thr3ads.net - rspec users - [rspec-users] given

If this information is useful, please help other people find it:
Share via:

Yurii Rashkovskii

2007-Sep-21 13:31 UTC

[rspec-users] given_it

Hello,

Just decided to check whether I am doing something that makes sense  
or not. I was thinking about how cool would it be to re-use examples  
(just like we reuse story scenarios with GivenScenario). I was not  
sure if this possibility already exists in rspec (and, honestly, was   
lazy to check), so I have created this helper:

def given_it(name)
   example_definition = behaviour.example_definitions.find{|i|  
i.description == name }
   instance_eval(&example_definition.example_block)
end

so it is possible to write things like

it "should do something after another action" do
   given_it "should successfully do another action"
   do_something.should be_fine
end

Is there anything already in rspec that allows me to do the same  
stuff? Or was it a bad idea at all?

Yurii.

David Chelimsky

2007-Sep-21 19:49 UTC

head link

[rspec-users] given_it

Have you seen shared behaviours?

On 9/21/07, Yurii Rashkovskii <yrashk at gmail.com>
wrote:> Hello,
>
> Just decided to check whether I am doing something that makes sense
> or not. I was thinking about how cool would it be to re-use examples
> (just like we reuse story scenarios with GivenScenario). I was not
> sure if this possibility already exists in rspec (and, honestly, was
> lazy to check), so I have created this helper:
>
> def given_it(name)
>    example_definition = behaviour.example_definitions.find{|i|
> i.description == name }
>    instance_eval(&example_definition.example_block)
> end
>
> so it is possible to write things like
>
> it "should do something after another action" do
>    given_it "should successfully do another action"
>    do_something.should be_fine
> end
>
> Is there anything already in rspec that allows me to do the same
> stuff? Or was it a bad idea at all?
>
> Yurii.
> _______________________________________________
> rspec-users mailing list
> rspec-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rspec-users
>

Yurii Rashkovskii

2007-Sep-21 20:01 UTC

head link

[rspec-users] given_it

Sure! But they are not the same stuff. What I was looking for is

it "should do #1" do
  ...
end
it "should do X after #1" do
  given_it "should do #1"
  should do_x
end
it "should do Y after #1" do
  given_it "should do #1"
  should do_y
end
it "should do Z after #1 Y" do
  given_it "should do Y after #1"
  should do_z
end

Or am I missing somethiing?


On Sep 21, 2007, at 10:49 PM, David Chelimsky wrote:
> Have you seen shared behaviours?
>
> On 9/21/07, Yurii Rashkovskii <yrashk at gmail.com> wrote:
>> Hello,
>>
>> Just decided to check whether I am doing something that makes sense
>> or not. I was thinking about how cool would it be to re-use examples
>> (just like we reuse story scenarios with GivenScenario). I was not
>> sure if this possibility already exists in rspec (and, honestly, was
>> lazy to check), so I have created this helper:
>>
>> def given_it(name)
>>    example_definition = behaviour.example_definitions.find{|i|
>> i.description == name }
>>    instance_eval(&example_definition.example_block)
>> end
>>
>> so it is possible to write things like
>>
>> it "should do something after another action" do
>>    given_it "should successfully do another action"
>>    do_something.should be_fine
>> end
>>
>> Is there anything already in rspec that allows me to do the same
>> stuff? Or was it a bad idea at all?
>>
>> Yurii.
>> _______________________________________________
>> 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

David Chelimsky

2007-Sep-21 20:03 UTC

head link

[rspec-users] given_it

On 9/21/07, Yurii Rashkovskii <yrashk at gmail.com>
wrote:> Sure! But they are not the same stuff. What I was looking for is
>
> it "should do #1" do
>   ...
> end
> it "should do X after #1" do
>   given_it "should do #1"
>   should do_x
> end
> it "should do Y after #1" do
>   given_it "should do #1"
>   should do_y
> end
> it "should do Z after #1 Y" do
>   given_it "should do Y after #1"
>   should do_z
> end
Sorry mate - that just seems like endless confusion - the examples
should never rely on each other that way - different animal from
scenarios that involve steps.
>
> Or am I missing somethiing?
>
>
> On Sep 21, 2007, at 10:49 PM, David Chelimsky wrote:
>
> > Have you seen shared behaviours?
> >
> > On 9/21/07, Yurii Rashkovskii <yrashk at gmail.com> wrote:
> >> Hello,
> >>
> >> Just decided to check whether I am doing something that makes
sense
> >> or not. I was thinking about how cool would it be to re-use
examples
> >> (just like we reuse story scenarios with GivenScenario). I was not
> >> sure if this possibility already exists in rspec (and, honestly,
was
> >> lazy to check), so I have created this helper:
> >>
> >> def given_it(name)
> >>    example_definition = behaviour.example_definitions.find{|i|
> >> i.description == name }
> >>    instance_eval(&example_definition.example_block)
> >> end
> >>
> >> so it is possible to write things like
> >>
> >> it "should do something after another action" do
> >>    given_it "should successfully do another action"
> >>    do_something.should be_fine
> >> end
> >>
> >> Is there anything already in rspec that allows me to do the same
> >> stuff? Or was it a bad idea at all?
> >>
> >> Yurii.
> >> _______________________________________________
> >> 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
>

lancecarlson at gmail.com

2007-Sep-21 20:11 UTC

head link

[rspec-users] given_it

This concept violates oo if you want to test whether or not r method calls
another method you would do that in the test and label it occordingly. What you
are defining is such a narrow use case I think it would be abused
Sent via BlackBerry from T-Mobile

-----Original Message-----
From: "David Chelimsky" <dchelimsky at gmail.com>

Date: Fri, 21 Sep 2007 22:03:30 
To:rspec-users <rspec-users at rubyforge.org>
Subject: Re: [rspec-users] given_it


On 9/21/07, Yurii Rashkovskii <yrashk at gmail.com>
wrote:> Sure! But they are not the same stuff. What I was looking for is
>
> it "should do #1" do
>   ...
> end
> it "should do X after #1" do
>   given_it "should do #1"
>   should do_x
> end
> it "should do Y after #1" do
>   given_it "should do #1"
>   should do_y
> end
> it "should do Z after #1 Y" do
>   given_it "should do Y after #1"
>   should do_z
> end
Sorry mate - that just seems like endless confusion - the examples
should never rely on each other that way - different animal from
scenarios that involve steps.
>
> Or am I missing somethiing?
>
>
> On Sep 21, 2007, at 10:49 PM, David Chelimsky wrote:
>
> > Have you seen shared behaviours?
> >
> > On 9/21/07, Yurii Rashkovskii <yrashk at gmail.com> wrote:
> >> Hello,
> >>
> >> Just decided to check whether I am doing something that makes
sense
> >> or not. I was thinking about how cool would it be to re-use
examples
> >> (just like we reuse story scenarios with GivenScenario). I was not
> >> sure if this possibility already exists in rspec (and, honestly,
was
> >> lazy to check), so I have created this helper:
> >>
> >> def given_it(name)
> >>    example_definition = behaviour.example_definitions.find{|i|
> >> i.description == name }
> >>    instance_eval(&example_definition.example_block)
> >> end
> >>
> >> so it is possible to write things like
> >>
> >> it "should do something after another action" do
> >>    given_it "should successfully do another action"
> >>    do_something.should be_fine
> >> end
> >>
> >> Is there anything already in rspec that allows me to do the same
> >> stuff? Or was it a bad idea at all?
> >>
> >> Yurii.
> >> _______________________________________________
> >> 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

Yurii Rashkovskii

2007-Sep-21 20:12 UTC

head link

[rspec-users] given_it

>> Sure! But they are not the same stuff. What I was looking for is
>>
>> it "should do #1" do
>>   ...
>> end
>> it "should do X after #1" do
>>   given_it "should do #1"
>>   should do_x
>> end
>> it "should do Y after #1" do
>>   given_it "should do #1"
>>   should do_y
>> end
>> it "should do Z after #1 Y" do
>>   given_it "should do Y after #1"
>>   should do_z
>> end
>
> Sorry mate - that just seems like endless confusion - the examples
> should never rely on each other that way - different animal from
> scenarios that involve steps.
I agree that they should not rely on each other _indirectly_. But why  
they should not _reuse_ each other? I find it very useful for  
testing, say, different aspects of, say, controller''s behavior. I  
simply do not want to put all checks into one example, like it should  
do this, this and this and in result we should have this, this and  
that. How do you propose to test different aspects of single actions  
in a nice way?

lancecarlson at gmail.com

2007-Sep-21 20:14 UTC

head link

[rspec-users] given_it

If you want to reuse stuff in your tests put those test methods in a separate
module and include them into you specs
Sent via BlackBerry from T-Mobile

-----Original Message-----
From: Yurii Rashkovskii <yrashk at gmail.com>

Date: Fri, 21 Sep 2007 23:12:50 
To:rspec-users <rspec-users at rubyforge.org>
Subject: Re: [rspec-users] given_it


>> Sure! But they are not the same stuff. What I was looking for is
>>
>> it "should do #1" do
>>   ...
>> end
>> it "should do X after #1" do
>>   given_it "should do #1"
>>   should do_x
>> end
>> it "should do Y after #1" do
>>   given_it "should do #1"
>>   should do_y
>> end
>> it "should do Z after #1 Y" do
>>   given_it "should do Y after #1"
>>   should do_z
>> end
>
> Sorry mate - that just seems like endless confusion - the examples
> should never rely on each other that way - different animal from
> scenarios that involve steps.
I agree that they should not rely on each other _indirectly_. But why  
they should not _reuse_ each other? I find it very useful for  
testing, say, different aspects of, say, controller''s behavior. I  
simply do not want to put all checks into one example, like it should  
do this, this and this and in result we should have this, this and  
that. How do you propose to test different aspects of single actions  
in a nice way?
_______________________________________________
rspec-users mailing list
rspec-users at rubyforge.org
http://rubyforge.org/mailman/listinfo/rspec-users

Yurii Rashkovskii

2007-Sep-21 20:17 UTC

head link

[rspec-users] given_it

Then I will have tens of methods that in fact has nothing really  
different from reused examples

On Sep 21, 2007, at 11:14 PM, lancecarlson at gmail.com wrote:
> If you want to reuse stuff in your tests put those test methods in  
> a separate module and include them into you specs
> Sent via BlackBerry from T-Mobile
>

David Chelimsky

2007-Sep-21 20:19 UTC

head link

[rspec-users] given_it

On 9/21/07, Yurii Rashkovskii <yrashk at gmail.com>
wrote:> Then I will have tens of methods that in fact has nothing really
> different from reused examples
But, as methods, they are more clearly differentiated for reuse.

The problem is that if I reuse an example, and then decide to change
that example, I may accidentally change the meaning of other examples.
This is possible w/ helper methods too, but I think it is less likely.

You can keep pursuing this if you like, but I can tell you right now
that it is very likely not going to happen. We already have a few
different ways to deal w/ reuse and this one seems more confusing to
me than helpful.

FWIW,
David
>
> On Sep 21, 2007, at 11:14 PM, lancecarlson at gmail.com wrote:
>
> > If you want to reuse stuff in your tests put those test methods in
> > a separate module and include them into you specs
> > Sent via BlackBerry from T-Mobile
> >
> _______________________________________________
> rspec-users mailing list
> rspec-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rspec-users
>

Yurii Rashkovskii

2007-Sep-21 20:26 UTC

head link

[rspec-users] given_it

Ok, let me try to explain. I have some controllers that are doing  
some kind of "main" action and some additional ones. These actions  
are belong to different aspects of logic and thus
I''d prefer to test them separately to have clean and nice code that  
can be easily modified.

Then, what''s better to read?

it "should do main action" do
   do_main_action
end

it "should set some value in session if there was no another specific  
value in session before" do
   session[:a] = 1
   do_main_action
   session[:b].should == b
end

def do_main_action
       post :create, ...
end

or

it "should do main action" do
   post :create, ...
end

it "should set some value in session if there was no another specific  
value in session before" do
   session[:a] = 1
   given_it "should do main action"
   session[:b].should == b
end

I personally like second one.

P.S. Examples above are surely artificial.



On Sep 21, 2007, at 11:11 PM, lancecarlson at gmail.com wrote:
> This concept violates oo if you want to test whether or not r  
> method calls another method you would do that in the test and label  
> it occordingly. What you are defining is such a narrow use case I  
> think it would be abused
> Sent via BlackBerry from T-Mobile
>
> -----Original Message-----
> From: "David Chelimsky" <dchelimsky at gmail.com>
>
> Date: Fri, 21 Sep 2007 22:03:30
> To:rspec-users <rspec-users at rubyforge.org>
> Subject: Re: [rspec-users] given_it
>
>
> On 9/21/07, Yurii Rashkovskii <yrashk at gmail.com> wrote:
>> Sure! But they are not the same stuff. What I was looking for is
>>
>> it "should do #1" do
>>   ...
>> end
>> it "should do X after #1" do
>>   given_it "should do #1"
>>   should do_x
>> end
>> it "should do Y after #1" do
>>   given_it "should do #1"
>>   should do_y
>> end
>> it "should do Z after #1 Y" do
>>   given_it "should do Y after #1"
>>   should do_z
>> end
>
> Sorry mate - that just seems like endless confusion - the examples
> should never rely on each other that way - different animal from
> scenarios that involve steps.
>
>>
>> Or am I missing somethiing?
>>
>>
>> On Sep 21, 2007, at 10:49 PM, David Chelimsky wrote:
>>
>>> Have you seen shared behaviours?
>>>
>>> On 9/21/07, Yurii Rashkovskii <yrashk at gmail.com> wrote:
>>>> Hello,
>>>>
>>>> Just decided to check whether I am doing something that makes
sense
>>>> or not. I was thinking about how cool would it be to re-use  
>>>> examples
>>>> (just like we reuse story scenarios with GivenScenario). I was
not
>>>> sure if this possibility already exists in rspec (and,
honestly,
>>>> was
>>>> lazy to check), so I have created this helper:
>>>>
>>>> def given_it(name)
>>>>    example_definition = behaviour.example_definitions.find{|i|
>>>> i.description == name }
>>>>    instance_eval(&example_definition.example_block)
>>>> end
>>>>
>>>> so it is possible to write things like
>>>>
>>>> it "should do something after another action" do
>>>>    given_it "should successfully do another action"
>>>>    do_something.should be_fine
>>>> end
>>>>
>>>> Is there anything already in rspec that allows me to do the
same
>>>> stuff? Or was it a bad idea at all?
>>>>
>>>> Yurii.
>>>> _______________________________________________
>>>> 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
> _______________________________________________
> rspec-users mailing list
> rspec-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rspec-users

Yurii Rashkovskii

2007-Sep-21 20:30 UTC

head link

[rspec-users] given_it

David,

On Sep 21, 2007, at 11:19 PM, David Chelimsky wrote:
> On 9/21/07, Yurii Rashkovskii <yrashk at gmail.com> wrote:
>> Then I will have tens of methods that in fact has nothing really
>> different from reused examples
>
> But, as methods, they are more clearly differentiated for reuse.
>
> The problem is that if I reuse an example, and then decide to change
> that example, I may accidentally change the meaning of other examples.
> This is possible w/ helper methods too, but I think it is less likely.
I see no difference, really. You can break anything, either helper or  
example. But when using helpers, you''d likely to have worse code  
(just IMHO)
>
> You can keep pursuing this if you like, but I can tell you right now
> that it is very likely not going to happen. We already have a few
> different ways to deal w/ reuse and this one seems more confusing to
> me than helpful.
>
I am not going to change anything, that''s simply not my intention. My  
intention was to check what people think about this approach. I just  
found it very convenient
for my coding experience. I do not need spend time extracting  
helpers, I do not need to keep single line examples that call  
helpers. I just write down pieces of logic, that''s it.

Anyway, thank you for your feedback.

Yurii.

Dan North

2007-Sep-21 20:32 UTC

head link

[rspec-users] given_it

I have to agree with David. (It''s in the contract ;)

Scenarios are reusable as steps in order to chain multiple scenarios 
together (to describe the various coarse-grained stages of a workflow, 
state machine or wizard, for example).

At an example level it is just confusing. I''ve noticed that if I find 
myself wanting to do that, it''s an indication that the objects are too 
tightly coupled or dependent on one another. There is usually a cleaner 
abstraction in there waiting to be discovered that will mean you don''t 
need to chain examples in this way.

fwiw GivenScenario was introduced into JBehave to work around the 
constraint that scenarios were only allowed to be of the form 
Given/When/Then in the early days. Then we relaxed scenarios to allow 
mixing up steps in any order (given.. when.. then.. when.. then..) which 
made them read much more naturally. GivenScenarios aren''t used much now
other than in the specific situations I described above (state machines, 
etc).

Cheers,
Dan

David Chelimsky wrote:> On 9/21/07, Yurii Rashkovskii <yrashk at gmail.com> wrote:
>   
>> Sure! But they are not the same stuff. What I was looking for is
>>
>> it "should do #1" do
>>   ...
>> end
>> it "should do X after #1" do
>>   given_it "should do #1"
>>   should do_x
>> end
>> it "should do Y after #1" do
>>   given_it "should do #1"
>>   should do_y
>> end
>> it "should do Z after #1 Y" do
>>   given_it "should do Y after #1"
>>   should do_z
>> end
>>     
>
> Sorry mate - that just seems like endless confusion - the examples
> should never rely on each other that way - different animal from
> scenarios that involve steps.
>
>   
>> Or am I missing somethiing?
>>
>>
>> On Sep 21, 2007, at 10:49 PM, David Chelimsky wrote:
>>
>>     
>>> Have you seen shared behaviours?
>>>
>>> On 9/21/07, Yurii Rashkovskii <yrashk at gmail.com> wrote:
>>>       
>>>> Hello,
>>>>
>>>> Just decided to check whether I am doing something that makes
sense
>>>> or not. I was thinking about how cool would it be to re-use
examples
>>>> (just like we reuse story scenarios with GivenScenario). I was
not
>>>> sure if this possibility already exists in rspec (and,
honestly, was
>>>> lazy to check), so I have created this helper:
>>>>
>>>> def given_it(name)
>>>>    example_definition = behaviour.example_definitions.find{|i|
>>>> i.description == name }
>>>>    instance_eval(&example_definition.example_block)
>>>> end
>>>>
>>>> so it is possible to write things like
>>>>
>>>> it "should do something after another action" do
>>>>    given_it "should successfully do another action"
>>>>    do_something.should be_fine
>>>> end
>>>>
>>>> Is there anything already in rspec that allows me to do the
same
>>>> stuff? Or was it a bad idea at all?
>>>>
>>>> Yurii.
>>>> _______________________________________________
>>>> 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
>   
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://rubyforge.org/pipermail/rspec-users/attachments/20070921/120f9874/attachment.html

Reasonably Related Threads

Search for more seemingly similar threads

rspec users - Sep 2007 - given_it

[rspec-users] given_it

[rspec-users] given_it

[rspec-users] given_it

[rspec-users] given_it

[rspec-users] given_it

[rspec-users] given_it

[rspec-users] given_it

[rspec-users] given_it

[rspec-users] given_it

[rspec-users] given_it

[rspec-users] given_it

[rspec-users] given_it

Reasonably Related Threads