junit-devel

[Junit-devel] OSGi-ify JUnit / bug 2720888

[lukewpatterson <luk...@gm...>]

In bug 2720888 [1], David Saff recommended that I bring the discussion to
this dev list.

Please take a look at that bug to read the discussion so far... basically,
I'm hoping that the current OSGi-ification work for JUnit can be
consolidated and brought into the core.

I'm not trying to rush anyone to a decision.  Please provide any
comments/feedback and I'll try to work through solutions for any objections.

Since JUnit is a core dependency of so many projects, I believe that
OSGi-ification will have a beneficial ripple effect.  The barriers to entry
will be lowered if the testing infrastructure is OSGi-ready.


Best Regards,

Luke Patterson


[1]
http://sourceforge.net/tracker/?func=detail&aid=2720888&group_id=15278&atid=365278
-- 
View this message in context: http://www.nabble.com/OSGi-ify-JUnit---bug-2720888-tp22976430p22976430.html
Sent from the JUnit - Dev mailing list archive at Nabble.com.

Re: [Junit-devel] OSGi-ify JUnit / bug 2720888

[Mike F. <bi...@io...>]

I heard alot about OSGi as of late... many developers at work seem to be
behind it as a system. I will say that I've very little knowlege of the
subject.

However, from what I have seen/heard... OSGi is in the interprocess/service
communication space.  To that extent, I really don't see how I'd ever want
my tests framework to be exposed as a service.  Maybe I'm wrong here, I'd
love to hear arguments for it.

Could someone provide some simple use cases, where exposing JUnit as a
service would provide some value.

Big Mike

On Thu, Apr 9, 2009 at 2:12 PM, lukewpatterson <luk...@gm...>wrote:

>
> In bug 2720888 [1], David Saff recommended that I bring the discussion to
> this dev list.
>
> Please take a look at that bug to read the discussion so far... basically,
> I'm hoping that the current OSGi-ification work for JUnit can be
> consolidated and brought into the core.
>
> I'm not trying to rush anyone to a decision.  Please provide any
> comments/feedback and I'll try to work through solutions for any
> objections.
>
> Since JUnit is a core dependency of so many projects, I believe that
> OSGi-ification will have a beneficial ripple effect.  The barriers to entry
> will be lowered if the testing infrastructure is OSGi-ready.
>
>
> Best Regards,
>
> Luke Patterson
>
>
> [1]
>
> http://sourceforge.net/tracker/?func=detail&aid=2720888&group_id=15278&atid=365278
> --
> View this message in context:
> http://www.nabble.com/OSGi-ify-JUnit---bug-2720888-tp22976430p22976430.html
> Sent from the JUnit - Dev mailing list archive at Nabble.com.
>
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by:
> High Quality Requirements in a Collaborative Environment.
> Download a free trial of Rational Requirements Composer Now!
> http://p.sf.net/sfu/www-ibm-com
> _______________________________________________
> Junit-devel mailing list
> Jun...@li...
> https://lists.sourceforge.net/lists/listinfo/junit-devel
>

Re: [Junit-devel] OSGi-ify JUnit / bug 2720888

[Stephen S. <st...@st...>]

I'm interested in having this discussion, but like Mike I'm a little puzzled as 
to the problem we're trying to solve. In addition to use cases, could someone 
provide some tests?

Big Steve.


Mike Forsberg wrote:
> I heard alot about OSGi as of late... many developers at work seem to be 
> behind it as a system. I will say that I've very little knowlege of the 
> subject.
> 
> However, from what I have seen/heard... OSGi is in the 
> interprocess/service communication space.  To that extent, I really 
> don't see how I'd ever want my tests framework to be exposed as a 
> service.  Maybe I'm wrong here, I'd love to hear arguments for it.
> 
> Could someone provide some simple use cases, where exposing JUnit as a 
> service would provide some value.
> 
> Big Mike
> 
> On Thu, Apr 9, 2009 at 2:12 PM, lukewpatterson <luk...@gm... 
> <mailto:luk...@gm...>> wrote:
> 
> 
>     In bug 2720888 [1], David Saff recommended that I bring the
>     discussion to
>     this dev list.
> 
>     Please take a look at that bug to read the discussion so far...
>     basically,
>     I'm hoping that the current OSGi-ification work for JUnit can be
>     consolidated and brought into the core.
> 
>     I'm not trying to rush anyone to a decision.  Please provide any
>     comments/feedback and I'll try to work through solutions for any
>     objections.
> 
>     Since JUnit is a core dependency of so many projects, I believe that
>     OSGi-ification will have a beneficial ripple effect.  The barriers
>     to entry
>     will be lowered if the testing infrastructure is OSGi-ready.
> 
> 
>     Best Regards,
> 
>     Luke Patterson
> 
> 
>     [1]
>     http://sourceforge.net/tracker/?func=detail&aid=2720888&group_id=15278&atid=365278
>     <http://sourceforge.net/tracker/?func=detail&aid=2720888&group_id=15278&atid=365278>
>     --
>     View this message in context:
>     http://www.nabble.com/OSGi-ify-JUnit---bug-2720888-tp22976430p22976430.html
>     Sent from the JUnit - Dev mailing list archive at Nabble.com.
> 
> 
>     ------------------------------------------------------------------------------
>     This SF.net email is sponsored by:
>     High Quality Requirements in a Collaborative Environment.
>     Download a free trial of Rational Requirements Composer Now!
>     http://p.sf.net/sfu/www-ibm-com
>     _______________________________________________
>     Junit-devel mailing list
>     Jun...@li...
>     <mailto:Jun...@li...>
>     https://lists.sourceforge.net/lists/listinfo/junit-devel
> 
> 
> 
> ------------------------------------------------------------------------
> 
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by:
> High Quality Requirements in a Collaborative Environment.
> Download a free trial of Rational Requirements Composer Now!
> http://p.sf.net/sfu/www-ibm-com
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Junit-devel mailing list
> Jun...@li...
> https://lists.sourceforge.net/lists/listinfo/junit-devel
> 
> 
> ------------------------------------------------------------------------
> 
> 
> No virus found in this incoming message.
> Checked by AVG - www.avg.com 
> Version: 8.0.238 / Virus Database: 270.11.49/2050 - Release Date: 04/09/09 10:27:00
> 

Re: [Junit-devel] OSGi-ify JUnit / bug 2720888

[lukewpatterson <luk...@gm...>]

On Thu, Apr 9, 2009 at 3:19 PM, Mike Forsberg <bi...@io...> wrote:
> ... exposing JUnit as a service ...
>

I'll clarify the level of OSGi support I'm referring to.  I was only hoping
for compliance with the modularity/classloading portion of the OSGi spec,
not the dynamic services portion.

JUnit would not code to any OSGi API.  No JUnit API would change.  Existing
consumer code which compiles to JUnit wouldn't need to change.

The "modularity/classloading" portion I referred to above basically means
that JUnit just needs proper OSGi entries added to the META-INF/manifest.mf. 
With the proper entries, the JUnit jar would become a "bundle" in the OSGi
world.  A "bundle" is basically a "module".  i.e. a module contains
packages, packages contain classes, classes contain fields/methods

The manifest would explicitly declare (at the module level), which java
packages JUnit imports (depends on) and which java packages JUnit exports
(provides to others).

In order for the JUnit jar to be considered a first-class citizen (can
import/export packages directly) in the OSGi world, it needs to have valid
manifest entries.  Currently, when other projects are running in an OSGi
container, they can't directly run against the official JUnit jar.  They
must custom-wrap the JUnit jar so its packages can be available at runtime.

If you have some free minutes, please consider reading sections 1 through
1.4.3 of Neil Bartlett's "OSGi in practice" book [1].  I think he explains
the core concepts pretty well (a lot better than I could).


Thanks,

Luke


[1] http://neilbartlett.name/blog/osgibook/
-- 
View this message in context: http://www.nabble.com/OSGi-ify-JUnit---bug-2720888-tp22976430p22988877.html
Sent from the JUnit - Dev mailing list archive at Nabble.com.