[Cucumber] --must-not-pass option

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
8 messages Options
Reply | Threaded
Open this post in threaded view
|

[Cucumber] --must-not-pass option

Perryn Fowler
Hi there,

I recently created a fork of Cucumber for use on my current project,
and Aslak asked me to start a discussion here to see if it is
something we would want to merge back into a Cucumber release.

I have attempted to explain the reasoning behind my changes here :
http://www.jroller.com/perryn/entry/bdd_on_a_multi_disciplined

The rake tasks mentioned are really just window dressing for a new
--must-not-pass option that fails the suite if any individual Scenario
passes.

Thoughts?

Pez

--
-----------------------
Perryn Fowler
ThoughtWorks
_______________________________________________
rspec-devel mailing list
[hidden email]
http://rubyforge.org/mailman/listinfo/rspec-devel
Reply | Threaded
Open this post in threaded view
|

Re: [Cucumber] --must-not-pass option

Aslak Hellesoy

Hi there,

I recently created a fork of Cucumber for use on my current project,
and Aslak asked me to start a discussion here to see if it is
something we would want to merge back into a Cucumber release.

I have attempted to explain the reasoning behind my changes here :
http://www.jroller.com/perryn/entry/bdd_on_a_multi_disciplined

The rake tasks mentioned are really just window dressing for a new
--must-not-pass option that fails the suite if any individual Scenario
passes.

Thoughts?

Thanks for a well-written blog post that explains the rationale behind it. I especially like the tie-in to WIP, since I'm a huge Kanban fan.
Let me mull it over for a couple of days.

Would anyone else find this useful?

Aslak
 

Pez

--
-----------------------
Perryn Fowler
ThoughtWorks
_______________________________________________
rspec-devel mailing list
[hidden email]
http://rubyforge.org/mailman/listinfo/rspec-devel


_______________________________________________
rspec-devel mailing list
[hidden email]
http://rubyforge.org/mailman/listinfo/rspec-devel
Reply | Threaded
Open this post in threaded view
|

Re: [Cucumber] --must-not-pass option

Ben Mabey-2
aslak hellesoy wrote:

>
>     Hi there,
>
>     I recently created a fork of Cucumber for use on my current project,
>     and Aslak asked me to start a discussion here to see if it is
>     something we would want to merge back into a Cucumber release.
>
>     I have attempted to explain the reasoning behind my changes here :
>     http://www.jroller.com/perryn/entry/bdd_on_a_multi_disciplined
>
>     The rake tasks mentioned are really just window dressing for a new
>     --must-not-pass option that fails the suite if any individual Scenario
>     passes.
>
>     Thoughts?
>
>
> Thanks for a well-written blog post that explains the rationale behind
> it. I especially like the tie-in to WIP, since I'm a huge Kanban fan.
> Let me mull it over for a couple of days.
>
> Would anyone else find this useful?
>
> Aslak

I like it.  I've been using a similar workflow with the tags @proposed
and @in_progress, and profiles 'proposed', 'in_progress', and 'done'.
I never thought of having the in-progress profile fail if they pass, but
it makes sense.  So, I really like the idea of having it in Cucumber.

I don't think that the rake tasks should be included in Cucumber, but
the --must-not-pass option is nice.  However, should we change it to
"--must-fail"?

-Ben

>  
>
>
>     Pez
>
>     --
>     -----------------------
>     Perryn Fowler
>     ThoughtWorks
>     _______________________________________________
>     rspec-devel mailing list
>     [hidden email] <mailto:[hidden email]>
>     http://rubyforge.org/mailman/listinfo/rspec-devel
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> rspec-devel mailing list
> [hidden email]
> http://rubyforge.org/mailman/listinfo/rspec-devel

_______________________________________________
rspec-devel mailing list
[hidden email]
http://rubyforge.org/mailman/listinfo/rspec-devel
Reply | Threaded
Open this post in threaded view
|

Re: [Cucumber] --must-not-pass option

Ben Mabey-2
Ben Mabey wrote:

> aslak hellesoy wrote:
>>
>>     Hi there,
>>
>>     I recently created a fork of Cucumber for use on my current project,
>>     and Aslak asked me to start a discussion here to see if it is
>>     something we would want to merge back into a Cucumber release.
>>
>>     I have attempted to explain the reasoning behind my changes here :
>>     http://www.jroller.com/perryn/entry/bdd_on_a_multi_disciplined
>>
>>     The rake tasks mentioned are really just window dressing for a new
>>     --must-not-pass option that fails the suite if any individual
>> Scenario
>>     passes.
>>
>>     Thoughts?
>>
>>
>> Thanks for a well-written blog post that explains the rationale
>> behind it. I especially like the tie-in to WIP, since I'm a huge
>> Kanban fan.
>> Let me mull it over for a couple of days.
>>
>> Would anyone else find this useful?
>>
>> Aslak
>
> I like it.  I've been using a similar workflow with the tags @proposed
> and @in_progress, and profiles 'proposed', 'in_progress', and 'done'.
> I never thought of having the in-progress profile fail if they pass,
> but it makes sense.  So, I really like the idea of having it in Cucumber.
>
> I don't think that the rake tasks should be included in Cucumber, but
> the --must-not-pass option is nice.  However, should we change it to
> "--must-fail"?

Actually, now that I think about it maybe "--must-not-pass" is better
since it could also be pending and be fine.

>
> -Ben
>
>>  
>>
>>
>>     Pez
>>
>>     --
>>     -----------------------
>>     Perryn Fowler
>>     ThoughtWorks
>>     _______________________________________________
>>     rspec-devel mailing list
>>     [hidden email] <mailto:[hidden email]>
>>     http://rubyforge.org/mailman/listinfo/rspec-devel
>>
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> rspec-devel mailing list
>> [hidden email]
>> http://rubyforge.org/mailman/listinfo/rspec-devel
>
> _______________________________________________
> rspec-devel mailing list
> [hidden email]
> http://rubyforge.org/mailman/listinfo/rspec-devel

_______________________________________________
rspec-devel mailing list
[hidden email]
http://rubyforge.org/mailman/listinfo/rspec-devel
Reply | Threaded
Open this post in threaded view
|

Re: [Cucumber] --must-not-pass option

Perryn Fowler
In reply to this post by Ben Mabey-2
> I don't think that the rake tasks should be included in Cucumber, but the
> --must-not-pass option is nice.

Yeah, I mainly included them and their tests and features as a kind of
documentation in the code base for why --must-not-pass exists.

I found it difficult to explain why on earth I would want that option
without talking about my need for in_progress and regression builds,
so I added them as an explicit use case.

perhaps it would have been better to stick with the blog

Pez
_______________________________________________
rspec-devel mailing list
[hidden email]
http://rubyforge.org/mailman/listinfo/rspec-devel
Reply | Threaded
Open this post in threaded view
|

Re: [Cucumber] --must-not-pass option

Aslak Hellesoy


On Tue, May 12, 2009 at 8:17 AM, Perryn Fowler <[hidden email]> wrote:
> I don't think that the rake tasks should be included in Cucumber, but the
> --must-not-pass option is nice.

Yeah, I mainly included them and their tests and features as a kind of
documentation in the code base for why --must-not-pass exists.

I found it difficult to explain why on earth I would want that option
without talking about my need for in_progress and regression builds,
so I added them as an explicit use case.

Personally I would love to name the switch --work-in-progress or --wip (-w)
The help text would explain the failure logic.

I think this name would be more educational too.

WDYT?

Aslak
 

perhaps it would have been better to stick with the blog

Pez
_______________________________________________
rspec-devel mailing list
[hidden email]
http://rubyforge.org/mailman/listinfo/rspec-devel


_______________________________________________
rspec-devel mailing list
[hidden email]
http://rubyforge.org/mailman/listinfo/rspec-devel
Reply | Threaded
Open this post in threaded view
|

Re: [Cucumber] --must-not-pass option

Perryn Fowler
> Personally I would love to name the switch --work-in-progress or --wip (-w)
> The help text would explain the failure logic.
>
> I think this name would be more educational too.
>
> WDYT?

Yeah, I like that. I'll change it and remove the rake tasks on my fork
then send you another pull request



--
-----------------------
Perryn Fowler
ThoughtWorks
_______________________________________________
rspec-devel mailing list
[hidden email]
http://rubyforge.org/mailman/listinfo/rspec-devel
Reply | Threaded
Open this post in threaded view
|

Re: [Cucumber] --must-not-pass option

Aslak Hellesoy
On Wed, May 13, 2009 at 5:34 AM, Perryn Fowler <[hidden email]> wrote:

>> Personally I would love to name the switch --work-in-progress or --wip (-w)
>> The help text would explain the failure logic.
>>
>> I think this name would be more educational too.
>>
>> WDYT?
>
> Yeah, I like that. I'll change it and remove the rake tasks on my fork
> then send you another pull request
>

I have done a diff on your master branch - from the commit prior to
your first commit up to your last one.

github pull perryn
git checkout perryn/master
git diff bc51edfa5..849634ca

The diff is here:

http://gist.github.com/111636

This is a fairly substantial diff (1300+ lines), and I can't just
merge that in. It changes a lot more than it needs to in order to
implement this fairly simple feature (--must-not-pass/--wip).

Things I believe should be in a separate branch/diff:

* Rake task (ref Ben's comment)
* Mucking with the backtrace formatting
* Random whitespace

Could you please create a new branch (from my current master) and
implement *only* the --wip option? Then I'll pull it in.

Cheers,
Aslak

>
>
> --
> -----------------------
> Perryn Fowler
> ThoughtWorks
> _______________________________________________
> rspec-devel mailing list
> [hidden email]
> http://rubyforge.org/mailman/listinfo/rspec-devel
>
_______________________________________________
rspec-devel mailing list
[hidden email]
http://rubyforge.org/mailman/listinfo/rspec-devel