb-users] build trigger chaining (build same revision
	for entire chain)
Message-ID: <4A172FE1.70900@NASA.gov>

Hi,

Does anyone know how to (or have a patch for) chaining a
series of builds so that each build in the chain builds
the triggering build's revision and not the latest revision
available from source control?

We have a series of builds that, if they all pass, the
final one builds a release.

This was my patch to 1.2.1,

*** cruisecontrolrb-1.2.1/app/models/successful_build_trigger.rb       2007-11-04 15:06:49.000000000 -0500
--- cruisecontrolrb-1.2.1-hacked/app/models/successful_build_trigger.rb        2008-04-04 06:58:28.000000000 -0400
***************
*** 16,22 ****
         []
       else
         @last_successful_build = new_last_successful_build
!       [@triggered_project.last_locally_known_revision]
       end
     end

--- 16,22 ----
         []
       else
         @last_successful_build = new_last_successful_build
!       [Revision.new(@last_successful_build.revision)]
       end
     end

but I'm failing to see how to translate this to 1.3.x.

Thanks,
--
Bil Kleb
http://fun3d.larc.nasa.gov

From thewoolleyman at gmail.com  Sat May 23 02:13:11 2009
From: thewoolleyman at gmail.com (Chad Woolley)
Date: Fri, 22 May 2009 23:13:11 -0700
Subject: [Cruisecontrolrb-users] build trigger chaining (build same
	revision for entire chain)
In-Reply-To: <4A172FE1.70900@NASA.gov>
References: <4A172FE1.70900@NASA.gov>
Message-ID: <e3af97510905222313p8a59602xfec9c4780b486999@mail.gmail.com>

On Fri, May 22, 2009 at 4:06 PM, Bil Kleb <Bil.Kleb at nasa.gov> wrote:
> but I'm failing to see how to translate this to 1.3.x.

Check Project#build_if_necessary
It says:  return build(source_control.latest_revision, reasons)

This now hardcodes to build the latest revision.

To accomplish what you want, maybe you could add a #revision_to_build
property on SuccessfulBuildTrigger, and use that instead of
latest_revision.  This would be an API change to the trigger, though,
so you'd probably want to default to the latest revision if there was
no #revision_to_build property or it was nil.

-- Chad

From micah at riseup.net  Sun May 31 10:55:05 2009
From: micah at riseup.net (Micah Anderson)
Date: Sun, 31 May 2009 10:55:05 -0400
Subject: [Cruisecontrolrb-users] Builds failing but not being reported as
	failed
Message-ID: <87y6sduxti.fsf@pond.riseup.net>


I've got a build for a project that clearly has failures in the
build.log, but shows up as a successful build in cruise (and the RSS
feed). I'm trying to determine what is causing this and how to get those
builds to show up as failed. 

I have a build.sh, which performs the different steps for building,
pretty basic stuff, it looks like this:

#!/bin/sh

export RAILS_ENV=test
patch -p1 < ../greencloth.diff

echo "loading default schema"
rake db:schema:load 

echo "running rake db:test:prepare..."
rake db:test:prepare 

echo "running rake db:fixtures:load..."
rake db:fixtures:load

echo "running rake ts:index..."
rake ts:index 

echo "running rake ts:start..."
rake ts:start

echo "running rake test:functionals..."
rake test:functionals

echo "running rake test..."
rake test

echo "running rake test:tools:all..."
rake test:tools:all

echo "running rake test:units:rcov..."
rake test:units:rcov
if [ -d $CC_BUILD_ARTIFACTS ]
then
        echo "moving unit test coverage build artifacts to the proper
        place..."
        mv coverage/units $CC_BUILD_ARTIFACTS/unit_test_coverage
fi

export SHOW_ONLY=controllers
echo "running test:functionals:rcov..."
rake test:functionals:rcov

if [ -d $CC_BUILD_ARTIFACTS ]
then
        echo "moving functional test coverage artifacts to the proper
        place..."
        mv coverage/functionals
        $CC_BUILD_ARTIFACTS/functional_test_coverage
fi

echo "running rake test:integration..."
rake test:integration 

A failed build, that did not get reported as failed, can be seen at
http://labs.riseup.net:3333/builds/crabgrass/eaa4d1

Any help would be appreciated!
micah