Skip to main content


Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » Eclipse Projects » Gemini » Gemini artifactId naming inconsistent between sub-projects
Gemini artifactId naming inconsistent between sub-projects [message #550531] Fri, 30 July 2010 16:39 Go to next message
Werner Keil is currently offline Werner KeilFriend
Messages: 1087
Registered: July 2009
Senior Member
Hi,

While discussing naming conventions for Maven artifactId and groupId with a
fellow Eclipse committer to another project, we found Gemini is using 2
different patterns.
Here the difference in short again, to save you reading through all quotes
which I still include for those interested:

<groupId>org.eclipse.gemini.blueprint</groupId>
<artifactId>gemini-blueprint-io</artifactId>
and
<artifactId>gemini-blueprint</artifactId>
for the parent POM.

Gemini (Naming) uses the OSGi bundle style like this:
<groupId>org.eclipse.gemini.naming</groupId>
<artifactId>org.eclipse.gemini.naming.impl.bundle-Incubation </artifactId>
and
<artifactId>org.eclipse.gemini.naming</artifactId>

Mike suggested to bring it to your attention and try synchronize that.

As interested party and committer to other OSGi/Tycho/Maven enabled projects
I appreciate an outcome so we can apply this "blueprint" to our projects,
too.

TIA,
Werner

----- Original Message -----
From: "Mike Keith" <michael.keith@oracle.com>
To: "Werner Keil" <werner.keil@gmx.net>
Sent: Wednesday, July 28, 2010 8:31 PM
Subject: Re: Fw: Gemini Blueprint 1.0.0.M1 ships


> Hi Werner,
>
> You are referencing POM files from two different Gemini subprojects:
> Gemini Blueprint and Gemini Naming.
> These have two different project leads, and they likely just didn't
> coordinate the naming rules with each other. Perhaps you could post a
> new thread in the Gemini forum to point this out?
>
> Thanks,
> -Mike
>
> On 7/28/2010 2:39 PM, Werner Keil wrote:
>> Hi Mike,
>>
>> Thanks for the quick response.
>>
>> The difference is mainly how the projects use Maven's artifactId.
>>
>> The 1.0.0.M2 release of Gemini Blueprint (one of the sub-projects it
>> seems) proposes this as in previously attached POM:
>> <groupId>org.eclipse.gemini.blueprint</groupId>
>> <artifactId>gemini-blueprint-io</artifactId>
>> and
>> <artifactId>gemini-blueprint</artifactId>
>> for the parent POM.
>>
>> Gemini itself uses the OSGi bundle style like this:
>> <groupId>org.eclipse.gemini.naming</groupId>
>> <artifactId>org.eclipse.gemini.naming.impl.bundle-Incubation </artifactId>
>> and
>> <artifactId>org.eclipse.gemini.naming</artifactId>
>>
>> for the parent POM.
>>
>> Not a huge deal, but if Gemini Blueprint is supposed to be applied by
>> users and developers, why is it different from what Gemini itself uses?
>>
>> Thanks,
>> Werner
>>
>> ----- Original Message ----- From: "Mike Keith"
>> <michael.keith@oracle.com>
>> To: "Werner Keil" <werner.keil@gmx.net>
>> Sent: Wednesday, July 28, 2010 7:00 PM
>> Subject: Re: Fw: Gemini Blueprint 1.0.0.M1 ships
>>
>>
>>> Hi Werner,
>>>
>>> I am the overall Gemini lead and Costin Leau from VMware is the lead
>>> of the Gemini Blueprint subproject.
>>>
>>> I am not sure what you mean when you say that "the blueprint differs
>>> from the actual Gemini code". Which blueprint are you referring to?
>>> What is going in a different direction from what else?
>>>
>>> -Mike
>>>
>>> On 7/28/2010 1:34 PM, Werner Keil wrote:
>>>> Mike,
>>>>
>>>> I am not sure, if you're the overall Gemini lead or if it is Adrian
>>>> or someone else at vmware?
>>>> (I forgot his name, but yesterday a new colleague, I think Adam
>>>> attended the EC call, making SE/EE look rather good again ;-)
>>>>
>>>> Do you know why the blueprint differs from the actual Gemini code
>>>> especially with regards to how groupId and artifactId are formatted?
>>>> Is there a method behind it, or just different teams and their leads
>>>> going in slightly different directions (i.E. Oracle and vmware ;-)
>>>>
>>>> Thanks and Regards,
>>>> Werner
>>>>
>>>> ----- Original Message ----- From: "Werner Keil" <werner.keil@gmx.net>
>>>> Newsgroups: eclipse.gemini,eclipse.tycho
>>>> Sent: Tuesday, July 27, 2010 7:36 PM
>>>> Subject: Re: Gemini Blueprint 1.0.0.M1 ships
>>>>
>>>>
>>>>> Hi,
>>>>>
>>>>> What's the reason a POM from the Blueprint (hopefully outdated ;-)
>>>>> promotes
>>>>> one form of Maven artifactId, while Gemini itself uses another?!
>>>>>
>>>>> I lead an Eclipse project (UOMo) where the only non-Eclipse
>>>>> dependency is
>>>>> also shaped by a group of experts around unitsofmeasure.org.
>>>>>
>>>>> The discussion about the final artifact name lead us to your
>>>>> Blueprints
>>>>> which as of M2 still use a "dash" style naming of artifactId:
>>>>> <groupId>org.eclipse.gemini.blueprint</groupId>
>>>>> <artifactId>gemini-blueprint-io</artifactId>
>>>>>
>>>>> Gemini itself uses the OSGi bundle style like this:
>>>>> <groupId>org.eclipse.gemini.naming</groupId>
>>>>> <artifactId>org.eclipse.gemini.naming.impl.bundle-Incubation </artifactId>
>>>>>
>>>>>
>>>>> Any reasons why this is inconsistent?
>>>>>
>>>>> Hopefully Tycho may also shed some light on this and what it
>>>>> promotes as
>>>>> best practice for artifactId naming, especially for OSGi-enabled
>>>>> JAR-bundles?
>>>>>
>>>>> Thanks,
>>>>> Werner
>>>>>
>>>>>
>>>>>
>>
Re: Gemini artifactId naming inconsistent between sub-projects [message #550815 is a reply to message #550531] Wed, 04 August 2010 12:24 Go to previous messageGo to next message
Costin Leau is currently offline Costin LeauFriend
Messages: 45
Registered: February 2010
Member
Hi,

I can speak only about Gemini Blueprint. We wanted to transition of the existing Spring DM users as smooth as possible to Gemini and that's why the artifact names preserve the previous naming convention.

Even though the project is incubation, my understanding of the rules is that the "Incubation" term needs to appear in the bundle name but not necessarily in the actual artifact name.
Thus, to avoid unnecessary disruption in the future, I opted for no trailing -Incubation.

However, if my understanding is incorrect or there's a different consensus, I'll be happy to address this problem in the next milestone package.

Cheers,
Re: Gemini artifactId naming inconsistent between sub-projects [message #550911 is a reply to message #550815] Wed, 04 August 2010 15:42 Go to previous messageGo to next message
Werner Keil is currently offline Werner KeilFriend
Messages: 1087
Registered: July 2009
Senior Member
Hi,

Thanks for the reply. It is a bit odd, since Spring Framework has
implemented the new package-style naming convention almost everywhere.

Especially for a blueprint which has template character for potential
adoptors, I don't see why it should wait till the very last to publish such
improvements.

However, the decicion whether or not it should be stream-lined is with the
Gemini lead I assume.

Cheers,
Werner
Re: Gemini artifactId naming inconsistent between sub-projects [message #551110 is a reply to message #550815] Thu, 05 August 2010 09:39 Go to previous message
Werner Keil is currently offline Werner KeilFriend
Messages: 1087
Registered: July 2009
Senior Member
Btw, I fully agree, the mix of "." and "-" notation with trailing
"-Incubation" should be fixed if possible.

SpringSource generally uses an all "." pattern, e.g. "2.3.3.M2" for STS. So
even with the word "Incubation" it would better be
"org.eclipse.gemini.naming.impl.bundle.Incubation" if the word was still
needed ;-)
Previous Topic:Gemini Blueprint 1.0.0.M1 ships
Next Topic:Develop WABs with PDE
Goto Forum:
  


Current Time: Fri Apr 19 23:28:27 GMT 2024

Powered by FUDForum. Page generated in 0.03203 seconds
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.2.
Copyright ©2001-2010 FUDforum Bulletin Board Software

Back to the top