successive calls to Capistrano::CLI

[ben.koski]

One more headscratcher for tonight:

I'm trying to embed Capistrano into a rails app. I'm using cap 2.4
and capistrano-ext to handle stages. The Capistrano::CLI docs make it
seem simple enough to call

Capistrano::CLI.parse(['staging', "deploy:db:update"]).execute!

or

config = Capistrano::Configuration.new
config.logger_level = Capistrano::Logger::TRACE
config.load "Capfile"
config.load "config/deploy/staging"
config.set(:stage) {'staging'}
config.deploy.db.update

both of these work _beautifully_ the first time. Problem is, if I try
to execute this twice it blows up on "ArgumentError: defining a task
named `symlink' would shadow an existing method with that name" in
capistrano/configuration/namespaces.rb.

Makes sense to raise when trying to redefine something, but the
strange thing is that these definitions seem to persist even after
config or event the class making the call to CLI has gone out of
scope--kind of. If I try to run config.deploy.db.update a second time
without running the config.load, the tasks aren't there.

Any suggestions?

[Jamis Buck]

Could you post your entire script?

- Jamis

[ben.koski]

Got the Capistrano::CLI.parse().execute! command working in a process
forked from the rails app--which I needed to do anyway--but it's still
strange that it blows up when invoked twice from the same process.
The capfiles are reasonable staightforward--though I am using the
ec2onrails gem.

If anyone sees something obvious, let me know!

Here's the code to invoke the cap:
----------------
require 'capistrano/cli'

class DeployManager
def self.deploy_data

Capistrano::CLI.parse(['staging',
'deploy:db:update']).execute!

end
end
------------------


Capfile:
----------------
load 'deploy' if respond_to?(:namespace) # cap2 differentiator

set :stages, %w(staging production)
require 'capistrano/ext/multistage'
---------------

config/deploy/staging.rb:
---------------
set :application, "myapp"

set :repository, "svn://myrepo/#{application}/trunk"
set :deploy_via, :copy
set :copy_stategy, :export

# NOTE: for some reason Capistrano requires you to have both the
public and
# the private key in the same folder, the public key should have the
# extension ".pub".
ssh_options[:keys] = ["#{ENV['HOME']}/.ssh/mykey.pem"]

role :web, "myserver"
role :app, "myserver"
role :db, "myserver", :primary => true

set :rails_env, "staging"

# EC2 on Rails config. Many of these can be omitted if not needed,
check
# the documentation.
set :ec2onrails_config, {
# S3 bucket and "subdir" used by the ec2onrails:db:restore task
:restore_from_bucket => "foo",
:restore_from_bucket_subdir => "bar",

# Set a root password for MySQL. Run "cap
ec2onrails:db:set_root_password"
# to enable this. This is optional, and after doing this the
# ec2onrails:db:drop task won't work, but be aware that MySQL
accepts
# connections on the public network interface (you should block the
MySQL
# port with the firewall anyway).
:mysql_root_password => "foobar",

# Any extra Ubuntu packages to install if desired
# :packages => ["logwatch", "imagemagick"],

# Any extra RubyGems to install if desired: can be "gemname" or if
a
# particular version is desired "gemname -v 1.0.1"
# :rubygems => ["rmagick", "rfacebook -v 0.9.7"],

# Set the server timezone. run "cap -e
ec2onrails:server:set_timezone" for
# details
:timezone => "US/Eastern",

# Files to deploy to the server (they'll be owned by root). It's
intended
# mainly for customized config files for new packages installed via
the
# ec2onrails:server:install_packages task. Subdirectories and files
inside
# here will be placed in the same structure relative to the root of
the
# server's filesystem.
#:server_config_files_root => "../server_config",

# If config files are deployed, some services might need to be
restarted
:services_to_restart => %w(apache2 postfix sysklogd),

# Set an email address to forward admin mail messages to
:admin_mail_forward_address => "te...@example.com",
}

require 'ec2onrails/recipes'

namespace :deploy do
namespace :db do

# main update task -- run this to upload and then swap databases
task :update, :roles => :db do
# exec my task
end
end
end
--------------

[Jamis Buck]

Are you seeing this error when your rails app is running in
development mode, or production mode? If dev mode, I suspect it's
rails' autoload magic that is causing problems. I just threw together
a simple little test script to try and duplicate the problem, and it
works just fine:

~ $ cat y.rb
require 'capistrano/cli'

Capistrano::CLI.parse(%w(-f capfile.test symlink)).execute!
Capistrano::CLI.parse(%w(-f capfile.test symlink)).execute!

No warnings printed or errors raised.

- Jamis