Included Modules

Class/Module Index [+]



This class was once named ‘Engine’, but since ruote 2.x and its introduction of workers, the methods here are those of a “dashboard”. The real engine being the set of workers.

The methods here allow to launch processes and to query about their status. There are also methods for fixing issues with stalled processes or processes stuck in errors.

NOTE : the methods launch and reply are implemented in Ruote::ReceiverMixin (this Engine class has all the methods of a Receiver).





Public Class Methods

new(worker_or_storage, opts=true) click to toggle source

Creates an engine using either worker or storage.

If a storage instance is given as the first argument, the engine will be able to manage processes (for example, launch and cancel workflows) but will not actually run any workflows.

If a worker instance is given as the first argument and the second argument is true, engine will start the worker and will be able to both manage and run workflows.

If the second options is set to { :join => true }, the worker will be started and run in the current thread (and the initialize method will not return).

# File lib/ruote/dashboard.rb, line 67
def initialize(worker_or_storage, opts=true)

  @context = worker_or_storage.context
  @context.dashboard = self

  @variables =

  workers = { |ser|
    ser.respond_to?(:run) && ser.respond_to?(:run_in_thread)

  return unless opts && workers.any?

  # let's isolate a worker to join

  worker = if opts.is_a?(Hash) && opts[:join]
    workers.find { |wor| == 'worker' } || workers.first

  (workers - Array(worker)).each { |wor| wor.run_in_thread }
    # launch their thread, but let's not join them if worker
    # and let's not return

Public Instance Methods

add_service(name, path_or_instance, classname=nil, opts=nil) click to toggle source

Adds a service locally (will not get propagated to other workers).

tracer =
@dashboard.add_service('tracer', tracer)


  'tracer', 'ruote/exp/tracer', 'Ruote::Exp::Tracer')

This method returns the service instance it just bound.

# File lib/ruote/dashboard.rb, line 987
def add_service(name, path_or_instance, classname=nil, opts=nil)

  @context.add_service(name, path_or_instance, classname, opts)
add_tracker(wfid, action, tracker_id, conditions, msg) click to toggle source

/!\ warning: advanced method.

Adds a tracker to the ruote engine.


  • wfid: When nil will track any workflow execution, when set will only react on msgs for the given wfid.

  • action: A string like “apply”, “reply” or “receive”, the action being tracked May begin with a “pre_” prefix.

  • tracker_id: When nil, ruote chooses a tracker_id, else its the unique identifier for the new tracker.

  • conditions: A Hash of keys pointing to arrays of expected values. For example { ‘tree.0’ ~=> [ ‘alfred’, ‘knuth’ ] } will trigger if the first element of msg equals alfred or knuth.

  • msg: The msg to place in the msg queue if the tracker matches the msg, the reaction.

Returns the tracker_id.

# File lib/ruote/dashboard.rb, line 1192
def add_tracker(wfid, action, tracker_id, conditions, msg)

  @context.tracker.add_tracker(wfid, action, tracker_id, conditions, msg)
apply_mutation(wfid, pdef) click to toggle source

Computes mutation and immediately applies it…

See compute_mutation

Return the mutation instance (forensic?)

# File lib/ruote/dashboard.rb, line 406
def apply_mutation(wfid, pdef), wfid,
attach(fei_or_fe, definition, opts={}) click to toggle source

Given a flow expression id, locates the corresponding ruote expression and attaches a subprocess to it.

Accepts the fei as a Hash or as a FlowExpressionId instance.

By default, the workitem of the expression you attach to provides the initial workitem for the attached branch. By using the :fields/:workitem or :merge_fields options, one can change that.

Returns the fei of the attached [root] expression (as a FlowExpressionId instance).

# File lib/ruote/dashboard.rb, line 194
def attach(fei_or_fe, definition, opts={})

  fe = Ruote.extract_fexp(@context, fei_or_fe).h
  fei = fe['fei']

  cfei = fei.merge(
    'expid' => "#{fei['expid']}_0",
    'subid' => Ruote.generate_subid(fei.inspect))

  tree =
  tree[0] = 'sequence'

  fields = fe['applied_workitem']['fields']
  if fs = opts[:fields] || opts[:workitem]
    fields = fs
  elsif fs = opts[:merge_fields]
    'launch', # "apply" is OK, but "launch" stands out better
    'parent_id' => fei,
    'fei' => cfei,
    'tree' => tree,
    'workitem' => { 'fields' => fields },
    'attached' => true)
cancel(wi_or_fei_or_wfid, opts={}) click to toggle source

Given a workitem or a fei, will do a cancel_expression, else it’s a wfid and it does a cancel_process.

A note about opts

They will get passed as is in the underlying ‘msg’, it can be useful to flag the message for historical purposes as in

dashboard.cancel(wfid, 'reason' => 'cleanup', 'user' => current_user)
# File lib/ruote/dashboard.rb, line 234
def cancel(wi_or_fei_or_wfid, opts={})

  do_misc('cancel', wi_or_fei_or_wfid, opts)
Also aliased as: cancel_process, cancel_expression
cancel_expression(wi_or_fei_or_wfid, opts={}) click to toggle source
Alias for: cancel
cancel_process(wi_or_fei_or_wfid, opts={}) click to toggle source
Alias for: cancel
compute_mutation(wfid, pdef) click to toggle source

Returns a Mutation instance listing all the operations necessary to transform the current tree of the process (wfid) into the given definition tree (pdef).

See also apply_mutation

# File lib/ruote/dashboard.rb, line 395
def compute_mutation(wfid, pdef), wfid,
configuration(config_key) click to toggle source

Returns a configuration value.

dashboard.configure('ruby_eval_allowed', true)

p dashboard.configuration('ruby_eval_allowed')
  # => true
# File lib/ruote/dashboard.rb, line 1013
def configuration(config_key)

configure(config_key, value) click to toggle source

Sets a configuration option. Examples:

# allow remote workflow definitions (for subprocesses or when launching
# processes)
@dashboard.configure('remote_definition_allowed', true)

# allow ruby_eval
@dashboard.configure('ruby_eval_allowed', true)
# File lib/ruote/dashboard.rb, line 1001
def configure(config_key, value)

  @context[config_key] = value
error(wi_or_fei) click to toggle source

Given a workitem or a fei (or a String version of a fei), returns the corresponding error (or nil if there is no other).

# File lib/ruote/dashboard.rb, line 493
def error(wi_or_fei)

  fei = Ruote.extract_fei(wi_or_fei)
  err ='errors', "err_#{fei.sid}")

  err ? : nil
errors(wfid=nil) click to toggle source

Returns an array of current errors (hashes)

Can be called in two ways :



dashboard.errors(:skip => 100, :limit => 100)
# File lib/ruote/dashboard.rb, line 477
def errors(wfid=nil)

  wfid, options = wfid.is_a?(Hash) ? [ nil, wfid ] : [ wfid, {} ]

  errs = wfid.nil? ?'errors', nil, options) :'errors', wfid)

  return errs if options[:count]

  errs.collect { |err| }
get_trackers(wfid=nil) click to toggle source

Returns a hash { tracker_id => tracker_hash } enumerating all the trackers in the ruote system.

# File lib/ruote/dashboard.rb, line 1216
def get_trackers(wfid=nil)['trackers']
history() click to toggle source

A shortcut for engine.context.history

# File lib/ruote/dashboard.rb, line 114
def history

join() click to toggle source

Joins the worker thread. If this engine has no nested worker, calling this method will simply return immediately.

# File lib/ruote/dashboard.rb, line 695
def join

  worker.join if worker
kill(wi_or_fei_or_wfid, opts={}) click to toggle source

Given a workitem or a fei, will do a kill_expression, else it’s a wfid and it does a kill_process.

(also see notes about opts for cancel)

# File lib/ruote/dashboard.rb, line 247
def kill(wi_or_fei_or_wfid, opts={})

  do_misc('cancel', wi_or_fei_or_wfid, opts.merge('flavour' => 'kill'))
Also aliased as: kill_process, kill_expression
kill_expression(wi_or_fei_or_wfid, opts={}) click to toggle source
Alias for: kill
kill_process(wi_or_fei_or_wfid, opts={}) click to toggle source
Alias for: kill
launch_single(process_definition, fields={}, variables={}, root_stash=nil) click to toggle source

Quick note : the implementation of launch is found in the module Ruote::ReceiverMixin that the engine includes.

Some processes have to have one and only one instance of themselves running, these are called ‘singles’ (‘singleton’ is too object-oriented).

When called, this method will check if an instance of the pdef is already running (it uses the process definition name attribute), if yes, it will return without having launched anything. If there is no such process running, it will launch it (and register it).

Returns the wfid (workflow instance id) of the running single.

# File lib/ruote/dashboard.rb, line 139
def launch_single(process_definition, fields={}, variables={}, root_stash=nil)

  tree =
  name = tree[1]['name'] || (tree[1].find { |k, v| v.nil? } || []).first

    'process definition is missing a name, cannot launch as single'
  ) unless name

  singles ='variables', 'singles') || {
    '_id' => 'singles', 'type' => 'variables', 'h' => {}
  wfid, timestamp = singles['h'][name]

  return wfid if wfid && (ps(wfid) || - timestamp < 1.0)
    # return wfid if 'singleton' process is already running

  wfid = @context.wfidgen.generate

  singles['h'][name] = [ wfid, ]

  r =

  return launch_single(tree, fields, variables, root_stash) unless r.nil?
    # the put failed, back to the start...
    # all this to prevent races between multiple engines,
    # multiple launch_single calls (from different Ruby runtimes)

  # ... green for launch
    'wfid' => wfid,
    'tree' => tree,
    'workitem' => { 'fields' => fields },
    'variables' => variables,
    'stash' => root_stash)

leftovers() click to toggle source

Warning : expensive operation.

Leftovers are workitems, errors and schedules belonging to process instances for which there are no more expressions left.

Better delete them or investigate why they are left here.

The result is a list of documents (hashes) as found in the storage. Each of them might represent a workitem, an error or a schedule.

If you want to delete one of them you can do
# File lib/ruote/dashboard.rb, line 554
def leftovers

  wfids ={})

  wis ='workitems').compact
  ers ='errors').compact
  scs ='schedules').compact
    # some slow storages need the compaction... [c]ouch...

  (wis + ers + scs).reject { |doc| wfids.include?(doc['fei']['wfid']) }
load_definition(path) click to toggle source

Loads (and turns into a tree) the process definition at the given path.

# File lib/ruote/dashboard.rb, line 702
def load_definition(path)
logger() click to toggle source

A shortcut for engine.context.logger

# File lib/ruote/dashboard.rb, line 121
def logger

noisy=(b) click to toggle source

A debug helper :

dashboard.noisy = true

will let the dashboard (in fact the worker) pour all the details of the executing process instances to STDOUT.

# File lib/ruote/dashboard.rb, line 1228
def noisy=(b)

  @context.logger.noisy = b
on_error() click to toggle source

Returns the process tree that is triggered in case of error.

Note that this ‘on_error’ doesn’t trigger if an on_error is defined in the process itself.

Returns nil if there is no ‘on_error’ set.

# File lib/ruote/dashboard.rb, line 1081
def on_error['trackers']['on_error']['msg']['tree']

on_error=(target) click to toggle source

Sets a participant or subprocess to be triggered when an error occurs in a process instance.

dashboard.on_error = participant_name

dashboard.on_error = subprocess_name

dashboard.on_error = Ruote.process_definition do

Note that this ‘on_error’ doesn’t trigger if an on_error is defined in the process itself.

# File lib/ruote/dashboard.rb, line 1118
def on_error=(target)

    nil, # do not track a specific wfid
    'error_intercepted', # react on 'error_intercepted' msgs
    'on_error', # the identifier
    nil, # no specific condition
    { 'action' => 'launch',
      'wfid' => 'replace',
      'tree' => target.is_a?(String) ?
        [ 'define', {}, [ [ target, {}, [] ] ] ] : target,
      'workitem' => 'replace',
      'variables' => 'compile' })
on_terminate() click to toggle source

Returns the process tree that is triggered in case of process termination.

Note that a termination process doesn’t raise a termination process when it terminates itself.

Returns nil if there is no ‘on_terminate’ set.

# File lib/ruote/dashboard.rb, line 1096
def on_terminate['trackers']['on_terminate']['msg']['tree']

on_terminate=(target) click to toggle source

Sets a participant or a subprocess that is to be launched/called whenever a regular process terminates.

dashboard.on_terminate = participant_name

dashboard.on_terminate = subprocess_name

dashboard.on_terminate = Ruote.define do

Note that a termination process doesn’t raise a termination process when it terminates itself.

on_terminate processes are not triggered for on_error processes. on_error processes are triggered for on_terminate processes as well.

# File lib/ruote/dashboard.rb, line 1151
def on_terminate=(target)

  msg = {
    'action' => 'launch',
    'tree' => target.is_a?(String) ?
      [ 'define', {}, [ [ target, {}, [] ] ] ] : target,
    'workitem' => 'replace' }

    nil,             # do not track a specific wfid
    'terminated',    # react on 'error_intercepted' msgs
    'on_terminate',  # the identifier
    nil,             # no specific condition
    msg)             # the message that gets triggered
participant(name) click to toggle source

Returns an instance of the participant registered under the given name. Returns nil if there is no participant registered for that name.

# File lib/ruote/dashboard.rb, line 970
def participant(name)

  @context.plist.lookup(name.to_s, nil)
participant_list() click to toggle source

Returns a list of Ruote::ParticipantEntry instances.

dashboard.register_participant :alpha, MyParticipant, 'message' => 'hello'

# interrogate participant list
list = dashboard.participant_list
participant = list.first
p participant.regex
  # => "^alpha$"
p participant.classname
  # => "MyParticipant"
p participant.options
  # => {"message"=>"hello"}

# update participant list
participant.regex = '^alfred$'
dashboard.participant_list = list
# File lib/ruote/dashboard.rb, line 925
def participant_list

participant_list=(pl) click to toggle source

Accepts a list of Ruote::ParticipantEntry instances or a list of

regex, [ classname, opts

] arrays.

See Engine#participant_list

Some examples :

dashboard.participant_list = [
  [ '^charly$', [ 'Ruote::StorageParticipant', {} ] ],
  [ '.+', [ 'MyDefaultParticipant', { 'default' => true } ]

This method writes the participant list in one go, it might be easier to use than to register participant one by ones.

# File lib/ruote/dashboard.rb, line 945
def participant_list=(pl)

  @context.plist.list = pl
pause(wi_or_fei_or_wfid, opts={}) click to toggle source

Given a wfid, will [attempt to] pause the corresponding process instance. Given an expression id (fei) will [attempt to] pause the expression and its children.

The only known option for now is :breakpoint => true, which lets the engine only pause the targetted expression.

fei and :breakpoint => true

By default, pausing an expression will pause that expression and all its children.

engine.pause(fei, :breakpoint => true)

will only flag as paused the given fei. When the children of that expression will reply to it, the execution for this branch of the process will stop, much like a break point.

# File lib/ruote/dashboard.rb, line 285
def pause(wi_or_fei_or_wfid, opts={})

  opts = Ruote.keys_to_s(opts)

    ':breakpoint option only valid when passing a workitem or a fei'
  ) if opts['breakpoint'] and wi_or_fei_or_wfid.is_a?(String)

  do_misc('pause', wi_or_fei_or_wfid, opts)
process(wfid) click to toggle source

Returns a ProcessStatus instance describing the current status of a process instance.

# File lib/ruote/dashboard.rb, line 436
def process(wfid)

  ProcessStatus.fetch(@context, [ wfid ], {}).first
process_ids() click to toggle source

Returns a [sorted] list of wfids of the process instances currently running in the engine.

This operation is substantially less costly than Engine#processes (though the ‘how substantially’ depends on the storage chosen).

# File lib/ruote/dashboard.rb, line 533
def process_ids{})
Also aliased as: process_wfids
process_wfids() click to toggle source
Alias for: process_ids
processes(opts={}) click to toggle source

Returns an array of ProcessStatus instances.

WARNING : this is an expensive operation, but it understands :skip and :limit, so pagination is our friend.

Please note, if you’re interested only in processes that have errors, Engine#errors is a more efficient means.

To simply list the wfids of the currently running, Engine#process_wfids is way cheaper to call.

# File lib/ruote/dashboard.rb, line 452
def processes(opts={})

  wfids =

  opts[:count] ? wfids.size : ProcessStatus.fetch(@context, wfids, opts)
ps(wfid=nil) click to toggle source

Returns a list of processes or the process status of a given process instance.

# File lib/ruote/dashboard.rb, line 462
def ps(wfid=nil)

  wfid == nil ? processes : process(wfid)
re_apply(fei, opts={}) click to toggle source

Re-applies an expression (given via its FlowExpressionId).

That will cancel the expression and, once the cancel operation is over (all the children have been cancelled), the expression will get re-applied.

The fei parameter may be a hash, a Ruote::FlowExpressionId instance, a Ruote::Workitem instance or a sid string.


:tree is used to completely change the tree of the expression at re_apply

  fei, :tree => [ 'participant', { 'ref' => 'bob' }, [] ])

:fields is used to replace the fields of the workitem at re_apply

  fei, :fields => { 'customer' => 'bob' })

:workitem is ok too

  fei, :workitem => { 'fields' => { 'customer' => 'bob' } })

:merge_in_fields is used to add / override fields

  fei, :merge_in_fields => { 'customer' => 'bob' })
# File lib/ruote/dashboard.rb, line 381
def re_apply(fei, opts={})
    'fei' => FlowExpressionId.extract_h(fei),
    're_apply' => Ruote.keys_to_s(opts))
register(*args, &block) click to toggle source

A shorter version of register_participant

dashboard.register 'alice', MailParticipant, :target => ''

or a block registering mechanism.

dashboard.register do
  alpha 'Participants::Alpha', 'flavour' => 'vanilla'
  participant 'bravo', 'Participants::Bravo', :flavour => 'peach'
  catchall ParticipantCharlie, 'flavour' => 'coconut'

Originally implemented in ruote-kit by Torsten Schoenebaum.

registration in block and :clear

By default, when registering multiple participants in block, ruote considers you’re wiping the participant list and re-adding them all.

You can prevent the clearing by stating :clear => false like in :

dashboard.register :clear => false do
  alpha 'Participants::Alpha', 'flavour' => 'vanilla'
  participant 'bravo', 'Participants::Bravo', :flavour => 'peach'
  catchall ParticipantCharlie, 'flavour' => 'coconut'
# File lib/ruote/dashboard.rb, line 877
def register(*args, &block)

  clear = args.first.is_a?(Hash) ? args.pop[:clear] : true

  if args.size > 0
    register_participant(*args, &block)
    proxy =, clear)
    block.arity < 1 ? proxy.instance_eval(&block) :
register_participant(regex, participant=nil, opts={}, &block) click to toggle source

Registers a participant in the engine.

Takes the form

dashboard.register_participant name_or_regex, klass, opts={}

With the form

dashboard.register_participant name_or_regex do |workitem|
  # ...

A BlockParticipant is automatically created.

name or regex

When registering participants, strings or regexes are accepted. Behind the scenes, a regex is kept.

Passing a string like “alain” will get ruote to automatically turn it into the following regex : /^alain$/.

For finer control over this, pass a regex directly

dashboard.register_participant /^user-/, MyParticipant
  # will match all workitems whose participant name starts with "user-"

some examples

dashboard.register_participant 'compute_sum' do |wi|
  wi.fields['sum'] = wi.fields['articles'].inject(0) do |s, (c, v)|
    s + c * v # sum + count * value
  # a block participant implicitely replies to the engine immediately

class MyParticipant
  def initialize(opts)
    @name = opts['name']
  def on_workitem
    workitem.fields['rocket_name'] = @name
  def on_cancel
    # do nothing

  /^moon-.+/, MyParticipant, 'name' => 'Saturn-V')

# computing the total for a invoice being passed in the workitem.
class TotalParticipant
  include Ruote::LocalParticipant

  def on_workitem
    workitem['total'] = workitem.fields['items'].inject(0.0) { |t, item|
      t + item['count'] * PricingService.lookup(item['id'])

  def on_cancel
dashboard.register_participant 'total', TotalParticipant

Remember that the options (the hash that follows the class name), must be serializable via JSON.

require_path and load_path

It’s OK to register a participant by passing its full classname as a String.

  'auditor', 'AuditParticipant', 'require_path' => 'part/audit.rb')
  'auto_decision', 'DecParticipant', 'load_path' => 'part/dec.rb')

Note the option load_path / require_path that point to the ruby file containing the participant implementation. ‘require’ will load and eval the ruby code only once, ‘load’ each time.

:override => false

By default, when registering a participant, if this results in a regex that is already used, the previously registered participant gets unregistered.

dashboard.register_participant 'alpha', AaParticipant
dashboard.register_participant 'alpha', BbParticipant, :override => false

This can be useful when the accept? method of participants are in use.

Note that using the register(&block) method, :override => false is automatically enforced.

dashboard.register do
  alpha AaParticipant
  alpha BbParticipant

:position / :pos => ‘last’ / ‘first’ / ‘before’ / ‘after’ / ‘over’

One can specify the position where the participant should be inserted in the participant list.

dashboard.register_participant 'auditor', AuditParticipant, :pos => 'last'
  • last : it’s the default, places the participant at the end of the list

  • first : top of the list

  • before : implies :override => false, places before the existing participant with the same regex

  • after : implies :override => false, places after the last existing participant with the same regex

  • over : overrides in the same position (while the regular, default overide removes and then places the new participant at the end of the list)

# File lib/ruote/dashboard.rb, line 834
def register_participant(regex, participant=nil, opts={}, &block)

  if participant.is_a?(Hash)
    opts = participant
    participant = nil

  pa = @context.plist.register(regex, participant, opts, block)
    'regex' => regex.is_a?(Regexp) ? regex.inspect : regex.to_s)

remove_process(wfid) click to toggle source

Removes a process by removing all its schedules, expressions, errors, workitems and trackers.

Warning: will not trigger any cancel behaviours at all, just removes the process.

# File lib/ruote/dashboard.rb, line 261
def remove_process(wfid)
remove_tracker(fei_sid_or_id, wfid=nil) click to toggle source

/!\ warning: advanced method.

Removes a tracker from the ruote system.

The first arg is a FlowExpressionId, in its instance form, hash form or shortened (sid) string form. It can also be any string (any tracker id).

The second arg is optional, it’s a wfid. It’s useful for some storage implementations (like ruote-swf) and helps determine how to grab the tracker list. Most of the ruote deployments don’t need that arg set.

# File lib/ruote/dashboard.rb, line 1208
def remove_tracker(fei_sid_or_id, wfid=nil)

  @context.tracker.remove_tracker(fei_sid_or_id, wfid)
replay_at_error(err) click to toggle source

Replays at a given error (hopefully the cause of the error got fixed before replaying…)

# File lib/ruote/dashboard.rb, line 322
def replay_at_error(err)

  err = error(err) unless err.is_a?(Ruote::ProcessError)

  msg = err.msg.dup

  if tree = msg['tree']
    # as soon as there is a tree, it means it's a re_apply

      'tree' => tree,
      'replay_at_error' => true,
      'workitem' => msg['workitem'])


    action = msg.delete('action')

    msg['replay_at_error'] = true
      # just an indication # remove error, msg) # trigger replay
respark(wfid, opts={}) click to toggle source

This method re_apply all the leaves of a process instance. It’s meant to be used against stalled workflows to give them back the spark of life.

Stalled workflows can happen when msgs get lost. It also happens with some storage implementations where msgs are stored differently from expressions and co.

By default, it doesn’t re_apply leaves that are in error. If the ‘errors_too’ option is set to true, it will re_apply leaves in error as well. For example:

$dashboard.respark(wfid, 'errors_too' => true)
# File lib/ruote/dashboard.rb, line 425
def respark(wfid, opts={})
    'wfid' => wfid,
    'respark' => Ruote.keys_to_s(opts))
resume(wi_or_fei_or_wfid, opts={}) click to toggle source

Given a wfid will [attempt to] resume the process instance. Given an expression id (fei) will [attempt to] to resume the expression and its children.

Note : this is supposed to be called on paused expressions / instances, this is NOT meant to be called to unstuck / unhang a process.

resume(wfid, :anyway => true)

Resuming a process instance is equivalent to calling resume on its root expression. If the root is not paused itself, this will have no effect.

dashboard.resume(wfid, :anyway => true)

will make sure to call resume on each of the paused branch within the process instance (tree), effectively resuming the whole process.

# File lib/ruote/dashboard.rb, line 314
def resume(wi_or_fei_or_wfid, opts={})

  do_misc('resume', wi_or_fei_or_wfid, opts)
schedules(wfid=nil) click to toggle source

Returns an array of schedules. Those schedules are open structs with various properties, like target, owner, at, put_at, …

Introduced mostly for ruote-kit.

Can be called in two ways :



dashboard.schedules(:skip => 100, :limit => 100)
# File lib/ruote/dashboard.rb, line 514
def schedules(wfid=nil)

  wfid, options = wfid.is_a?(Hash) ? [ nil, wfid ] : [ wfid, {} ]

  scheds = wfid.nil? ?'schedules', nil, options) :'schedules', /!#{wfid}-\d+$/)

  return scheds if options[:count]

  scheds.collect { |s| Ruote.schedule_to_h(s) }.sort_by { |s| s['wfid'] }
shutdown() click to toggle source

Shuts down the engine, mostly passes the shutdown message to the other services and hope they’ll shut down properly.

# File lib/ruote/dashboard.rb, line 569
def shutdown

storage() click to toggle source

Returns the storage this engine works with passed at engine initialization.

# File lib/ruote/dashboard.rb, line 98
def storage
storage_participant() click to toggle source

A convenience method for

sp =

simply do

sp = dashboard.storage_participant
# File lib/ruote/dashboard.rb, line 958
def storage_participant

  @storage_participant ||=
Also aliased as: worklist
unregister(name_or_participant) click to toggle source
unregister_participant(name_or_participant) click to toggle source

Removes/unregisters a participant from the engine.

# File lib/ruote/dashboard.rb, line 892
def unregister_participant(name_or_participant)

  re = @context.plist.unregister(name_or_participant)

  raise('participant not found')) unless re
    'regex' => re.to_s)
Also aliased as: unregister
update_expression(fei, opts) click to toggle source

Warning: advanced method.

Currently only used by mutations.

# File lib/ruote/dashboard.rb, line 1237
def update_expression(fei, opts)

  fei = Ruote.extract_fei(fei)
  fexp = Ruote::Exp::FlowExpression::fetch(@context, fei)

    "no expression found with fei #{fei.sid}"
  ) unless fexp

  if t = opts[:tree]
    fexp.h.updated_tree = opts[:tree]

  r =

    "expression #{fei.sid} is gone"
  ) if r == true

  return update_expression(fei, opts) unless r.nil?
wait_for(*items) click to toggle source

This method expects there to be a logger with a wait_for method in the context, else it will raise an exception.

WARNING: wait_for() is meant for environments where there is a unique worker and that worker is nested in this engine. In a multiple worker environment wait_for doesn't see events handled by 'other' workers.

This method is only useful for test/quickstart/examples environments.

  # will make the current thread block until a workitem is delivered
  # to the participant named 'alpha'

  # will make the current thread block until the processed whose
  # wfid is given (String) terminates or produces an error.

  # will make the current thread block until 5 messages have been
  # processed on the workqueue...

  # will return as soon as the engine/storage is empty, ie as soon
  # as there are no more processes running in the engine (no more
  # expressions placed in the storage)

  # will return as soon as any process has a 'terminated' event.

It’s OK to wait for multiple wfids:

engine.wait_for('20100612-bezerijozo', '20100612-yakisoba')

If one needs to wait for something else than a wfid but needs to break in case of error:

engine.wait_for(:alpha, :or_error)

ruote 2.3.0 and wait_for(event)

Ruote 2.3.0 introduced the ability to wait for an event given its name. Here is a quick list of event names and a their description:

  • ‘launch’ - [sub]process launch

  • ‘terminated’ - process terminated

  • ‘ceased’ - orphan process terminated

  • ‘apply’ - expression application

  • ‘reply’ - expression reply

  • ‘dispatched’ - emitted workitem towards participant

  • ‘receive’ - received workitem from participant

  • ‘pause’ - pause order

  • ‘resume’ - pause order

  • ‘dispatch_cancel’ - emitting a cancel order to a participant

  • ‘dispatch_pause’ - emitting a pause order to a participant

  • ‘dispatch_resume’ - emitting a resume order to a participant

Names that are past participles are for notification events, while plain verbs are for action events. Most of the time, a notitication is emitted has the result of an action event, workers don’t take any action on them, but services that are listening to the ruote activity might want to do something about them.

ruote 2.3.0 and wait_for(hash)

For more precise testing, wait_for accepts hashes, for example:

r = dashboard.wait_for('action' => 'apply', 'exp_name' => 'wait')

will block until a wait expression is applied.

If you know ruote msgs, you can pinpoint at will:

r = dashboard.wait_for(
  'action' => 'apply',
  'exp_name' => 'wait',
  'fei.wfid' => wfid)

what wait_for returns

wait_for returns the intercepted event. It's useful when testing/ spec'ing, as in:

it 'completes successfully' do

  definition = Ruote.define :on_error => 'charly' do

  wfid = @board.launch(definition)

  r = @board.wait_for(wfid)
    # wait until process terminates or hits an error

  r['workitem'].should_not == nil
  r['workitem']['fields']['alpha'].should == 'was here'
  r['workitem']['fields']['bravo'].should == 'was here'
  r['workitem']['fields']['charly'].should == nil

:timeout option

One can pass a timeout value in seconds for the wait_for call, as in:

dashboard.wait_for(wfid, :timeout => 5 * 60)

The default timeout is 60 (seconds). A nil or negative timeout disables the timeout.

# File lib/ruote/dashboard.rb, line 685
def wait_for(*items)

  opts = (items.size > 1 && items.last.is_a?(Hash)) ? items.pop : {}

  @context.logger.wait_for(items, opts)
worker() click to toggle source

Returns the worker nested inside this engine (passed at initialization). Returns nil if this engine is only linked to a storage (and the worker is running somewhere else (hopefully)).

# File lib/ruote/dashboard.rb, line 107
def worker

worker_info() click to toggle source

Returns the hash containing info about each worker connected to the storage.

# File lib/ruote/dashboard.rb, line 1021
def worker_info

  ('variables', 'workers') || {})['workers']
worker_state() click to toggle source

Returns the state the workers are supposed to be in right now. It’s usually ‘running’, but it could be ‘stopped’ or ‘paused’.

# File lib/ruote/dashboard.rb, line 1029
def worker_state

  doc ='variables', 'worker') ||
    { 'type' => 'variables', '_id' => 'worker', 'state' => 'running' }

worker_state=(state) click to toggle source

Sets the [desired] worker state. The workers will check that target state at their next beat and switch to it.

Setting the state to ‘stopped’ will force the workers to stop as soon as they notice the new state.

Setting the state to ‘paused’ will force the workers to pause. They will not process msgs until the state is set back to ‘running’.

By default the [engine] option ‘worker_state_enabled’ is not set, so calling this method will result in a error, unless ‘worker_state_enabled’ was set to true when the storage was initialized.

# File lib/ruote/dashboard.rb, line 1053
def worker_state=(state)

    "'worker_state_enabled' is not set, cannot change state"
  ) unless @context['worker_state_enabled']

  state = state.to_s

    "#{state.inspect} not in #{WORKER_STATES.inspect}"
  ) unless WORKER_STATES.include?(state)

  doc ='variables', 'worker') ||
    { 'type' => 'variables', '_id' => 'worker', 'state' => 'running' }

  doc['state'] = state && worker_state=(state)
worklist() click to toggle source

worklist or storage_participant

Alias for: storage_participant

Protected Instance Methods

do_misc(action, wi_or_fei_or_wfid, opts) click to toggle source

Used by pause and resume.

# File lib/ruote/dashboard.rb, line 1263
def do_misc(action, wi_or_fei_or_wfid, opts)

  opts = Ruote.keys_to_s(opts)

  target = Ruote.extract_id(wi_or_fei_or_wfid)

  if action == 'resume' && opts['anyway']
    # determines the roots of the branches that are paused
    # sends the resume message to them.

    exps = ps(target) { |fexp| fexp.state == 'paused' }
    feis = exps.collect { |fexp| fexp.fei }

    roots = exps.inject([]) { |a, fexp|
      a << fexp.fei.h unless feis.include?(fexp.parent_id)

    roots.each { |fei|'resume', 'fei' => fei) }

  elsif target.is_a?(String)
    # action targets a process instance (a string wfid)
      "#{action}_process", opts.merge('wfid' => target))

      action, opts.merge('fei' => target))


Generated with the Darkfish Rdoc Generator 2.