rspec users - Dec 2007 - Change in isolation behaviour 1.08 - 1.10 ?

Hi,

I just moved from 1.08 to 1.10 and now have one example failing, which,
under 1.08, passed. Is the due to a change in behaviour?

Here''s my spec (removed some passing examples)

require File.dirname(__FILE__) + ''/../spec_helper''

describe "A user" do

  before(:each) do
    @user = User.new
    @valid_user = User.new(
      :email    => ''bert.valid at lostboys.nl'',
      :fname    => ''bert'',
      :lname    => ''valid'',
      :jobtitle => "programmer"
    )
  end

  it "should have a unique email address" do
    @valid_user.save.should == true
    @user.email = @valid_user.email
    @user.should have(1).error_on(:email)
  end
  
  it "should allow two users with the same name" do
    @valid_user.save.should == true
    @user.fname = @valid_user.fname
    @user.lname = @valid_user.lname
    @user.should have(:no).error_on(:fname)
    @user.should have(:no).error_on(:lname)
  end
  
end

The second example now fails, as the inserted record from the first example
is not rolled back.

I can just put a User.delete_all in an after(:each) block but I think it
would be nicer if Rspec wrapped each example in a db transaction.
Wasn''t
this they way things worked in 1.08 ?

Jeroen

On Dec 17, 2007 2:13 PM, Jeroen Houben <jeroen.houben at lostboys.nl>
wrote:
> Hi,
>
> I just moved from 1.08 to 1.10 and now have one example failing, which,
> under 1.08, passed. Is the due to a change in behaviour?
>
> Here''s my spec (removed some passing examples)
>
> require File.dirname(__FILE__) + ''/../spec_helper''
>
> describe "A user" do
>
>   before(:each) do
>     @user = User.new
>     @valid_user = User.new(
>       :email    => ''bert.valid at lostboys.nl'',
>       :fname    => ''bert'',
>       :lname    => ''valid'',
>       :jobtitle => "programmer"
>     )
>   end
>
>   it "should have a unique email address" do
>     @valid_user.save.should == true
>     @user.email = @valid_user.email
>     @user.should have(1).error_on(:email)
>   end
>
>   it "should allow two users with the same name" do
>     @valid_user.save.should == true
>     @user.fname = @valid_user.fname
>     @user.lname = @valid_user.lname
>     @user.should have(:no).error_on(:fname)
>     @user.should have(:no).error_on(:lname)
>   end
>
> end
>
> The second example now fails, as the inserted record from the first example
> is not rolled back.
>
stack trace please
> I can just put a User.delete_all in an after(:each) block but I think it
> would be nicer if Rspec wrapped each example in a db transaction.
Wasn''t
> this they way things worked in 1.08 ?
>
> Jeroen
>
> _______________________________________________
> rspec-users mailing list
> rspec-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rspec-users
>

jeroen$ ./script/spec spec/models/user_spec.rb -b
...F..................P..

Pending:
A valid user should add the correct error on dob when using an bogus dob
assignment  (with only a month) (met max dob is beetje lastig op win32 lijkt
het :-s)

1)
''A user should allow two users with the same name'' FAILED
expected: true,
     got: false (using ==)
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/expectations.rb:52:i
n `fail_with''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/matchers/operator_ma
tcher.rb:46:in `fail_with_message''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/matchers/operator_ma
tcher.rb:56:in `__delegate_method_missing_to_target''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/matchers/operator_ma
tcher.rb:12:in `==''
./spec/models/user_spec.rb:36:
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/example/example_meth
ods.rb:79:in `instance_eval''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/example/example_meth
ods.rb:79:in `run_with_description_capturing''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/matchers.rb:144:in
`capture_generated_description''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/example/example_meth
ods.rb:78:in `run_with_description_capturing''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/example/example_meth
ods.rb:19:in `execute''
/usr/local/lib/ruby/1.8/timeout.rb:48:in `timeout''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/example/example_meth
ods.rb:16:in `execute''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/example/example_grou
p_methods.rb:280:in `execute_examples''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/example/example_grou
p_methods.rb:279:in `each''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/example/example_grou
p_methods.rb:279:in `execute_examples''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/example/example_grou
p_methods.rb:120:in `run''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/runner/example_group
_runner.rb:22:in `run''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/runner/example_group
_runner.rb:21:in `each''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/runner/example_group
_runner.rb:21:in `run''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/runner/options.rb:87
:in `run_examples''
/Users/jeroen/lostboys.nl/vendor/plugins/rspec/lib/spec/runner/command_line.
rb:19:in `run''
./script/spec:4:

Finished in 0.391897 seconds

25 examples, 1 failure, 1 pending


On 12/17/07 3:17 PM, "aslak hellesoy" <aslak.hellesoy at
gmail.com> wrote:
> On Dec 17, 2007 2:13 PM, Jeroen Houben <jeroen.houben at lostboys.nl>
wrote:
>> Hi,
>> 
>> I just moved from 1.08 to 1.10 and now have one example failing, which,
>> under 1.08, passed. Is the due to a change in behaviour?
>> 
>> Here''s my spec (removed some passing examples)
>> 
>> require File.dirname(__FILE__) + ''/../spec_helper''
>> 
>> describe "A user" do
>> 
>>   before(:each) do
>>     @user = User.new
>>     @valid_user = User.new(
>>       :email    => ''bert.valid at lostboys.nl'',
>>       :fname    => ''bert'',
>>       :lname    => ''valid'',
>>       :jobtitle => "programmer"
>>     )
>>   end
>> 
>>   it "should have a unique email address" do
>>     @valid_user.save.should == true
>>     @user.email = @valid_user.email
>>     @user.should have(1).error_on(:email)
>>   end
>> 
>>   it "should allow two users with the same name" do
>>     @valid_user.save.should == true
>>     @user.fname = @valid_user.fname
>>     @user.lname = @valid_user.lname
>>     @user.should have(:no).error_on(:fname)
>>     @user.should have(:no).error_on(:lname)
>>   end
>> 
>> end
>> 
>> The second example now fails, as the inserted record from the first
example
>> is not rolled back.
>> 
> 
> stack trace please
> 
>> I can just put a User.delete_all in an after(:each) block but I think
it
>> would be nicer if Rspec wrapped each example in a db transaction.
Wasn''t
>> this they way things worked in 1.08 ?
>> 
>> Jeroen
>> 
>> _______________________________________________
>> rspec-users mailing list
>> rspec-users at rubyforge.org
>> http://rubyforge.org/mailman/listinfo/rspec-users
>> 
> _______________________________________________
> rspec-users mailing list
> rspec-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rspec-users

On Dec 17, 2007 2:13 PM, Jeroen Houben <jeroen.houben at lostboys.nl>
wrote:
> Hi,
>
> I just moved from 1.08 to 1.10 and now have one example failing, which,
> under 1.08, passed. Is the due to a change in behaviour?
>
> Here''s my spec (removed some passing examples)
>
> require File.dirname(__FILE__) + ''/../spec_helper''
>
> describe "A user" do
>
>   before(:each) do
>     @user = User.new
>     @valid_user = User.new(
>       :email    => ''bert.valid at lostboys.nl'',
>       :fname    => ''bert'',
>       :lname    => ''valid'',
>       :jobtitle => "programmer"
>     )
>   end
>
>   it "should have a unique email address" do
>     @valid_user.save.should == true
>     @user.email = @valid_user.email
>     @user.should have(1).error_on(:email)
>   end
>
>   it "should allow two users with the same name" do
>     @valid_user.save.should == true
>     @user.fname = @valid_user.fname
>     @user.lname = @valid_user.lname
>     @user.should have(:no).error_on(:fname)
>     @user.should have(:no).error_on(:lname)
>   end
>
> end
>
> The second example now fails, as the inserted record from the first example
> is not rolled back.
>
I don''t see a fixtures :users in your spec. Do you have this in your
spec_helper?
Have you rerun script/generate rspec after you upgraded? There are
some changes in the spec_helper.rb file between the two releases.

Aslak
> I can just put a User.delete_all in an after(:each) block but I think it
> would be nicer if Rspec wrapped each example in a db transaction.
Wasn''t
> this they way things worked in 1.08 ?
>
> Jeroen
>
> _______________________________________________
> rspec-users mailing list
> rspec-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rspec-users
>

On Dec 17, 2007 8:47 AM, aslak hellesoy <aslak.hellesoy at gmail.com>
wrote:
> On Dec 17, 2007 2:13 PM, Jeroen Houben <jeroen.houben at lostboys.nl>
wrote:
>
> > Hi,
> >
> > I just moved from 1.08 to 1.10 and now have one example failing,
which,
> > under 1.08, passed. Is the due to a change in behaviour?
> >
> > Here''s my spec (removed some passing examples)
> >
> > require File.dirname(__FILE__) + ''/../spec_helper''
> >
> > describe "A user" do
> >
> >   before(:each) do
> >     @user = User.new
> >     @valid_user = User.new(
> >       :email    => ''bert.valid at lostboys.nl'',
> >       :fname    => ''bert'',
> >       :lname    => ''valid'',
> >       :jobtitle => "programmer"
> >     )
> >   end
> >
> >   it "should have a unique email address" do
> >     @valid_user.save.should == true
> >     @user.email = @valid_user.email
> >     @user.should have(1).error_on(:email)
> >   end
> >
> >   it "should allow two users with the same name" do
> >     @valid_user.save.should == true
> >     @user.fname = @valid_user.fname
> >     @user.lname = @valid_user.lname
> >     @user.should have(:no).error_on(:fname)
> >     @user.should have(:no).error_on(:lname)
> >   end
> >
> > end
> >
> > The second example now fails, as the inserted record from the first
example
> > is not rolled back.
> >
>
> I don''t see a fixtures :users in your spec. Do you have this in
your
> spec_helper?
> Have you rerun script/generate rspec after you upgraded? There are
> some changes in the spec_helper.rb file between the two releases.
This is actually a regression in 1.1.0 that has been fixed in trunk.
We''ll do a 1.1.1 release later today, but go ahead and update to trunk
for the time being and you should be good.

Cheers,
David
>
> Aslak
>
>
> > I can just put a User.delete_all in an after(:each) block but I think
it
> > would be nicer if Rspec wrapped each example in a db transaction.
Wasn''t
> > this they way things worked in 1.08 ?
> >
> > Jeroen
> >
> > _______________________________________________
> > rspec-users mailing list
> > rspec-users at rubyforge.org
> > http://rubyforge.org/mailman/listinfo/rspec-users
> >
> _______________________________________________
> rspec-users mailing list
> rspec-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rspec-users
>

> 
> This is actually a regression in 1.1.0 that has been fixed in trunk.
> We''ll do a 1.1.1 release later today, but go ahead and update to
trunk
> for the time being and you should be good.
> 
Great, thanks. Will I get 1.1.1 automatically if I''m using
svn:externals and
the CURRENT tag? 

jeroen

On Dec 17, 2007 8:52 AM, Jeroen Houben <jeroen.houben at lostboys.nl>
wrote:
> >
> > This is actually a regression in 1.1.0 that has been fixed in trunk.
> > We''ll do a 1.1.1 release later today, but go ahead and update
to trunk
> > for the time being and you should be good.
> >
>
> Great, thanks. Will I get 1.1.1 automatically if I''m using
svn:externals and
> the CURRENT tag?
When we do the release, yes, but not until then.
>
>
> jeroen
>
> _______________________________________________
> rspec-users mailing list
> rspec-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rspec-users
>

> 
> I don''t see a fixtures :users in your spec. Do you have this in
your
> spec_helper?
Not using fixtures in this particular spec
> Have you rerun script/generate rspec after you upgraded?
Yes, but manually copied over the changes in spec_helper.rb
> There are some changes in the spec_helper.rb file between the two releases.
 
... Just received the message by David. This will be fixed in 1.1.1

Thanks for your help!

Jeroen

On Dec 17, 2007 8:55 AM, Jeroen Houben <jeroen.houben at lostboys.nl>
wrote:
>
> >
> > I don''t see a fixtures :users in your spec. Do you have this
in your
> > spec_helper?
>
> Not using fixtures in this particular spec
>
> > Have you rerun script/generate rspec after you upgraded?
>
> Yes, but manually copied over the changes in spec_helper.rb
>
> > There are some changes in the spec_helper.rb file between the two
releases.
>
> ... Just received the message by David. This will be fixed in 1.1.1
Jeroen - just for good measure, would you kindly update from trunk and
verify that it does in fact fix your issue?
>
> Thanks for your help!
>
>
> Jeroen
>
> _______________________________________________
> rspec-users mailing list
> rspec-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rspec-users
>

>> 
>> ... Just received the message by David. This will be fixed in 1.1.1
> 
> Jeroen - just for good measure, would you kindly update from trunk and
> verify that it does in fact fix your issue?
Finished in 3.42683 seconds

176 examples, 0 failures, 3 pending

:)

On Dec 17, 2007 9:06 AM, Jeroen Houben <jeroen.houben at lostboys.nl>
wrote:
> >>
> >> ... Just received the message by David. This will be fixed in
1.1.1
> >
> > Jeroen - just for good measure, would you kindly update from trunk and
> > verify that it does in fact fix your issue?
>
> Finished in 3.42683 seconds
>
> 176 examples, 0 failures, 3 pending
Thanks. I''ll get the release out by tonight.

Cheers,
David
>
> :)
>
>
> _______________________________________________
> rspec-users mailing list
> rspec-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rspec-users
>

On Dec 17, 2007 4:12 PM, David Chelimsky <dchelimsky at gmail.com>
wrote:
> On Dec 17, 2007 9:06 AM, Jeroen Houben <jeroen.houben at lostboys.nl>
wrote:
> > >>
> > >> ... Just received the message by David. This will be fixed in
1.1.1
> > >
> > > Jeroen - just for good measure, would you kindly update from
trunk and
> > > verify that it does in fact fix your issue?
> >
> > Finished in 3.42683 seconds
> >
> > 176 examples, 0 failures, 3 pending
>
> Thanks. I''ll get the release out by tonight.
>
Awesome David.

Do we have any specs that will catch this regression if it''s
reintroduced?

Aslak
> Cheers,
> David
>
>
> >
> > :)
> >
> >
> > _______________________________________________
> > rspec-users mailing list
> > rspec-users at rubyforge.org
> > http://rubyforge.org/mailman/listinfo/rspec-users
> >
> _______________________________________________
> rspec-users mailing list
> rspec-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rspec-users
>

On Dec 17, 2007 10:55 AM, aslak hellesoy <aslak.hellesoy at gmail.com>
wrote:
> On Dec 17, 2007 4:12 PM, David Chelimsky <dchelimsky at gmail.com>
wrote:
> > On Dec 17, 2007 9:06 AM, Jeroen Houben <jeroen.houben at
lostboys.nl> wrote:
> > > >>
> > > >> ... Just received the message by David. This will be
fixed in 1.1.1
> > > >
> > > > Jeroen - just for good measure, would you kindly update from
trunk and
> > > > verify that it does in fact fix your issue?
> > >
> > > Finished in 3.42683 seconds
> > >
> > > 176 examples, 0 failures, 3 pending
> >
> > Thanks. I''ll get the release out by tonight.
> >
>
> Awesome David.
>
> Do we have any specs that will catch this regression if it''s
reintroduced?
I didn''t look and don''t have time right now. The fix was
committed in
r3158. Can  you check the diff?
>
> Aslak
>
>
> > Cheers,
> > David
> >
> >
> > >
> > > :)
> > >
> > >
> > > _______________________________________________
> > > rspec-users mailing list
> > > rspec-users at rubyforge.org
> > > http://rubyforge.org/mailman/listinfo/rspec-users
> > >
> > _______________________________________________
> > rspec-users mailing list
> > rspec-users at rubyforge.org
> > http://rubyforge.org/mailman/listinfo/rspec-users
> >
> _______________________________________________
> rspec-users mailing list
> rspec-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rspec-users
>

On Dec 17, 2007 5:58 PM, David Chelimsky <dchelimsky at gmail.com>
wrote:
> On Dec 17, 2007 10:55 AM, aslak hellesoy <aslak.hellesoy at
gmail.com> wrote:
> > On Dec 17, 2007 4:12 PM, David Chelimsky <dchelimsky at
gmail.com> wrote:
> > > On Dec 17, 2007 9:06 AM, Jeroen Houben <jeroen.houben at
lostboys.nl> wrote:
> > > > >>
> > > > >> ... Just received the message by David. This will
be fixed in 1.1.1
> > > > >
> > > > > Jeroen - just for good measure, would you kindly update
from trunk and
> > > > > verify that it does in fact fix your issue?
> > > >
> > > > Finished in 3.42683 seconds
> > > >
> > > > 176 examples, 0 failures, 3 pending
> > >
> > > Thanks. I''ll get the release out by tonight.
> > >
> >
> > Awesome David.
> >
> > Do we have any specs that will catch this regression if it''s
reintroduced?
>
> I didn''t look and don''t have time right now. The fix was
committed in
> r3158. Can  you check the diff?
>
I''ve added some more functional specs (r3172) that should prevent any
transaction related bugs from coming back in.

Jeroen - if you find more bugs, please report what Rails and Ruby
versions you are using - it took me a while to realise that you were
on Rails 1.2.3 (David told me)
>
> >
> > Aslak
> >
> >
> > > Cheers,
> > > David
> > >
> > >
> > > >
> > > > :)
> > > >
> > > >
> > > > _______________________________________________
> > > > rspec-users mailing list
> > > > rspec-users at rubyforge.org
> > > > http://rubyforge.org/mailman/listinfo/rspec-users
> > > >
> > > _______________________________________________
> > > rspec-users mailing list
> > > rspec-users at rubyforge.org
> > > http://rubyforge.org/mailman/listinfo/rspec-users
> > >
> > _______________________________________________
> > rspec-users mailing list
> > rspec-users at rubyforge.org
> > http://rubyforge.org/mailman/listinfo/rspec-users
> >
> _______________________________________________
> rspec-users mailing list
> rspec-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rspec-users
>

On 12/17/07 7:14 PM, "aslak hellesoy" <aslak.hellesoy at
gmail.com> wrote:
> On Dec 17, 2007 5:58 PM, David Chelimsky <dchelimsky at gmail.com>
wrote:
>> On Dec 17, 2007 10:55 AM, aslak hellesoy <aslak.hellesoy at
gmail.com> wrote:
>>> On Dec 17, 2007 4:12 PM, David Chelimsky <dchelimsky at
gmail.com> wrote:
>>>> On Dec 17, 2007 9:06 AM, Jeroen Houben <jeroen.houben at
lostboys.nl> wrote:
>>>>>>> 
>>>>>>> ... Just received the message by David. This will
be fixed in 1.1.1
>>>>>> 
>>>>>> Jeroen - just for good measure, would you kindly update
from trunk and
>>>>>> verify that it does in fact fix your issue?
>>>>> 
>>>>> Finished in 3.42683 seconds
>>>>> 
>>>>> 176 examples, 0 failures, 3 pending
>>>> 
>>>> Thanks. I''ll get the release out by tonight.
>>>> 
>>> 
>>> Awesome David.
>>> 
>>> Do we have any specs that will catch this regression if
it''s reintroduced?
>> 
>> I didn''t look and don''t have time right now. The fix
was committed in
>> r3158. Can  you check the diff?
>> 
> 
> I''ve added some more functional specs (r3172) that should prevent
any
> transaction related bugs from coming back in.
> 
> Jeroen - if you find more bugs, please report what Rails and Ruby
> versions you are using - it took me a while to realise that you were
> on Rails 1.2.3 (David told me)
Don''t know if it makes a difference, but I was on 1.2.6

Cheers,

Jeroen