0.1.9 breaks with to_sym for nil:NilClass about transitions

[matthewford]

My statemachine: https://github.com/alphagov/whitehall/blob/postgresql/app/models/edition/workflow.rb

[matthewford]

Specifically when using postgresql

[troessner]

First of all, why are you doing:

include ::Transitions
include ActiveRecord::Transitions

?

I think

include ActiveModel::Transitions

should be ok.

Second, "specifically" means "not on mysql" in this case?
I am asking because we had quite a problems that occured using postgresql, but not mysql.

[matthewford]

Yes, with our mysql branch I'm seeing no issues.

[troessner]

Interesting. I'll look into this but i can't promise anything until the weekend because of work. @barelyknown are you free to look into this?

[matthewford]

Cheers, No rush.

[barelyknown]

I may be able to take a look tomorrow morning.

[matthewford]

More info: apparently we see this on <3.2 on Mysql as well so it might be an active model issue

[troessner]

@barelyknown any updates on this?

[barelyknown]

I haven't looked into it yet. Totally forgot after the initial email thread.

[dbelwood]

FWIW, I'm getting the same error with this line -
Subscription.select("id, user_id").where("updated_at >= ?", Time.at(Benchprep::Commands::UpdateUserTotalRevenue.get_last_run_at).to_formatted_s(:db)).order(:id)

But not this -
Subscription.select("id, user_id, state").where("updated_at >= ?", Time.at(Benchprep::Commands::UpdateUserTotalRevenue.get_last_run_at).to_formatted_s(:db)).order(:id)

The lack of a state column seems to be the culprit. The second line works in 0.1.3, but not 0.1.9

[troessner]

Sorry guys for having let this issue remain open so long. I'll try to take care of it asap.
But first, can we quickly gather what we know about this issue?

So am I getting this right @dbelwood and @matthewford :

• The issue appears only with AR < 3.2
• Happens on mysql / postgresql
• Does not appear with 0.1.3, but with 0.1.9 it does

Can you confirm this?

[matthewford]

I haven't tried on 0.1.3, but I can confirm the other two

[troessner]

Thanks, I will definitely look into this this week.

[fairchild]

fyi, i had the same error (running postgresql) and eliminated the error by removing the ranked_model gem that I was also using. I did not dig in enough to figure out why, but this might help figure out why an error is being thrown.

[troessner]

Alright guys,

I need a little help to reproduce this problem. Please have a look at this feature branch i just created:

master...issue_95

It uses:

• a custom select without the state column
• mysql
• activerecord 3.1.0:

( $ (issue_95)) bundle list | grep -i record
activerecord (3.1.0)

So I'd have expected the associated test to fail. Only it doesn't.

How can I reproduce the behaviour you reported?

@fairchild thanks for the additional info, but I think I'll try to reproduce the less complicated set up..:)

[corroded]

We have also experienced this in 0.1.5 (I upgrade to 0.1.9) hoping it fixes it but the issue still occurs. We also notice that it complains about the state attribute missing:

1.9.3 (main):0 > Customer.select :id
  Customer Load (0.4ms)  SELECT id FROM "customers"
=> [#<Customer id: 1>, #<Customer id: 6454>, #<Customer id: 6455>, #<Customer id: 6456>, #<Customer id: 6457>, #<Customer id: 6458>, #<Customer id: 6459>, #<Customer id: 6460>, #<Customer id: 6461>, #<Customer id: 6462>, #<Customer id: 6463>, #<Customer id: 6464>, #<Customer id: 6465>, #<Customer id: 6466>, #<Customer id: 6467>, #<Customer id: 12906>, #<Customer id: 12907>]
1.9.3 (main):0 > cc = _
=> [#<Customer id: 1>, #<Customer id: 6454>, #<Customer id: 6455>, #<Customer id: 6456>, #<Customer id: 6457>, #<Customer id: 6458>, #<Customer id: 6459>, #<Customer id: 6460>, #<Customer id: 6461>, #<Customer id: 6462>, #<Customer id: 6463>, #<Customer id: 6464>, #<Customer id: 6465>, #<Customer id: 6466>, #<Customer id: 6467>, #<Customer id: 12906>, #<Customer id: 12907>]
1.9.3 (main):0 > cc.inspect
=> "[#<Customer id: 1>, #<Customer id: 6454>, #<Customer id: 6455>, #<Customer id: 6456>, #<Customer id: 6457>, #<Customer id: 6458>, #<Customer id: 6459>, #<Customer id: 6460>, #<Customer id: 6461>, #<Customer id: 6462>, #<Customer id: 6463>, #<Customer id: 6464>, #<Customer id: 6465>, #<Customer id: 6466>, #<Customer id: 6467>, #<Customer id: 12906>, #<Customer id: 12907>]"

1.9.3 (main):0 > jj = Job.select :id
  Job Load (0.3ms)  SELECT id FROM "jobs"
=> #<ActiveRecord::Relation:0x3fe9da7a31b4>
1.9.3 (main):0 > jj.inspect
  Job Load (0.5ms)  SELECT id FROM "jobs"
ActiveModel::MissingAttributeError: missing attribute: state
from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/transitions-0.1.5/lib/active_model/transitions.rb:66:in `set_initial_state'
1.9.3 (main):0 > jj.map
  Job Load (0.4ms)  SELECT id FROM "jobs"
  Job Load (0.2ms)  SELECT id FROM "jobs"
ActiveModel::MissingAttributeError: missing attribute: state
from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/transitions-0.1.5/lib/active_model/transitions.rb:66:in `set_initial_state'
1.9.3 (main):0 > jj.first
  Job Load (0.4ms)  SELECT id FROM "jobs" LIMIT 1
ActiveModel::MissingAttributeError: missing attribute: state
from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/transitions-0.1.5/lib/active_model/transitions.rb:66:in `set_initial_state'

1.9.3 (main):0 > ii = Invoice.select :id
  Invoice Load (0.4ms)  SELECT id FROM "invoices"
=> #<ActiveRecord::Relation:0x3fe9daeecbb4>
1.9.3 (main):0 > ii.inspect
  Invoice Load (0.3ms)  SELECT id FROM "invoices"
ActiveModel::MissingAttributeError: missing attribute: state
from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/transitions-0.1.5/lib/active_model/transitions.rb:66:in `set_initial_state'

as you can see from above, we are trying to select distinct ids from jobs. It failed for us so we tried it on other models. From the console logs above, you can see that it works perfectly fine on customers, but fails on jobs and invoices. Jobs and invoices both have transitions, while customer is a plain AR model.

when we try any selector for Job or Invoice (using either first or last, we get that MissingAttributeError. One thing to note that might help:

1.9.3 (main):0 > jj.pluck 1
   (0.4ms)  SELECT id, 1 FROM "jobs"
=> [1, 2, 3, 4, 5, 7, 17]

Don't know why pluck gives us that while map, first or last doesn't. Examples above are from 0.1.5. But we also tried 0.1.9:

1.9.3p194 :002 > Job.select :id
Creating scope :all. Overwriting existing method Job.all.
  Job Load (0.4ms)  SELECT id FROM "jobs"
(Object doesn't support #inspect)
 =>
1.9.3p194 :003 > Job.select(:id).first
  Job Load (0.3ms)  SELECT id FROM "jobs" LIMIT 1
NoMethodError: undefined method `to_sym' for nil:NilClass
    from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/transitions-0.1.9/lib/active_model/transitions.rb:91:in `set_initial_state'
    from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/activesupport-3.2.11/lib/active_support/callbacks.rb:405:in `_run__2958716729451653381__initialize__4056324975497312426__callbacks'
    from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/activesupport-3.2.11/lib/active_support/callbacks.rb:405:in `__run_callback'
    from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/activesupport-3.2.11/lib/active_support/callbacks.rb:385:in `_run_initialize_callbacks'
    from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/activesupport-3.2.11/lib/active_support/callbacks.rb:81:in `run_callbacks'
    from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/activerecord-3.2.11/lib/active_record/base.rb:523:in `init_with'
    from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/activerecord-3.2.11/lib/active_record/inheritance.rb:68:in `instantiate'
    from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/activerecord-3.2.11/lib/active_record/querying.rb:38:in `block (2 levels) in find_by_sql'
    from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/activerecord-3.2.11/lib/active_record/querying.rb:38:in `collect!'
    from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/activerecord-3.2.11/lib/active_record/querying.rb:38:in `block in find_by_sql'
    from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/activerecord-3.2.11/lib/active_record/explain.rb:40:in `logging_query_plan'
    from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/activerecord-3.2.11/lib/active_record/querying.rb:37:in `find_by_sql'
    from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/newrelic_rpm-3.6.5.130/lib/new_relic/agent/method_tracer.rb:518:in `block in find_by_sql_with_trace_ActiveRecord_self_name_find_by_sql'
    from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/newrelic_rpm-3.6.5.130/lib/new_relic/agent/method_tracer.rb:268:in `trace_execution_scoped'
    from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/newrelic_rpm-3.6.5.130/lib/new_relic/agent/method_tracer.rb:513:in `find_by_sql_with_trace_ActiveRecord_self_name_find_by_sql'
    from /Users/dev/.rvm/gems/ruby-1.9.3-p194@spanner/gems/activerecord-3.2.11/lib/active_record/relation.rb:171:in `exec_queries'

Same result with map, last and [0]. BUT!

1.9.3p194 :018 >   Job.select(:id).pluck 1
   (0.3ms)  SELECT id, 1 FROM "jobs"
 => [1, 2, 3, 4, 5, 7, 17]

I hope those logs I pasted aren't just jargon and will actually help in solving the problem.

I might try checking out the feature branch either tonight or tomorrow and see how it goes.

[troessner]

Ah, thanks a lot, got it!

bdc1040

-> Ok, i can reproduce this now, I'll fix this a.s.a.p

[corroded]

sweet! Thanks! hopefully it's not a major one...

[corroded]

hi @troessner, just a question if there's a fix already for this? Thanks!

[troessner]

I just released 0.1.10 which fixes this problem as far as I can see - please try it out and let me know if that works for you.
If not, feel free to re-open this issue.

[troessner]

Sooooo, any feedback, guys?

[matthewford]

Hi Timo,

I'll see if I can replicate the issue with the patch on Monday, thanks for
the fix.

Matthew Ford

Director of Bit Zesty

T: [phone]

This message contains confidential information and is intended only for the
individual named. If you are not the named addressee you should not
disseminate, distribute or copy this e-mail. Please notify Bit Zesty
immediately by e-mail if you have received this e-mail by mistake and
delete this e-mail from your system. Bit Zesty does not accept liability
for any errors or omissions in the contents of this message, which arise as
a result of e-mail transmission. Opinions expressed in this email are those
of Matthew Ford, and do not necessarily reflect those of Bit Zesty.

Bit Zesty Ltd, a company incorporated in England with registered company
number 06883289.

On 9 August 2013 12:00, Timo Rößner [email protected] wrote:

Sooooo, any feedback, guys?

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/95#issuecomment-22387752
.

[corroded]

sorry haven't tried it yet in our code since it wasn't scheduled. We will try it out on monday and give some feedback. Thanks for this!