Why does gherkin depend on Cucumber? As far as I can see, it only needs Cucumber for its tests, so that shouldn't be a production dependency.
>
>
> --
> You received this message because you are subscribed to the Google Groups "Cukes" group.
> To post to this group, send email to cu...@googlegroups.com.
> To unsubscribe from this group, send email to cukes+un...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/cukes?hl=en.
>
cheers,
Matt
If you say that one dependency can disappear that would even be
better. Circular dependencies are evil. So this is a bug?
On Sep 27, 1:52 am, Matt Wynne <m...@mattwynne.net> wrote:
> On 27 Sep 2010, at 03:05, moritz wrote:
>
>
>
>
>
> > There is a circular dependency between the cucumber and gherkin gems
> > that makes cleanup hard:
>
> > You have requested to uninstall the gem:
> > cucumber-0.8.5
> > ...
> > gherkin-2.1.5 depends on [cucumber (~> 0.8.5)]
> > ...
>
> > You have requested to uninstall the gem:
> > gherkin-2.1.5
> > cucumber-0.8.5 depends on [gherkin (~> 2.1.4)]
>
> > Even though, I have newer versions in the gem repository
>
> > cucumber (0.9.0, 0.8.5)
> > gherkin (2.2.4, 2.1.5)
>
> > neither of them would meet the '~>' requirement.
>
> > A similar circular dependency is present in the new gems as well.
> > Could this possibly be relaxed on one end using '>='? That would make
> > cleanup easier, as it is not always easy to discern that it is ok to
> > use 'Y' on the continue question.
>
> Why does gherkin depend on Cucumber? As far as I can see, it only needs Cucumber for its tests, so that shouldn't be a production dependency.
>
First, if you cleanup a bunch of gems, it is not immediately obvious
why the cucumber/gherkin combination would not uninstall,
and yes, it
requires more attention and more careful analysis, if you have to
answer 'Y'. Normally, I would always opt for 'n'.
For example, gherkin 2.2 came out first and even though it was locally
installed, I couldn't uninstall gherkin 2.1.5, because cucumber 0.8
depended on it (~> 2.1.4). That was as expected and correct.
But even after cucumber 0.9 being available, which is now depending on
gherkin 2.2, again via ~>, the circular dependency, using ~> in both
directions keeps the old pair from being easily removed.
I think that relaxing maybe the development dependency from ~> to >=
would fix the issue, but you would have to decide if that is
semantically possible or would break something.
To unsubscribe from this group, send email to cukes+un...@googlegroups.com.