click-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gilberto C. Andrade" <gilbert...@gmail.com>
Subject Re: Removing OgnlOps, ConcurrentReaderHashMap and WebappLoader
Date Thu, 17 Jun 2010 12:20:59 GMT
On Thu, Jun 17, 2010 at 9:13 AM, Adrian A. <a.adrian.tech@gmail.com> wrote:

>
>             Also regarding the javassist dependency, this shouldn't be a
>>>>>            problem
>>>>>            cause the scope of the dependency is "provided":
>>>>>            http://github.com/jkuhnert/ognl/issues/closed#issue/2
>>>>>
>>>>        I assume that is maven-speak :) Does provided mean optional?
>>>>
>>>
>>>    I think it's something like this - if the users "provides" it at
>>>    runtime, than it will be used, otherwise it's not required.
>>>
>> No, it is a dependence!
>>
>> It means ognl won't work if you (you environment) don't have it[1]. (for
>> example: javax.annotation is provided, if you don't have it nothing
>> works!)
>> [1]
>> http://www.sonatype.com/books/mvnref-book/reference/pom-relationships-sect-project-dependencies.html#pom-relationships-sect-dependency-scope
>>
> Isn't this because such code is also using javax.annotation directly?
> (In the case of OGNL, Click code is not using javassist directly).
>
> If it would be a runtime dependence, than would be quite strange
> considering that I asked the actual developer of OGNL to remove the
> dependency if it's possible, and his answer was:
> "But the scope is "provided", so it isn't a problem"  :).
>
> Adrian.
> P.S. I'm so thankful that Click is using ANT :).
>
> With maven it would be simple to test: just change the pom.xml version
element of that lib and let it do the work.

How do you do in ant?

Mime
View raw message